Motion pictures of specific historical eclipses exist dating back more than 375 years. In this post, I’ll present some of the oldest surviving examples for viewing in motion for the first time (as far as I’m aware): eleven from the seventeenth century, four from the eighteenth century, and six from the nineteenth century. That makes twenty-one in all, representing nineteen different astronomical events from the terrestrial vantage points of Poland, England, France, and the United States.
Of course, the phrase “motion picture” is open to multiple interpretations. What I mean by it here is a sequence of images or “frames” that illustrate successive phases in the changing appearance of their subject.
If we like, we can display these images in rapid succession to create an illusion of movement—a “moving picture”—but they can also be shown simultaneously side by side for comparison. These are two equally legitimate modes of eduction for the same category of data. That said, people sometimes use the alternative expression “motion sequence photography” to describe motion pictures that aren’t, or weren’t, meant to function as moving pictures. Pictures of eclipses that fall into this category have been captured photographically since at least 1854, so those plainly count as motion sequence photography. However, a more accurate term for the eclipse sequences as a whole would be “motion sequence imagery,” since the earlier cases weren’t photographic, even though many of them involved the use of cameras. Some of the examples I’ll be sharing here, including the ones from the seventeenth century, are actual records captured while the eclipses they represent were underway, and these are easy to assimilate conceptually to documentary film or raw scientific footage. Others, from the eighteenth century, were prepared in advance so that people could consult them while watching an eclipse “live,” and these are more closely akin to animated cartoons or video simulations.
It’s possible that someone else might already have set one or more of these eclipse sequences into motion, but my searches haven’t turned anything up along those lines, which leads me to believe it probably hadn’t been tried before. The sequences themselves are reasonably well known among historians of astronomy and/or photography, but their status as motion pictures, with the potential to become moving pictures, seems not to have been duly recognized. It’s true that they don’t document microchronic motion—that is, motion rapid enough to sustain a smooth illusion of movement when “played back” at original speed. That’s an important distinction I don’t mean to downplay, and I’ve even put forward a candidate of my own as the pioneer of microchronic motion sequence photography: James Ross of Edinburgh. But we shouldn’t overplay the microchronic distinction either. After all, time-lapse motion pictures are still legitimate motion pictures—I don’t anticipate any controversy on that point. Indeed, time-lapse video is often used today as a means of showing eclipses (see recent examples on YouTube: solar here, lunar here). For whatever reason, though, the idea of making historical eclipse imagery “move” according to the same logic seems not to have been obvious.
And that goes for early astronomical image sequences more generally. The image sequence Jules Janssen captured with his “photographic revolver” of the transit of Venus on December 9, 1874, is often presented as a moving picture (see e.g. here). Twenty-odd years ago, Albert Van Helden of Rice University also animated a thirty-six-frame sequence of sunspots as drawn by Galileo over the course of a month in 1613 (here; some other cool animations based on Galileo’s drawings, e.g., here and here, throw extra data into the mix and so aren’t “straight” actualizations of their sources). Otherwise, the only animations of old astronomical image sequences I know about are ones I produced myself—check out my article “Moon Phase Animations (AD 650-1650),” which, despite the title, isn’t limited to moon phases.
Even when they’re weighed alongside those other cases, I believe the eclipse sequences I’m going to present here as moving pictures will still stand out as noteworthy and exciting. The astronomical sources I’ve animated in the past represent generalized cyclical phenomena rather than the unique details of specific historical moments. Galileo’s sunspot records were historically specific, but they documented phenomena that weren’t easily perceptible to the naked eye. By contrast, eclipses were major events witnessed by large numbers of people who invested them with significance, whether as ill omens or natural marvels, and each one took a markedly different form. Those are the events these motion pictures enable us to watch today in the same vicarious sense we can watch the happenings documented in old newsreels, even if there’s not much more to see than one circle passing over another circle.
Exhibit #1: Records of Eclipses made by Johannes Hevelius using a Camera Obscura (1639-1666).
The camera obscura looms large in the technical history of image capture. Wikipedia defines it as “the natural optical phenomenon that occurs when an image of a scene at the other side of a screen (or for instance a wall) is projected through a small hole in that screen as a reversed and inverted image (left to right and upside down) on a surface opposite to the opening.” Artists of the early modern period sometimes used camerae obscurae as drawing aids by projecting scenes onto paper and then tracing the images, as we see in the following sketch by Canaletto (Giovanni Antonio Canal, 1697-1768):
During the first half of the nineteenth century, the photographic camera evolved from the camera obscura by substituting a surface that undergoes a photochemical reaction when exposed to light, thereby eliminating the need for a human being to trace the projected patterns by hand. That breakthrough gave us photography as we know it today. Then, during the second half of the century, advances in “instantaneous” photography made it possible to capture sequences of images illustrating the phases of rapid motion, as famously demonstrated in 1878 through Eadweard Muybridge’s photographic documentation of a galloping horse.
Meanwhile, the camera obscura had also enjoyed a long association with the viewing of eclipses. Indeed, the first published image of the phenomenon itself, found in Gemma Frisius’s De radio astronomico et geometrico (1545), shows it projecting an image of a solar eclipse onto a wall—and not just any solar eclipse, either, but a specific one that took place on January 24, 1544, as seen at Louvain.
The same technique is still recommended today for viewing eclipses safely in real time through pinhole cameras. But sometimes the two pre-photographic uses of the camera obscura—as drawing aid and eclipse viewer—were exploited simultaneously. During the sixteenth century, astronomers seem to have contented themselves with recording the maximum occultation of eclipses, with each record corresponding to one single moment in time. Here, for example, is an illustration that appears in De radio astronomico et geometrico just a few pages before the picture of the camera obscura:
But with the advent of the helioscope in the early seventeenth century—more specialized apparatus designed, as Wikipedia states, for “projecting an image of the sun onto a white sheet of paper suspended in a darkened room with the use of a telescope”—we begin to see astronomers tracing whole sequences of images that documented successive moments in time, such as Galileo’s daily drawings of sunspots in June and July 1613. I’m not sure who first applied this same principle to eclipses, but Johannes Hevelius (1611-1687) is well known for having used helioscopes to capture image sequences for a number of them between 1639 and 1666. One version of the apparatus he used is described and depicted in his Selenographia (published 1647, here and here) and also shown in Machinae Coelestis (published 1674, here), but in the 1660s, he introduced an improved model with a more elegant provision for following the sun’s movement, also depicted in Machinae Coelestis (reproduced below from this source; see also an English translation of his written description here).
The last image shows Hevelius in the act of using his improved helioscope to record one phase in the occultation of an eclipse, something that would probably have taken only a second or two to trace by hand. It’s worth emphasizing that his goal in making such records was not to capture detailed solitary pictures of the sun or the moon, but rather to secure successive images corresponding to successive moments in time so that he could analyze the differences among them. With that in mind, it seems entirely reasonable to refer to this device as a “motion picture camera,” and the practice of using it as the camera obscura equivalent of Eadweard Muybridge’s motion sequence photography. Hevelius was seeking to study the movements of celestial bodies just as Muybridge sought to study the movements of animal bodies, but unlike Muybridge, Hevelius didn’t need to wait for instantaneous photography to come along because the movements he was studying were far slower and simpler.
One fifty-nine-frame sequence depicting an illustration of the solar eclipse of June 1, 1639, as seen at Gdańsk, appeared in Johannes Hevelius’s Machinae Coelestis, pars posterior (1679), here. Each frame is numbered, with the sequence running from left to right and from top to bottom.
My source for the digitized plate is Gallica, which states that the illustrator is unidentified (“non identifié”), but it’s marked at the bottom as engraved by the author (“Autor sculpsit”). The accompanying text discusses specifically how the sequence was obtained on a white tablet (tabella alba) opposite the hole (foramen) of a camera obscura—there are a lot of Latin passive constructions here, but Hevelius is definitely describing an event in which he personally took part. The original drawings were presumably destroyed in Hevelius’s catastrophic house fire of September 26, 1679, as were most copies of the printed book. Here’s my animation:
Hevelius had seen eclipses before 1639, according to his own account, but this seems to have been his first experience recording one. Numerical measurements were taken simultaneously of the area of occultation in digits, a unit of measure equal to a twelfth of a diameter, and minutes, which were apparently sixtieths of a digit. In the resulting table of figures, Hevelius notes which frames in his image sequence correspond to which measurements, as well as the time at which each measurement was taken, to a precision of ten tierces—a tierce being one sixtieth of a second. However, measurements were secured less frequently than images, and presumably not at exactly the same moments; and the intervals between measurements are somewhat irregular, hinting that the intervals between images may have been irregular as well. That said, each frame in the image sequence seems to correspond to approximately two or three minutes, and I’ve displayed the frames at a constant duration of 0.06 second in my animation on the understanding that they were supposed to represent consistent time intervals. Hevelius writes that the distance of the tabella from the foramen varied during the eclipse, leading to changes in the apparent diameter of the sun, but he argues that this isn’t a problem because the ratio of diameters remains consistently correct. In my animation, this results in an effect much as though we were zooming in and out over the course of the recording.
Some other similar eclipse sequences can also be found in Machinae Coelestis, but their time bases are irregular, which makes animating them a little more complicated. Three of them are lunar eclipses rather than solar ones. Here’s an eleven-frame sequence showing a partial lunar eclipse observed at Gdańsk on January 20, 1647 (source), also published in Selenographia (here):
This time, Hevelius opted to capture one frame per digit of occultation, corresponding to the times 9:19, 9:26, 9:30, 9:44, 10:00, 10:29, 10:53, 11:06, 11:16, 11:22, and 11:28 AM, and thereby representing irregular intervals of 7, 4, 14, 16, 29, 24, 13, 10, 6, and 6 minutes respectively. In my animation, I’ve set 0.01 second—the shortest duration I have available to work with—equal to two minutes, basing the duration of each frame on the interval between the time when it was captured and the time when the next frame was captured. Setting 0.01 second equal to one minute would have allowed me to establish an exactly correct time ratio, but I found that slowing things down by a factor of two spoiled the illusion of movement. The duration of the final frame is arbitrary.
An alternative strategy for handling the irregular time base would be to morph between successive source images while interpolating a variable number of frames each time based on the ratio of time intervals. Indeed, morphing between eclipse sequence images would allow us to slow down “playback” more generally without sacrificing smoothness of illusion. I haven’t tried this with an eclipse sequence yet, but I imagine the results would be comparable to these.
There are two more lunar eclipse records where the previous one came from. These present an image for every half-digit of occultation within their time spans and, in one case, another frame showing peak occultation. Here, animated on the same terms as the previous sequence, are an eight-frame sequence showing a lunar eclipse of August 27-28, 1654 (source)—
—and a thirteen-frame sequence showing a lunar eclipse of October 30, 1659 (source):
The majority of eclipse sequences depicted in Machinae Coelestis are solar. Hevelius occasionally succeeded in documenting an entire solar eclipse, as he had in 1639, but in many cases he was only able to secure a partial record; the accompanying Latin text often blames missing segments on cloudy weather. For the solar eclipses, I’ve set 0.01 second equal to one minute, so that they’re “played back” at twice the speed of the lunar eclipses. Both absolute timings are arbitrary, of course, as is true of any time-lapse video; my goal is simply to sustain a passable illusion of movement with a time base that shares a reasonably accurate time ratio with the original event.
A sixteen-frame sequence showing the solar eclipse of March 30, 1661 (source):
And finally an eighteen-frame sequence showing the solar eclipse of July 2, 1666 (source):
In each case except the very first one from 1639, Hevelius printed his eclipse sequences beneath larger-scale illustrations of the same data superimposed (source):
Exhibit #2: Predictive Illustration of the Solar Eclipse of April 22, 1715 (old style).
During the eighteenth century, image sequences that documented eclipses gave way to image sequences that predicted in advance what they would look like and when. Edmond Halley (1656-1742), the namesake of Halley’s Comet, accurately predicted the timing of the solar eclipse of May 3, 1715, to within four minutes, leading it to be known as “Halley’s Eclipse.” That same event also seems to have marked the introduction of a distinctive twenty-five-frame format for predictive illustrations of eclipses: a 5×5 grid with the moment of totality shown at the center and a time base running primarily from right to left and secondarily from top to bottom. As far as I’m aware, this format first appears in a broadside by “H. M.” entitled “A plain Description of the Sun’s appearance, in the Increase and Decrease of the Eclipse, which will happen on Fryday (in the morning) April the 22d, 1715.” Note that the date is given here in the “old style,” rather than according to the revised Gregorian calendar. A nice digital scan of this broadside appeared in Lambeth Palace Library Twitter feed, here, and shortly thereafter in Whewell’s Gazette, here; that version only shows the top portion of the broadside, but a lower-resolution image of the whole thing also turns up online, e.g., here, and here on page 6, with transcription of the accompanying text on page 7.
Below is my animation, presented at the full resolution of the scan available online, with the perimeter of the sun manually aligned in each of the twenty-five frames to match its position in the first frame.
I’ve set the GIF to display each frame for six hundredths of a second. The interval between frames was originally supposed to represent a consistent duration of five minutes and forty seconds, running from 8:08 AM through 10:24 AM.
Exhibit #3: Predictive Illustrations of the Solar Eclipse of May 11, 1724 (old style).
The same format was resurrected for the solar eclipse of May 22, 1724 (Gregorian), known in England at the time as May 11, 1724. It appeared, first of all, on another broadside, this one engraved by Emanuel Bowen and issued by Thomas Bowles. This time, my source image comes from Barry Lawrence Ruderman Antique Maps, LLC., who sold a copy of the broadside sometime in the recent past and continue to maintain the scan on their website. They were aware of only other extant copy, in Harvard University’s Houghton Library.
Another illustration closely resembling the preceding one—and presumably based on it (or vice versa)—appeared in Parker’s London News, on page three of the issues for May 4th, May 8th, and May 11th, 1724, skipping the issue for May 6th. For a description and transcription of the accompanying text, see here.
Copies of the three issues of Parker’s London News are available in the British Library’s 17th-18th Century Burney Collection Newspapers digital database. The same printing plate seems to have been used throughout the run, but due to vagaries of inking, creasing, aging, and digitization, each of the three issues in the database has different imperfections. For my animation, I’ve repeated each frame three times in a row as taken respectively from the issues for May 4th, May 8th, and May 11th. Since there are twenty-five frames in the sequence, the resulting animated GIF is seventy-five frames long.
Not bad, considering that this is based on black-and-white scans from a newspaper database that prioritizes sheer legibility of text.
Exhibit #4: Predictive Illustration of the Solar Eclipse of April 1, 1764.
The next illustration is also predictive but appears in a different format: two rows of six, running left to right. It was prepared by Nicole-Reine Lepaute (1723-1788), and my source for it is Gallica.
Lepaute’s images represent the situation at fifteen-minute intervals with one exception: the times indicated are 9:15, 9:30, 9:45, 10:00, 10:15, 10:30, 10:39 (midpoint), 11:00, 11:15, 11:30, 11:45, and 12:00. In my animation, I display the fifteen-minute frames for 0.05 second, the nine-minute frame for 0.03 second, and the twenty-one-minute frame for 0.07 second (thus, 0.01 second represents exactly three minutes).
Exhibit #5: Predictive illustration of the Solar Eclipse of June 16, 1806.
The practice of publishing predictive sequence imagery of eclipses carried over into America as well. The eclipse of June 16, 1806 was the subject of a three-frame sequence on the frontispiece of a comparatively common pamphlet entitled Darkness at Noon (Boston, 1806).
Three frames wouldn’t make for much of an animation, but the same eclipse was also the subject of a more promising twelve-frame sequence in a broadside entitled “Approaching Solar Eclipse,” published by John Poulson at Philadelphia. The best scan of this broadside which I’ve found online was used to advertise a copy sold by Goldberg Auctions for $3,173 back in 2009 (collectors take note: at the time of this writing, there’s another copy for sale for £4,000 from the Altea Gallery in London).
Exhibit #6: Predictive illustration of the Solar Eclipse of September 18, 1838.
Exhibit #7: Photographic Motion Pictures of the Solar Eclipse of May 26, 1854.
The solar eclipse of May 26, 1854 was the subject of at least one predictive sequence, published in the Burlington Free Press of May 19th. However, the invention of photography had then recently rekindled interest in capturing images of eclipses as they occurred, an impulse that seems otherwise to have lain dormant since Hevelius’s day. Gian Alessandro Majocchi reportedly took a pair of daguerreotypes of the solar eclipse of July 8, 1842, just before and just after totality, which aren’t known to survive; and solitary photographs exist of the solar eclipse of July 28, 1851 taken by John Whipple Adams at Harvard University and Johann Julius Friedrich Berkowski at the Royal Observatory in Königsberg (Prussia). But to the best of my knowledge, it was the eclipse of May 26, 1854, that gave us the oldest surviving sequence photographs of an eclipse. The Metropolitan Museum of Art holds a set of daguerreotypes taken by the brothers William and Frederick Langenheim, about which they state:
On May 26, 1854, the Langenheim brothers made eight sequential photographs of the first total eclipse of the sun visible in North America since the invention of photography. Although six other daguerreotypists and one calotypist are known to have documented the event, only these seven daguerreotypes survive. In the northern hemisphere, the moon always shadows the sun from right to left during a solar eclipse; these images therefore seem odd because they are, like all uncorrected daguerreotypes, reversed laterally as in a mirror.
Here’s an animation of the Langenheim image sequence (each frame = 0.1 second):
But I have to take exception to the museum’s claim that “only these seven daguerreotypes survive” out of the several photographic attempts to document the eclipse of May 26, 1854. Two plates of photographically captured images of this same eclipse, furnished courtesy of the Smithsonian Institution, were used to illustrate Astronomical Journal #77 (Volume 4, Number 6), dated December 5, 1854. One, apparently distributed with the issue itself, was engraved from daguerreotypes taken by Stephen Alexander at Ogdensburg, New York, as described in an accompanying article—although I can’t square the numbers of the images with the references in the article. The other plate is an actual photographic print of nineteen scaled-down photographs taken during the eclipse. We read:
The photographic sheet upon which Professor [W. H. C.] BARTLETT’s observations are reproduced, and which will be distributed with the succeeding number, contains his nineteen observations, within the space occupied by one in the original impression, and may be more satisfactorily examined by transmitted than by reflected light. It is hoped that it may prove the pioneer of a new class of accurate and permanently recorded astronomical observations.
In an accompanying article, Bartlett wrote:
Through the kindness and professional skill of Mr. VICTOR PREVOST, of New York, who was induced to come to West Point for the purpose, I was enabled…to obtain nineteen photographs of as many different phases. The corresponding times of these phases were carefully noted. The photographs were made by means of a small camera adapted to the eye-end of a refracting telescope of which the aperture was six inches and the solar focal length eight feet. The impressions were taken within the interval required to remove an opaque cap from the object-glass and replace it again without loss of time, so that the effect may, for all practical purposes, be said to have been instantaneous.
For a biography of the photographer, see Julie Mellby’s article “Victor Prevost: Painter, Lithographer, Photographer,” in History of Photography 35 (2011): 221-239. The photographs Prevost took were not daguerreotypes but calotypes: translucent wax paper negatives from which multiple positive salt prints could be made. Curiously enough, the plate in the Astronomical Journal is a negative, with a dark sun shown against a bright background, and different copies of the Astronomical Journal contain different versions of the plate with slightly different title frames and layout. For my animation experiments, I scanned the plate in the print copy held by the Indiana University Libraries, which matches the plate from the University of Toronto library available digitally through Archive.org, but not the plate in the Google Books copy from the University of Minnesota. Prevost had probably resorted to furnishing negative composites because of some technical challenge involved in large-scale duplication; in 1854, the use of photographic prints in published works was still a rare novelty in itself. Below I show my original scan of the plate on the left, with an inverted positive version of it on the right.
Here’s an animation of the plate in its original form, with durations based on a table showing the time at which each photograph was taken (0.01 second = 1 minute). I couldn’t resist making Prevost’s original title frame a part of the presentation.
Next, here’s the same thing inverted:
Finally, here’s the sequence reduced to grayscale and cropped, with a little contrast enhancement, and with the title frame removed:
Take a moment to savor what this is: a photographically captured motion picture from the year 1854, published contemporaneously to document the unfolding of an actual historical event.
Granted, the animation could stand improvement. Because of the low contrast, for example, I had trouble getting the sun to line up from frame to frame, so it bobs around more than I’d like. But the most substantial improvement would probably come from tapping a better source for the images themselves, which were reduced considerably in scale for the plate published in the Astronomical Journal. According to Mellby (p. 232), Prevost’s waxed-paper negatives of the eclipse survive in a volume in the library of the United States Military Academy at West Point, accession number 7364. Albums of full-sized salt prints were also made from these negatives. Mellby reports that Prevost’s personal copy of the eclipse album is held by the National Museum of American History, but that it now contains only three of the nineteen images. However, the Library of Congress appears to have a complete album—see the catalog record here, plus this lo-res teaser of a negative showing image number sixteen uncropped:
Exhibit #8: Published Photographs of the Solar Eclipse of August 7, 1869.
The Astronomical Journal would have reached only a limited audience, but Harper’s Weekly had a circulation in the neighborhood of 200,000 when a set of engravings made from photographs of the solar eclipse of August 7, 1869, adorned the front page of the issue for August 28th.
The caption reads: “SOLAR ECLIPSE, AUGUST 7, 1869—PHASES OF THE ECLIPSE, AS SEEN AT SHELBYVILLE, KENTUCKY, FROM THE BEGINNING TO THE POINT OF TOTALITY. [PHOTOGRAPHED BY J. A. WHIPPLE.]” The photographer was John Adams Whipple (1822-1891), who had also taken a single photograph of a partial solar eclipse on July 28, 1851. Inside the issue, we read:
Our illustrations are from observations made at Shelbyville, Kentucky. The photographs were taken by J. A. WHIPPLE, of Boston…. The telescope used for photographing was one from Cambridge. The glass was 6 inches diameter with 7 1/2 feet focal length. It was arranged for photography, giving an image of the sun about three-quarters of an inch in diameter. The exposure during totality was for 40 seconds. Professor WINLOCK devised a plan by which the same motion that exposed the plate to the sun’s rays made an electrical connection with a chronograph; and thus the exact instant was recorded.
I just bought a clean print copy of this issue of Harper’s Weekly on eBay, but since I didn’t want to hold up my experiments by waiting for it, I also requested delivery of the print copy held by the Indiana University Libraries from the Auxiliary Library Facility, which I scanned as the basis for the animation below (each frame = 0.07 second; Figure 6 is disfigured by a slight tear, so I should probably redo this animation using my own copy at some point).
I’m not sure when the first sequence photographs of a lunar eclipse were taken, but a six-image sequence showing an eclipse of the moon on October 25, 1874—apparently photographed by Henry M. Parkhurst—was reproduced in the following day’s New York Herald:
Exhibit #9: Eadweard Muybridge’s Motion Picture of the Solar Eclipse of January 11, 1880.
In June 1878, Eadweard Muybridge famously captured his first photographic image sequence of a horse in motion. A year and a half later, he photographed the phases of the solar eclipse of January 11, 1880. Of the twenty-one frames he’s reported to have secured, the negatives for fifteen are now held by the Stanford Museum. As far as I can tell, the photographs themselves have never been published. However, Muybridge did publish a composite of seven of the images, captioned as “photographed for Hon. Leland Stanford.” The caption is a bit misleading, since the composite actually appears to be a photograph of drawings or paintings made by hand from the eclipse photographs. But that doesn’t seem to have lessened its twenty-first-century popularity: it was recently tapped as the cover art for Brian Catling’s fantasy novel, The Vorrh, and you can also buy art prints of it on Etsy.
Also included in the caption is a list of the times to which each of the seven pictures corresponds. The intervals are even more irregular than Hevelius’s at their worst, but here’s the obligatory animation (0.01 second = 2 minutes):
This isn’t one of the better sequences I’ve animated, although I might have better luck with the fifteen actual photographs at the Stanford Museum if these became available. Even so, it shows that Muybridge himself dabbled in eclipse sequence imagery, which gives us a further reason to factor older practices of recording eclipses into our understanding of the history of the motion picture in general.
As the nineteenth century drew to a close, efforts to photograph eclipses became increasingly automated, and by the early twentieth century regular motion picture cameras were being used successfully for this purpose. I won’t go further into these developments here, but for an overview, see Millicent Todd Bingham’s article “Solar Eclipse Photography: A Forerunner of the Motion Picture Camera” (1923), which picks up the thread with an experiment by David Todd in 1887, as well as Vitor Bonifácio‘s far more recent article “How the Movie Camera Failed to Become Part of the Standard Astronomical Observational Toolkit (1895-1914).”
The twenty-one eclipse sequences I’ve presented here as moving pictures are simply the ones I was able to identify and locate most easily. I’m sure there are others lurking out there, just waiting to be found and brought to life. But even with my current set of examples, there’s never a gap of more than fifty years separating any two adjacent sequences on the timeline; the longest gap stretches only from 1666 to 1715. And while some of my examples might conceivably have been prepared by people who had never seen or heard of eclipse sequence imagery before, I think it’s still safe to assume that an unbroken continuity of knowledge was kept alive throughout the period in question. Among other things, writers of the eighteenth and nineteenth centuries continued to cite and describe Hevelius’s Selenographia, marveling in particular at his maps of the surface of the moon. On those grounds, I think we’d be justified in speaking of a tradition or convention of using image sequences to represent eclipses, even if it manifested itself only sporadically in earlier centuries.
So let’s hear it for the eclipse movie! After all, how many other motion picture genres can boast a “filmography” spanning more than 375 years?