Within three years Land had produced his first artificial polarizing material. The secret was to use microscopic grains of needle-shaped crystals, which polarize light, and embed them in a lacquer. Land called his new product Polaroid glass. It “transmits,” he said, “almost all the useful light rays,” but it contains “a matrix of tiny crystals [that] combs out the tangled waves of light so that they are all vibrating on the same plane. The crystals are so small that you cannot see them. They are suspended in cellulose, all oriented in precisely the same direction.”
Land, now twenty and married, returned to Harvard, where he lasted for another two and a half years. He never took his degree. By this time he had met George Wheelwright III, a physics instructor of independent means six years Land’s senior, who suggested that they open their own laboratory. It became the Land-Wheelwright Laboratories, and in 1934 Land was issued his first patent, for manufacturing sheets of polarizing material. By 1937 the name of the company had been changed to Polaroid and Land had become chairman of the board, president, and director of research. Wheelwright ended up as vice-president. Wensberg does not tell us if this transition was an amiable one. Ironically, in view of a bitter lawsuit that Land and Polaroid eventually won in 1985 for patent infringement by Kodak on the Polaroid instant cameras, the first serious customer of the new company was Kodak, which wanted to use the material as filters for camera lenses.
The immediate future of the company had been secured by its alliance with the American Optical Company, which made sunglasses. Polaroid sunglasses became widely used, and still are. However, as Wensberg reports, by 1940 the market was saturated and Polaroid was reduced to selling its product for use in Wurlitzer jukeboxes, where it enhanced visual effects. No car manufacturer had ever made use of Polaroid in its automobile headlights. In fact, if the war had not come along the Polaroid Corporation might well have gone under. As it was, the company prospered during the war, making a variety of sun goggles for the military and working on the design of heat-seeking missiles.
None of this had anything to do with photography. But in December of 1943 Land had a second inspiration, now also a part of the Polaroid legend. On a rare vacation with his family in New Mexico he spent some time with his daughter Jennifer, then three years old, walking around Santa Fe and taking pictures, with Jennifer directing the picture taking. When they got back to where they were staying Jennifer asked her father about the pictures he had taken: “Why can’t I see them now?”
Our knowledge of the circumstances of this question, and what happened next, comes from Land. Since Wensberg quotes from what appears to be Land’s published account, one has the impression that he never was able to ask Land about it either. This is what Land wrote:
As I walked around that charming town [Santa Fe] I undertook the task of solving the puzzle [Jennifer] had set me. Within the hour, the camera, the film, and the physical chemistry became so clear to me that with a great sense of excitement I hurried over to the place where Donald Brown [Land’s patent attorney, who was conveniently in Santa Fe]…was staying, to describe to him in great detail a dry camera which would give a picture immediately after exposure.
To a Time interviewer Land put things slightly differently. The Time story reads, “He now claims jokingly that by the time he and Jennifer returned from their walk, he had solved all the problems ‘except for the ones that it has taken from 1943 to 1972 to solve.’ ”
In 1972 Land introduced the SX-70, the first color “instant” camera. The best simple explanation I know of how this camera works is given by Land in a talk he delivered in 1956 at the Franklin Institute. It has been reprinted in the Journal of the Franklin Institute under the somewhat ponderous title “From Imbibition to Exhibition.” ^1 In essence, the idea of one-step photography is to make a sandwich in which a negative and a positive sheet of paper encase an extremely thin layer—0.003 inches in the example Land gave in his 1956 lecture—of a chemical reagent used in developing photographs. The reagent is contained in what Land called a pod. When the pod is run through the camera the pressure on it breaks one end and the fluid runs out between the negative and positive papers.
It is not so difficult to imagine Land having thought up this part of the process. The rest of it is what took genius. In a conventional camera, the film is made up of silver halide, a compound of a halogen, a nonmetallic chemical element, and another element. When enough light falls on one of the grains of this material a speck of silver is produced. If chemical developer is applied, silver ions are induced to migrate through the photographic emulsion, where they deposit themselves on the exposed silver speck. This dark silver represents the exposure to light on the negative. The unexposed grains are then washed away with a solvent (“hypo”), and the result is the photographic negative we have all seen. If this sheet is placed in contact with a light-sensitive sheet—the positive paper on which the photograph is printed—and exposed to light, one soon has the picture. This is the normal two-step process that leads from exposing a negative in a camera to producing a photograph.
Land’s completely novel idea was to have the negative and the positive made at essentially the same instant. This can be done because the reagent in the pod can transport the unexposed silver ions from the negative across the very narrow gap to the positive. In practice the positive is doped with a catalyst that has a chemical affinity for silver. Thus instead of washing away the silver not attached to the negative—the usual procedure—this silver is attached to the positive, creating the positive image essentially simultaneously with the negative image. The dark spots on the negative correspond to the white spots on the positive since the silver that has attached itself to the negative cannot migrate across to the positive. The silver that the negative cannot retain becomes the basis for the image on the positive. In the original versions of Land’s camera, the exposed negative and positive had to be physically separated. There is a famous Life magazine picture of Land, the cover photograph of Wensberg’s book, showing Land looking at a picture of himself which he is revealing by separating the two sheets.
How much of this process Land envisaged that night in 1943 I do not know, but Nobel prizes have been given for less. (To cite two examples: in 1912 Nils Gustaf Dalén was given the Nobel Prize in Physics for a device that automatically regulated lights on lighthouses and buoys; and, more pertinently, in 1908 Gabriel Lippmann was given the Nobel Prize in Physics for his method of reproducing colors photographically.) In November 1948 the first Land camera went on sale for $95 at Jordan Marsh in Boston. It weighed five pounds, took sepia-colored pictures, and was an immediate success. Two years later Land introduced black-and-white film and it was a disaster. The film, once removed from the camera, continued to develop itself until the sharp images simply disappeared. To combat this the positive had to be chemically fixed—by hand painting it—which removed any illusion of the process being “one step.”
Land then made a decision of the kind that seemed characteristic of him. He decided to go for broke and develop an instant color process in which the resulting print would be fixed once and for all. (It remains unclear from Wensberg’s book whether the black-and-white problem was ever really solved.) This effort culminated in 1972 with the introduction of the SX-70, perhaps the most sophisticated invention involving the interaction between camera and film ever achieved. One of the innumerable problems that had to be solved was the battery. Each film pack contains its own battery that runs the automatic focus and the flash lamp. These batteries kept expiring before the film could be used—they still cannot be stored for long without going flat. Polaroid had to develop special batteries.
All of this is described very clearly in Wensberg’s book, as is the marketing of the camera. (Sir Laurence Olivier, chosen to sell the SX-70 on television, was stymied by the pronunciation of “SX,” until he decided it should be “Essex.”) Wensberg is also good on Land’s break with the Polaroid Corporation. He makes it clear that the same obsessiveness that was at work when he quit Harvard at the age of eighteen to go to New York to make a polarizer could be disastrous when his intuition turned out to be wrong.
Land insisted that the SX-70 have a complex viewfinder even though consumer surveys indicated that people didn’t like it. His attitude was that they would have to be educated to like it. He apparently never spoke again to the man who reported the market research to him. Even worse was the disaster of Polavision, a soundless system for taking home movies which was introduced at the time home video was taking hold. By 1977, when the system was introduced, Polaroid had spent millions on it, most of which was lost. By 1982 all of Land’s ties with the company were severed.
Land’s interest in color vision, not discussed at all in Wensberg’s book, goes back at least to the 1960s.2 Like many scientific ideas of great importance Land’s theory starts from a phenomenon we all recognize but do not attach much importance to until someone like Land calls our attention to how remarkable it really is. This is what Land and his collaborators call “color constancy.” It is a fact of common experience that colors retain something like the same appearance, regardless of how the colored object is illuminated. A blue object, for example, looks recognizably blue when viewed from sunrise to sunset. This ability of the brain to maintain the near constancy of colors, even as the illumination is varied, would give a species that had it an evolutionary advantage. An edible red berry, for example, would look like an edible red berry whether encountered at 8 AM or at 5 PM.
Land devised an experiment to exhibit this remarkable fact quantitatively. (He showed a version of it to us in Minnesota in 1982.) It involved boards Land called Mondrians because they resemble the work of that artist. Paper rectangles of different sizes and different colors are pasted on each board. For the sake of illustration we can imagine two adjacent rectangles, one red and one white. According to the scientific theory of color, each color is associated with a specific wavelength of light. When we say that the rectangle or any other object is red, we mean that it will absorb all the wavelengths of light except red ones and reflect back to our eyes only the red wavelength. Hence the object looks red to us. A white object, on the other hand, reflects back all the wavelengths impinging on it, so no color is singled out and the object looks white.
In Land’s experiment the rectangles are illuminated by three colored spotlights—say red, green, and blue. When this is done, the white rectangle reflects all three colors from the spotlights about equally, while the red rectangle absorbs much of the blue and green light and reflects the red. This is what we would expect.
The surprise comes when one begins to change the intensity of the light coming from the different colored spotlights. One can adjust the intensity of the spotlights so that the reflected light coming from the red rectangle has just the mixture of intensities that formerly were reflected from the white rectangle. Naively speaking, one could say the reflected light is now indistinguishable from white light. One might, therefore, be tempted to think that the red rectangle would now look white, just as one might be tempted to think that a red berry viewed at noon would have a different color from that of a red berry viewed at 9 AM. But this is not what happens, either for the berries or, more dramatically, for the Mondrians, where the light intensities can be precisely adjusted. In both cases the red object continues to look red.
This presents a difficult problem for any theory of color. We know that on the outer layer of the retina there are visual cells called “cones.” (There are also “rods,” which have a part in night vision.) Fundamentally there are three types of cones, each one sensitive to wavelengths appropriate to a given primary color. So, for example, approximately a third of the cones respond selectively to red light. (In reality there is some overlap in the sensitivity of the different cones, otherwise we could not see a color like orange, but that is a nuance that need not concern us here.) It would be tempting to say that we see red when our red cones are stimulated by red light and that the more they are stimulated the more red we see. But then how do we explain Land’s experiment in which a red square still looks red even though it reflects, under suitable illumination, the same mixture of light intensities that made a white square look white? The clear implication is that there must be more to color vision than such a naive processing of intensities.
That this is the case was made dramatically clear in a recent article in these pages by Oliver Sacks and Robert Wasserman (The New York Review, November 19, 1987). They describe their study of a painter they call Jonathan I., who lost his sense of color vision when he was injured in a car accident. Elaborate tests showed that his cones were intact. He did not have colorblindness in the usual sense of having defective cones. He was able therefore to register the intensities of different colors when they were presented to him, but he couldn’t see the colors. Still, the images he saw looked different to him when the intensities of the spotlights were changed, whereas a person with normal color vision would have noticed no change.
To see color the brain must therefore be able to do some kind of analysis of visual data beyond simply analyzing the intensities. Land, in collaboration with John J. McCann, has suggested such a model, which they call the “retinex theory.” Basically the idea is that by scanning the entire visual area—not, say, one rectangle but the entire Mondrian—the data processing system responsible for color vision produces three so-called lightnesses, one for short wavelengths, one for medium wavelengths, and one for long wavelengths. It is the combined effect of these lightnesses—a single point in color space, one could say—that determines what color we actually see. A reader who wants to know more about how, according to Land and McCann, the brain calculates the combined effect should consult the references I have given. I am not enough of an expert on the theory of color vision to describe just how Land’s model is to be compared with other models for color constancy. If Land ever writes his autobiography perhaps he will tell us. As I have mentioned, none of this is to be found in Mr. Wensberg’s book. Someday a truly serious biography of Land will be written in which Land’s Polaroid will be a footnote. Or perhaps Land will surprise us all by writing his own book.
Three accessible references to this matter are: E.H. Land and J.J. McCann, Journal of the Optical Society of America, Vol. 61, No. 1 (January 1971), pp. 1–11; E.H. Land, Scientific American 237, No. 6 (1977), pp. 108–128; and J.J. McCann, "Retinex Theory and Colour Constancy," in The Oxford Companion to the Mind, Richard L. Gregory, ed. (Oxford University Press, 1987), pp. 684–685.↩
Three accessible references to this matter are: E.H. Land and J.J. McCann, Journal of the Optical Society of America, Vol. 61, No. 1 (January 1971), pp. 1–11; E.H. Land, Scientific American 237, No. 6 (1977), pp. 108–128; and J.J. McCann, “Retinex Theory and Colour Constancy,” in The Oxford Companion to the Mind, Richard L. Gregory, ed. (Oxford University Press, 1987), pp. 684–685.↩