The eyes, it’s said, are the mirror of the soul. But they also serve as mirrors more literally: we can see things reflected in them. And when we’re looking at two eyes side by side, we can often see the same things reflected in both at once, from slightly different vantage points—nature’s own stereoviews! With that in mind, I’ve started looking around lately for historical photographs that show interesting reflections in their subjects’ eyes, with the idea of making the same kind of animations from them that I’ve made in the past from other, more obvious stereoscopic image pairs (see here, here, and here to bring yourself up to speed).
Before we go on, let’s take a quick look at a few photographs of reflections in single eyes, just so we’re clear about the sort of image I have in mind and the potential it carries. Here’s a composite of the nine coolest examples I was able to find on Wikimedia Commons (with thanks to the creators of these remarkable images):
These pictures do a fine job of illustrating how well eyes can reflect recognizable things, and how effectively these reflections can be photographed. I should point out that some of these reflections were apparently captured on purpose, while others seem to have been accidental.
I don’t know whether the reflection of the photographer’s own image in both eyes was an intentional piece of the composition, but the two reflections taken together can be used to produce a traditional stereoscopic effect or an animation like this:
In this case, we’re dealing with reflections in the eyes of two different people, but the same effect could naturally be achieved based on reflections in both of the eyes of a single person (which are presumably a lot more common in the photographic record). And this is the phenomenon I set out to try to find in old photographs, spurred on by these questions:
Did nineteenth-century photographers sometimes capture stereoscopic pairs of reflections in the eyes of their subjects? If so, what can we see in them? Do they offer us a look, however, distorted, at what the subjects were themselves looking at while being photographed?
The first old example of eye reflections that caught my attention appears on a cabinet card that turned up in an intact photograph album I found at an antique mall.
The original print has abysmal contrast and seems to have been used as an ink blotter at some point (I’ve presented an enhanced version on the right to bring out the detail a bit more effectively). The reverse of the card credits the image to “J. J. Rubottom, Photographer, Metamora, Ind.”; according to a document published by the Indiana Historical Society, “John J. Rubottom (1825–96) was a Bermuda-born photographer who operated a studio in Metamora, Indiana in the latter half of the nineteenth century.” If we zoom in on the subject’s eyes, we see a strikingly complex reflection repeated in both of them:
Here’s a morphed animation I created from the two reflections using this technique, yielding an illusion reminiscent of circling about a reflective Christmas tree globe:
So what are we seeing? I think the dark square visible near the center must represent the camera used to take the picture. To the right of it (from our perspective), we have what looks like a window complete with a rectangular grille; the camera stands in front of part of it. To the left of the camera we see more light coming, I suppose, through a window or door. Directly above the camera is another dark region that might just be a wall, but that could conceivably be the form of the photographer himself, perhaps under a hood. And so it is that we get a hazy and enigmatic glimpse into John J. Rubottom’s photographic studio, as reflected in the eyes of one of his customers.
That was my first experiment along these lines, and I liked the results well enough to begin looking around for other promising candidates. I soon learned that complex-looking patterns in eyes aren’t necessarily a good sign in and of themselves, since these can turn out on closer scrutiny to be artifacts of retouching. Instead, we need to be on the lookout for pairs of reflections that behave like the two images on a stereocard: the left image should show more of the scene to the left; the right image should show more of the scene to the right; and objects should appear further leftward within the right frame and further rightward within the left frame. Below are three photographic portraits I’ve found with eye reflections that fit those criteria: one tintype and two cabinet cards.
The left and middle examples center on black squares similar to the one in my first experiment, so I suspect these too may represent cameras. I’m not sure what’s going on in the last example, on the right, but there’s definitely a complex pattern of light in evidence there as well. High-resolution images these aren’t. But I’d argue that they’re all the more intriguing for their distortion and abstraction: they offer us a window on things we weren’t meant to see, glimpsed obscurely through the mists of time.
Photographs of eye reflections probably don’t predate intentional stereoviews, or at least not by very long, so they aren’t likely to push back the time depth of stereoscopy to any great extent. But we’re not necessarily limited to photographs, which leads me to my next question:
Can we find any paintings dating back before the invention of photography that display potentially stereoscopic pairs of eye reflections?
Why, yes we can! Here’s the famous “Portrait of a Young Woman” believed to have been painted by Petrus Christus between 1465 and 1470.
The reflection in the subject’s left eye—to the viewer’s right—is marred somewhat by craquelure, which makes it harder to interpret than it might otherwise be. Otherwise, though, the pair of reflections behaves remarkably like the photographic examples we considered above. An illuminated region appears further rightward in the subject’s right eye and further leftward in her left eye, just as we’d expect. Whereas the illuminated regions in the photographs were interrupted by big black squares which I’m interpreting as the silhouettes of cameras, it’s interrupted here by something narrower and more vertically extended. This thing—whatever it is—shows the horizontal displacement from side to side which we’d expect from an accurate rendering of its reflection in a pair of eyes, as well as an apparent vertical displacement (the subject’s left eye seems to be positioned a little higher than her right eye). Is it part of a window grille? Or could it be the silhouette of the painter against an open window or door? After all, its reflection in the subject’s left eye shows a vaguely headlike protrusion at the top (too high up to appear in her right eye, if I’m interpreting the correspondences between the reflections correctly—i.e., identifying the “blob” under the “head” with the “blob” at the top of the other reflection). Regardless of what the figure is, it seems to represent something in the original fifteenth-century setting. And we can create a smooth stereoscopic animation from it using image morphing software, just as we did with the photographic examples.
Postscript (October 11, 2015): For those of you with red-cyan glasses handy, here are anaglyphs of some of the paired eye reflections shown above as animations:
Credits for the montage of eye-reflection photographs shown at the top of this post, from left to right and then top to bottom: (1) Hazel eye with reflection of compact fluorescent lightbulb, by Ylem; (2) Eye with reflection of computer screen, by Veilleur; (3) Right female eye with reflection of glass door, by Bierenard; (4) “Castle using eye for wide angle lens,” by D L; (5) Reflection “looking up at the ceiling of the Bahá’í Centre of Learning, Hobart, Tasmania,” by Jalal Volker; (6) Cat’s eye with reflection of camera, by Stefano Mortellaro; (7) Eye of an Alexandrine parakeet with landscape reflection, by Tomfotograf; (8) Horse’s eye with reflection of photographer, by kallerna; and (9) Eye of an Australian magpie with reflection of photographer, by Toby Hudson. My composite is published under a Creative Commons Attribution/ShareAlike 3.0 Unported license, consistent with the licenses of the source images.