Accurate and minute measurement seems to the non-scientific imagination, a less lofty and dignified work than looking for something new. But nearly all the grandest discoveries of science have been but the rewards of accurate measurement and patient long-continued labour in the minute sifting of numerical results.

Baron William Thomson Kelvin^​*​

Introduction

History is fond of scientists whose contributions withstand the test of time and whose public personas elevate them above the field. But for every great moment of insight, revolutionary invention, or paradigm shift, there is a cadre of dedicated scientists asking fundamental questions about our understanding, moving the marker along the timeline of progress. It is attractive to focus on those who, by their genius or simply by being in the right place and right time, find themselves enshrined in the great halls of science. We live today in an age of digital photography and image processing. We should study and appreciate those who made it possible. However, it would be a mistake to overlook the contributions of our fathers and grandfathers and the process of invention that advanced analog photography. We are all connected by the thread of history.

In the previous article in this series, Orthochromatic Photography, Part 1, we examined how scientists and photographers expanded the sensitivity of plates to the full visual spectrum. In this article, we’ll discuss the innovative scientific techniques used to measure color and accurately characterize those new orthochromatic plates and screens.

There were no standard techniques or tools for measuring light, the sensitivity of plates, or the spectral characteristics of screens (filters) and other materials in the late 1800s^​1​. The scientific community was reluctant to consider the measurement of light a true science, worthy of formal institutional support. As a result, the scientists who developed orthochromatic photography in the 1800s—Vogel, Eder, Abney, Shepherd, Pierre Bouger, Johann Lambert, Benjamin Thompson, among others—were a motley group of chemists, photographers, and citizen scientists working independently to meet a common need for light measurement and expanded spectral sensitivity. As Johnston states, the period was “characterized by a lack of social cohesion and interaction between investigators, a collection of practices developed that came to value the brightness of light as a quantity. Their motivations and methods were particular, seldom involving social interactions tied to organized applications of light measurement or sharing of research results by like-minded individuals. Indeed, an investigator during this period who became aware of another’s work was as likely to discount it as to build upon it” (Johnston, 2001, p. 12).^​1​

There are three components necessary for characterizing any imaging system: 1) the light source, 2) the object, and 3) the detector. Light sources are characterized using photometric measurements. Objects are characterized using spectral reflectance or transmittance measurements. Detectors are characterized by a measure of spectral sensitivity, including that of plates or human vision. At the time orthochromatic photography was being developed, photoelectric light measuring techniques were in their infancy and rarely used in scientific measurement. Photometry (of a light or of an object) was a visual science. Absolute measurement relied on the judgement of human observers and, therefore, the measurement of lights and objects were impacted by the peculiarities of the human visual system.

The story of how orthochromatic photography was influenced by the development of absolute measurements is the story of those scientists who pioneered methods for accurate measurements of light and color; those who helped the world understand how a careful study of the interaction of light and matter could bring color to the world.

Early Sensitometry and Measuring Exposure

To take a properly exposed photograph on film using modern technology, you simply point the camera at your subject, look in the viewfinder and adjust the shutter speed or aperture to the optimal setting indicated on the exposure meter for the given film speed. None of these tools existed for photographers in the mid-1800s. There was no standard measurement of emulsion speed (what we know today as the “ISO” speed) or electronic sensors to measure. The science of sensitometry, the study of the relationship between photographic density and exposure, grew out the need for photographers to determine the sensitivity of their plates to different exposures.

Speed measurements date back to 1848 with Claudet and his sensitometer for daguerreotypes.^​2​ There were several later types of sensitometers developed in the mid-1800s. All required an exposure with a constant light source and constant development. An early sensitometer was patented in 1866 by Louis Bing and announced in 1876 by Colonel Wortley.^​3​ Taylor’s Sensitometer was first used in 1869.^​4​ It consisted of a box with shallow wide tubes. At the top of each tube was a thin material with different numbers of holes punched in it. The plate was placed at the bottom of the box. The intensity of light passing through each cylinder was proportional to the number of holes. The photographer would mark the cylinder with the smallest number of holes that produced a significant density on the plate.

Hermann Vogel introduced a sensitometer similar to Taylor’s which used a metal plate with holes placed over a wooden block in which large holes were drilled to channel the light.^​4​ Vogel described his instrument as follows:

This instrument is a double one, like the stereo box. The construction of the two sides is exactly the same. In the front part you see a plate made from thin brass, divided in twenty-four fields. In the first field is drilled one small hole three-fourths of a millimetre in diameter; in the second two holes, and so on up to twenty-four…under each of these above mentioned fields are drilled tubes, each tube exactly corresponding with each one field with holes.”^​5​

The luminance at the back of each tube was proportional to the number of holes. Plates were placed in the back of the sensitometer and exposed to light using a diffuse source. Having two sections allowed for the direct comparison of different plates or screens.

Mucklow and Spurge also released a sensitometer similar to Taylor’s, but with one hole of varying sizes rather than multiple holes (amplitude modulation rather than frequency modulation).^​4​

Sensitometers were a variation on actinometers, photometric devices that measured light intensity from the activity of photochemical reactions. The Woodbury Photometer was a small, portable device that might have been one of the first portable exposure meters.^​6​ Sensitive paper was inserted into the device and appeared through a hole in the center. Surrounding the hole were six sectors with different tints.

Abney described the workings of the Woodbury Photometer:

Here we have a piece of bromide paper exposed to light for a minute, then read off against one of these tinted circles, according as to whichever tint it agrees with; you know what is the intensity of the light, and, therefore what to give to a plate. A simple rule to remember is this, that if you use a bromide plate, only use a bromide paper for securing the tint; if you are using a chloride plate, use a chloride paper.^​7​

When measuring light intensity, photographers had to be aware of the spectral sensitivity of the particular silver salt used in the plate emulsion. There was no process standardization, so experimental means were required to determine the best operating procedure for each plate. Exposure calculations based on sensitometry were less accurate on days with high UV and some exposure compensation was required.

Edmund Bequerel created the first electrochemical photometer in 1841. “This device consisted of a wooden box divided in two compartments containing acidified water. In each compartment was immersed a silver plate covered with a thin silver chloride layer, prepared as a photographic plate. Each plate was connected to a very sensitive galvanometer. As in the previous device, a hatch gave the possibility of lighting one plate, leaving the second in its dark compartment”.^​8​ Bequerel’s device was used to test different salts (silver bromide, silver chloride, silver iodide) and determine which produced a stronger reaction. This was essential to the development of photographic plate materials from those same silver salts.

Władysław Małachowski, a Polish photographic scientist known professionally as Leon Warnerke, also created an actinometer exposure meter based on the movement of gasses from the decomposition of a sensitive liquid. A bottle of photosensitive liquid was connected to a horizontal glass tube containing a drop of mercury. As light hit the bottle, gasses formed in the bottle and propelled the mercury down the tube. The exposure calculation was based on the distance the mercury traveled down the tube.

Scientists were constantly battling error, especially in the exposures calculated using actinometers. William Abney wrote in critique of Warnerke’s actinometer, “Errors consequently are unavoidable in the use of this kind of photometer according to the different absorption powers of the liquid…You cannot absolutely say it will act two days alike, because in some days I find there is three hundred times as much blue as on an ordinary day in the present weather.” Abney continued, “For practical purposes Mr. Warnerke’s photometer is well adapted, but for exact registration my objection to it will hold good. It will answer admirably to time the exposure for negatives, but not strictly scientific measurement.”^​6​ Abney is saying that Warnerke’s photometer gives a good approximate exposure but is not precise enough for scientific measurement.

Actinometers were useful for quantifying the intensity of light and analyzing the reaction of certain sensitive chemicals (variable light source). Sensitometers were not much different in principle. The density on a negative resulting from a chemical reaction of the silver salts with light was related to exposure (luminance x time). Unlike the Woodbury Photometer, the light source in sensitometers was held constant and the chemical response was unknown.

Another early sensitometer from Wortley created a step wedge using “layers of paper, from one to sixteen in number” placed in a printing frame and in contact with the plate to be tested. The system was illuminated by a gas jet flame directed to the plate using a conical luminaire.^​3​ Using this type of design, which Wall referred to as a “step sensitometer,” Warnerke created one of the most popular sensitometers of the late 1800s and early 1900s.  

While sensitometers were used casually during the period in which wet plates were popular, the development of dry plates with higher sensitivities hastened the need for accurate measures of plate sensitivity and the need to study the effect of light and different developers on density production.

Warnerke created several iterations of sensitometers. His first design used a mold created from 25 squares of paper with thicknesses of 1 to 25 stacked sheets.^​3​ He printed a Woodbury Type gelatin on glass to create a plate of graduated tints. Numbers were printed on each square and plates were exposed in contact with the sensitometer plate. The plate sensitivity (speed) was the largest number square for which density appeared on the plate (larger numbers were more opaque, passing less light). Because this was a relative measurement, two plates could be compared side-by-side to determine their relative sensitivities.

Warnerke improved on his first sensitometer design and created the “Standard” sensitometer. He wrote, “It consists of a glass quarter plate having Woodburytypic impressions forming 25 squares, steadily increasing in opacity.”^​3​ The stacked paper was cemented together and molded in Spence metal, a metal-like compound “composed of metallic sulphides and sulphur.”^​9​ Wayne Homren wrote about Spence’s metal in the context of coin manufacturing: “It was not a metal! It was, in fact, a mixture of metal sulphides dissolved in sulphur. It had the useful properties of a low melting point (a little over 300 degrees Fahrenheit), expanding on cooling, and high resistance to acids (including aqua regia). It was primarily used for sealing joints in iron pipes and I haven’t found any references to its use for making medals.”^​10​

Spence metal was used because Woodburytype printing required a very flat surface. Pritchard wrote, “The difficulty of getting large metal surfaces truly flat is almost insurmountable, and if printing surfaces are not true in Woodburytype printing, an unevenness in inking results, and uniform pictures are unobtainable. But the easily-molten Spence metal will supply any number of flat surfaces without difficulty, if care is only taken to cast it upon a true level.”^​11​ A gelatin ink was used to print the step wedge and a preservative added to prevent decomposition.

Warnerke’s Standard sensitometer also required a standard light source. A common standard light source at the time was the “standard candle,” a candle with a specially defined wax formulation and wick that burned with a known luminance. According to Wall, a standard candle had to be exactly 1 m away from the sensitive surface and burn for 5 minutes before experimental use. The wick should be kept to 45 mm and trimmed if needed.^​2​ However, Standard candles required expert operation and a standard setup. The light from a standard candle was dependent on surrounding conditions. Intensity varied according to oxygen level, type of wick and the tightness of the wick’s weave. The user could expect a 10% variation in light, depending on temperature. Plus, candles were not acknowledged as “standard” anywhere, except in England, and consequently, were hard to find.

Warnerke understood the difficulties of Standard candles and instead used the light emitted from a plate uniformly coated with a phosphorescent paint as the source to his sensitometer.  He wrote:

A large quantity of phosphorescent calcium suphide is obtained in order to secure uniformity of the first product. This is mixed with melted (solid) paraffin and poured on to a previously warmed glass plate; while the paraffin is still in liquid condition it is shaken to distribute uniformly the powder which might sink to the bottom…The phosphorescent coating is covered with another glass, and the edges cemented.^​3​

The phosphorescent plate was excited (made to phosphoresce) by burning a 1-inch magnesium ribbon close to the plate. Warnerke experimented to determine the proper timing. The phosphorescent emission (the light emitted from the paint after excitation by the burning magnesium ribbon) started high then decreased quickly in the first minute until it stabilized. Therefore, the plate rested one minute after excitation by the burning magnesium, and then a one-minute exposure was made by placing the phosphorescent plate some distance from the photographic plate using a specially made frame.

Abney provided another account of the Warnerke sensitometer’s use. “First, a phosphorescent tablet is exposed to magnesium light, and then allowed to rest a minute, placed in contact with the sensitometer plate, which is in front of the plate to be tried. The exposure lasts for half a minute, and the plate is then developed. The last distinct number seen upon the plate by reflected light, before fixing, is read off.”^​7​ The numbers on the Warnerke device are akin to the exposure value units of modern photography, each number indicating a level of exposure or sensitivity. Though this system did not provide a specific “shutter speed” and “aperture,” it helped photographers determine the relative exposures required for two plates of different sensitivities. So, if they knew one plate required a 30 second exposure, then they could calculate the exposure for other plates based on that 30 second reference.

Warnerke’s sensitometer was popular among photographers due to its ease of use. However, there were arguments against it: the creation of the phosphorescent paint was not repeatable; the effect of emission decreased at a rate proportional to excitation time; a uniform coating was difficult to achieve; the temperature had to be constant, etc. Stolze also cited concerns about the longevity of the phosphorescent plate’s emission over repeated uses and reiterated concerns that preparation of the phosphorescent paint was not repeatable across operators or instrument batches.^​12​ The lampblack pigment in the sensitometer plate was also subject to changes in temperature and may otherwise change over time. In general, a sensitometer’s standard light had to be less dependent on variability in the manufacturing process.  Stolze wrote, “we must rather seek for a principle which makes the instrument used in testing for light as independent as possible of difficult-to-control chemical and physical properties of the substances employed, and which lays the principal stress upon linear proportions which always and everywhere may be determined with sufficient accuracy.”^​12​ Though critical, Stolze offered up his own suggestion for a standard light using an arc lamp, but it does not seem to have taken hold.

William Abney strongly promoted the adoption of sensitometers by both amateur and professional photographers. “To my mind, no photographer, be he amateur or professional, should be without one, or one similar to it.”^​7​ Abney was an advocate for scientific rigor in photography. He credited Warnerke as the originator of the “sensitometer,” perhaps referring to its wide commercial availability. However, Abney found this sensitometer’s convenient use of phosphorescent paint to be its greatest weakness. He found that the paint’s phosphorescent emission did not correspond with the spectral sensitivity of the silver salts commonly used in plate emulsions. Warnerke’s phosphorescent paint emitted primarily blueish-indigo light. Of the silver salts Abney tested—silver iodide, silver chloride, silver bromide, and silver bromo-iodide—only silver bromide was significantly sensitive to the indigo of the phosphorescent emission. The others required much greater exposures to show any response. This gave way to incorrect exposure values due to the mismatch between spectral sensitivity and phosphorescent emission.^​7​ The plot below shows the emission spectrum of the phosphorescent paint (top) compared to the spectral sensitivity of the four silver salts.

William de Wiveleslie Abney was perhaps the most important figure in the development of light and color measurement in the late 1800s. He was born on July 24, 1843 in Derbey England^​13​ and was among the breed of citizen scientist that Johnston referred to as “enthusiastic amateurs”. Johnston wrote, “By championing an unpopular subject using private funds, [enthusiastic amateurs] were able both to increase its exposure to particular communities and to nurture its development along individualistic lines” (Johnston, 2001, pp. 72-73).^​1​ Light measurement in the 1800s was not a discipline recognized by university physics and chemistry departments. Abney received his engineering and scientific training at the Royal Military Academy and was encouraged to build on his interest in photography. He was an instructor in chemistry and photography at the Chatham School of Military Engineering and conducted his experimental work in a lab at the South Kensington Museum. He was a prolific author of scientific papers, publishing articles on subjects including photography, sensitometry, optics, and photometry.^​14​ According to Johnston, “Abney was central in laying the foundations for photographic photometry and unique in having a broad interest in light measurement as well as an unparalleled desire to understand the scientific basis of photography” (Johnston, 2001, p. 73).^​1​

Abney was a staunch advocate for the advancement of photographic research, serving as President of the Royal Photographic Society four times in the period between 1892 and 1905 (Johnston, 2001, p. 73).^​1​ He helped push the Society to a direction of scientific research and promoted the development of technology accessible to the average photographer. The development of sensitometry was driven by the lack of information provided by photographic plate manufacturers to their customers. Manufacturers did not often provide exposure and development for their plates, which meant it was up to photographers to figure out how to use them. Abney sought to remove the guessing game of plate performance: “The speed of a plate is usually understood to be measured by the minimum exposure that will secure a good printing negative, and we have extant numbers of expressions which are supposed to indicate such exposure,” he wrote.^​15​ He recommended an easy method to determine the speed of plates for both professional and amateur photographers.

The development of sensitometry required an understanding of the fundamental relationship between exposure and the density production. Exposure (E) is the light intensity (I) multiplied by time (T). Intensity at the plate can be varied by changing the aperture or by changing the power of the light source. According to the equation, E = I x T, exposure should be the same if you double the intensity and half the time, or vice versa. This is called the law of “reciprocity.” However, Abney knew that the law of reciprocity often failed: “Now a pretty considerable number of experiments…proved that, except in the case of very rapid plates, this assumption [of reciprocity] is not valid, and that even with very rapid plates, when the intensity of light is beyond that of ordinary daylight, the supposition cannot be maintained.”^​15​ Within normal daylight conditions there was no reciprocity law failure, but with artificial illumination, which was much less intense than daylight, the effect was omnipresent. Slower, less sensitive plates were even more susceptible to this problem. This made it difficult to estimate exposures and characterize plates.

Another assumption in reciprocity law was that the rate density should change at the same rate when doubling intensity as when doubling time. This assumption was also false. At some point, when illumination was either high or low, changes in illumination did not result in proportional changes in density. The effect was less noticeable when illumination was fixed, and exposure times were varied. As an analogy, think about pouring water on a sponge. If you pour water on the sponge for one second, then you would expect that doubling the amount of water each time would result in the sponge absorbing double the amount of water. At some point you have so much water being poured on the sponge in that one second that it will splash off and not be absorbed.  On the other hand, if you pour water onto the sponge at a constant rate, and just increase the time of the pour, then you give the sponge more time to absorb the water. The doubling of water absorbed will only subside when the sponge has become saturated.

Abney also found that operating temperature had an impact on plate density and concluded that both operating temperature and the minimum intensity of light should be specified for certain plates. The only problem was that there was no reliable method for measuring the intensity of light. Abney stated “that the measure of intensity of light which, with a given exposure, will just not give a deposit, is the most useful test: for a great deal of opacity which may show gradation in the negative will be practically useless” due to the differences in dynamic range.”^​15​

Abney and Festing

Abney worked extensively to develop techniques for the measurement of light and color. He and his colleague, Major-General Edward Robert Festing, were awarded the Bakerian Medal in 1886 by the Royal Society which came with the honor of giving the Bakerian Lecture that year.  Festing, a few years older than Abney, also studied at the Royal Military Academy and became a member of the Royal Engineers.^​16​ Festing was skilled in mathematics and electrical science, but maintained a diverse interest (much like Abney). Abney and Festing’s Bakerian Lecture described the techniques they developed for “measuring the relative illuminating intensities of different parts of the spectrum”, which they termed “colour photometry”.^​17​ Photometric measurements were based on visual observation since photoelectric measurement systems were not yet available. Abney and Festing used a system in which observers would “compare the visual intensity of one ray with another of different color.” Basically, the user would see two patches of light reflecting off a white card, one colored and one white, and then adjust the intensity of the white until it had the same brightness as the colored patch. The device they constructed to make these measurements was a type of “visual photometer,” a photometer being a device that measures how intense the light of an object looks to a human observer (as opposed to a radiometer, which measure the absolute amount of light in physical units). The illustration below shows how an observer might have seen a green sample patch next to the reference white in Abney and Festing’s photometer.

The sample color always remains static during measurements, while the luminance of the reference white is varied until it shares the same luminance as the sample color. Different methods could be used to control the luminance of the reference white. Abney and Festing were inspired by the Rumford Photometer. According to Johnston,

In 1794, [Benjamin, a.k.a. Count Rumford] Thompson devised a visual photometer for measuring light intensity…His photometer consisted of a sheet of white paper and a cylinder of wood fixed vertically a few inches from it. The two light sources to be compared were placed on moveable stands some 6 to 8 feet from the paper and from each other. The observer compared the shadows of the cylinder cast by the two lights, and moved one or the other light further away until the densities of the shadows appeared to be exactly equal.”^​1​

(Johnston, 2001, p. 15)

They referred to Thompson’s method as the “Rumford System of Photometry” and found it convenient. In Abney and Festing’s first visual photometer (they later improved upon the design), the reference light luminance was varied by moving the light closer or farther from the reflecting screen. Moving it toward the screen increased the luminance and moving it farther from the screen decreased the luminance. The amount in which the luminance increased or decreased varied proportional to the inverse of the square of the distance (the inverse square law). In their device, shown below, light from one source was focused on the entrance slit to a collimator (S, in the figure). The light then passed through two adjacent prisms (P), dividing the spectrum. The spectrum from the prisms was focused by a “camera” on the ground glass at the viewing side of the camera (B). The dark slide was replaced with a card with a narrow slit that could be slid back and forth, aligning with different parts of the spectrum formed on the ground glass. Another lens (L) then collected the light from the sliding slit and focused it on a white screen (D) four feet away. The reference light (T) was at an acute angle from the camera and illuminated the same white card. A half-inch wood rod (E) was placed in front of the screen to cast a shadow of the reference source and of the spectrum. 

A simplified illustration of the device is shown below. You can see how the slit in front of the spectrum passes green light and illuminates half of the viewing field, with the rod casting a shadow on the adjacent half. The reference source illuminates the viewing field next to the green patch, with the rod performing the same function to prevent white light from affecting the green half.

Abney and Festing’s goal was to measure the human visual system’s spectral sensitivity curve (the luminance of light at each wavelength) and the effect of different factors on that curve. The luminance of the reference candle was calculated from the candle’s distance to the screen using the inverse square law.

Observers used two techniques to determine the specific luminance of each spectral point. Abney and Festing guessed that our sensitivity to color peaked around the green and tapered off in the red and blue. Two methods were used by observers to calculate sensitivity at different spectral positions. First, the candle position was held constant, and the user then adjusted the sliding slit until a color was found matching the luminance of the reference source at its given position. The authors promoted an oscillation method, moving the slit back and forth until the observer was sure the luminance of the sample at the selected wavelength matched the luminance of the reference candle. The slit was then moved farther from the screen and a second wavelength found that matched the luminance of the reference candle.

The second measurement method held the slit constant. The candle position was adjusted coarsely, then oscillated back and forth until the observer was sure of the matching luminance. This method was ideal for positions close to the maximum of the curve. Optical adjustments were made far into the violet due to the low light level.

The process of making precise visual judgments was no simple task. The human visual system has many tricks that help us survive in a world of widely varying stimuli, but they tend to create problems when trying to make precise experimental judgments. Perhaps the most debilitating effect is our tendency to second-guess our own perceptions, rather than go with our gut-instinct. Observers in these vision experiments were asked to make judgments as quickly as possible, without much consideration for thought, as Abney wrote below:

By gradually diminishing the range of the “too open to “too close” apertures we arrive at the aperture where the two colours appear equally bright. The two patches will cease to wink at the operator, if we may use such an unscientific expression, when equality in brightness is established. This operation of equalizing luminosities must be carried out quickly and without concentrated thought, for if an observer stops to think, a fancied equality of brightness may exist, which other properly carried out observations will show to be inexact.^​18​

The spectrum scale was calibrated using the known emission lines of various salts, such as magnesium and lithium. Lithium has emission lines in the red and blue, magnesium in green, and the sodium in orange and yellow. Calcium chloride was also used to fix the Fraunhofer H line at 396 nm. A magnesium wire, coated with lithium chloride was burned in front of the entrance slit to the photometer. The focused emission lines were then projected on the camera’s ground glass.

The standard candle turned out to an unsuitable source for the spectrometer. As we discussed before, there were many disadvantages to using the standard candle in sensitometers: lack of luminance, availability, luminance control, etc. These disadvantages are also true for its use in photometers. Abney and Festing also found the new Edison-style tungsten lamp were too weak. The reference source had to be bright, steady, common, of constant temperature, and of “constant total illuminating value.” The electric arc light was the artificial light of choice for Abney and Festing.

The carbon-arc lamp was a common electric light source since the 1870s due to its quality and whiteness of illumination. The lamp contained of two carbon rods with pointed tops set at some close distance from one another. Electrodes were attached the carbon rods. As current passed from one rod to another, ionized carbon particles traveled across the gap and gave off an intense light.^​19​ The figure below, from Sparavigna’s paper referencing earlier articles, show what is referred to as the “crater” of the lamp, the bright ball of light between the rods’ points.

The term “crater” comes from the depression formed in the top one carbon rod as carbon particles migrate to the other rod, forming a depression. The photograph below, from Jacob Abbott’s 1871 book, Light, shows the crater on the top rod and the carbon particles built up on the bottom rod.^​20​

The luminance of an arc lamp decreases as the tips move farther away from one another. However, the distance between the tips increases as the carbon burns away. Maintaining a constant luminance requires the constant movement of one of the charcoal tips to keep the two tips at a constant distance. Abbott reported an apparatus that regulated the distance between carbon elements in the arc lamp using a weight and an electromagnet to maintain a constant luminance.^​19​ However, it’s unclear whether Abney and Festing used this type of device in their electric arc illumination.

Performance of the arc lamp is also dependent on the frequency of the electric current. AC current produced by the electric motors had a constant frequency. A slow motor frequency would result a series of flashes, but if the frequency was fast enough then the flashes are not noticeable. You could also run an arc lamp from a DC current using batteries, but Abney and Festing used an electric motor to power their lamp, which ran at 467 W (11 A and 42.5 V).

Like their first design, a condenser was used to collimate the light from the arc lamp and form an image of the crater on the entrance slit. White cardboard was initially used for viewing the color patches but this was abandoned due to quality control issues. A board coated with zinc oxide in gelatin, a very pure and consistent white, was used instead. For the reference lamp, they needed symmetric and constant illumination. The gas flame was not used due to the need for a hose to supply the gas and the possibility of interrupting the gas flow. An ordinary candle or Siemens’s standard unit lamp burning amyl acetate (9 candelas) was the best available reference source.

Luminance measurements were made of spectral bands across the range of the visible spectrum. Abney and Festing found their results were similar, citing only a 2% observation error based on several replicated measurements. Feeling confident in their abilities to make measurements, they considered themselves “normal” observers. The curve area of each spectral luminance measurement was normalized to 100.

It was assumed that the observations were independent of the arc lamp’s spectral power. Abney and Festing tested the effect of the reference light color on their observations by filtering the reference light with various colored screens. They found no effect of the reference light’s color on the observations. They also tested the effect of luminance on the observations by varying the width of the entrance slit, but they could find no significant effect. Out of this research came an early measurement of the luminance efficiency function, which was then known as the “comparative luminosity” curve, reproduced below in an 1898 article by Sanger Shepherd.^​21​ This function was later revised in the 1920s and became known as the “V-lambda” function.

Perhaps the most lasting contribution of the Bakerian Lecture article was a validation of the law of additivity for light. Abney and Festing hypothesized that the luminance of component colors in a mixture should match the luminance of the mixture itself. R + G + B = (R + G + G). They used this principle to test the accuracy of their measurements. Rather than a single slit at the end of the prism camera, they used a card with three slits, positioned roughly over red, green, and blue regions of the spectrum. The light passing through the three slits was recombined by a lens into white and the luminance measured. The luminance of the three individual bands was also measured by closing all but one of the slits. If the law of additivity held, or conversely, if the measurement system was accurate and repeatable, the luminance of the three components should equal the luminance of the combined light.

They wrote: “It has been assumed, but, as far as we know, never been experimentally proved, that the impression on the eyes of a mixed light is equal to the sum of the impressions of each of the components of the light.”^​17​ The data supported the hypothesis that the luminance of the red, green, and blue channels added up to the luminance of the combined light. “We therefore feel satisfied that it is true that the impression caused by mixed light is equal to the sum of the impressions of its components, also that the method of measurement adopted is trustworthy.” Abney and Festing argued that demonstrating additivity validated their instrument, although a cynic might argue that they simply demonstrated how their measurement method was successful in proving the law to be true. Ironically, this principle of additivity of luminance became known in the annals of history as “Abney’s Law,” despite the team effort.

In 1853, the mathematician Hermann Grassmann, an advocate for Helmholtz’s three-color vision theory, published an article in which he proposed four laws for the mixture of colors.  Abney’s Law is a restatement of Grassmann’s Fourth Law, “dass die gesammte Lichtintensität der Mischung die Summe sey aus den Intersensitäten der gemischten Lichter,” which translates to “that the total light intensity of the mixture is the sum of the intensities of the mixed lights.”^​22​ Abney and Festing’s apparatus is also a variation of Grassmann’s apparatus in which spectral light and white light fell on two white panels, each of which could be rotated to change the relative luminance reaching the viewer.

David McAdam later shows that Abney’s Law does not apply to the sensation of brightness, but only to luminance (brightness being the perception of intensity while luminance being an absolute measure of perceived intensity). Cohen also felt another figure, Ogden Rood, deserved some credit for observing the additivity principle in 1878 using Maxwell Disks (Cohen, 2001, pp. 155-156).^​23​

Abney and Festing were intent on using the sun as the spectral source, but the constant variation in daylight made this impossible alongside a static reference lamp (or candle). The solution was to modify their apparatus such that the sun was both spectral source and reference. The reference source was just the re-directed reflection off the surface of the first prism, as shown in the diagram below. A mirror directed the reference beam to the screen where it formed the reference white patch.

Later In 1896, Abney and Festing modified their apparatus to use of 10 candlepower incandescent lamp. The intensity of the reference beam was modulated by rotating sectors on a disk spinning at  40 RPS (2400 RPM)^​17​ shown below in an illustration from Abney’s book, Colour Measurement and Mixture (Abney W. , 1891, p. 46).^​24​ The sectors could be opened and closed by means of the lever (C, in the illustration) while the disk was spinning. Opening the sectors increased the luminance of the reference patch and closing the sectors decreased the luminance.

Rotating sectors were preferred over moving the light source because the light could be kept in a fixed position. Filters were also common light modifiers, but they often had an unavoidable tint modifying the color of the light. Rotating sectors solved this problem as well.

Rotating sectors modify light intensity according to Talbot’s Law, which says that luminance through a rotating sectored disk is proportional to the angle covered by the sectors (e.g. 90 degrees is half the light of 180 degrees), described by the equation  where d is the combined angle of the openings. According to Johnston, the pioneering scientist Henry Fox Talbot, “postulated that the apparent brightness should be proportional to the fraction of the cut-out diameter of the wheel. Thus, to avoid one of the problems he saw with photometry—that of obtaining a quantifiable reference intensity—Talbot appropriated a new physical effect” (Johnston, 2001, p. 19).^​1​

Talbot’s Law was the subject of an investigation in 1906 by Edward Pechin Hyde, a scientist responsible for photometric research in the newly formed United States National Bureau of Standards.^​25​ Hyde served as a staff member of NIST until 1908, and went on to become President of the U. S. National Committee of the International Commission on Illumination (the CIE).^​26​ He investigated whether Talbot’s law did in fact describe the the intensity of light passing through rotating sectors, which by that time were commonly used in photometric and sensitometric measurement. His focus, though, was on photometric light measurement. Hyde concluded that Talbot’s Law applied to angles from 10 degrees to 288 degrees within a 0.3% error, while outside that range the error was more significant. Hyde also verified Talbot’s law for red, green, and blue light, but found they were less accurate than white light. Others before Hyde had also questioned the validity of Talbot’s Law when sectors were open at small angles. It’s unclear how Abney responded to these claims, though Johnston suggests he was dismissive as his research was heavily based on the use of rotating sectors (Johnston, 2001, p. 148).^​1​

Abney and Festing next focused their efforts on measuring the spectral reflectance of reflective objects, which they presented to the Royal Society in 1889.^​27​ They were inspired by the free-spinning tops used by Maxwell and Gorham^​28​. These rotating disks contained several layers of colored disks, each with a slit cut through a single radius. The cut allowed the disks to slide on top of one another in various degrees. “When two or three colours are combined by rotation to form a grey, and black and white sectors are combined to match that gray, in order to ascertain the total luminosity of each colour, the angular value of the sectors being known, it is necessary to refer to the luminosity to that of some standard reflecting surface, which is naturally a white one”.^​27​ In this statement, Abney and Festing are providing the definition of spectral reflectance. The percent of light reflected from each sample color wavelength, where 100% is white.

Abney and Festing started with the same color photometer setup used in “Colour Photometry – Part I”.^​17​ Initially half the receiving screen would be white and half the sample color, but they found there was too much observer variability with this design. They described the modifications below:

The collimator, prisms, and camera were at first kept as in the colour photometer; but for the camera lens was substituted a lens divided into equal segments, which could be centrally separated, as in a heliometer. The light coming through the last prism fell as a square patch on this divided lens and the two segments were separated so that two spectra fell on the focusing screen one above the other. A slit in a card was then passed across this double spectrum, and any required ray was isolated.^​27​

The main problem addressed by Abney was the need to have a white reference for the spectral light. Modern spectrophotometers are calibrated by measuring the spectral reflectance of a white object with a known spectral reflectance. Thus, there is a direct relationship between the signal from the solid-state detector and spectral reflectance. Abney and Festing did not have this computational luxury. They could measure luminance, akin to the raw signal measurements, but needed a way to compare directly to a white reference. This is where their double prism method came into play. The two identical spectra (from the split prisms) were made to fall on the white card and sample object using shadows cast by the rod, as in the colour photometer used to measure luminance. The device is illustrated (crudely) in the figure below with a cyan sample object.

The double prism diverted some of the spectral light to the reference patch. If the sample was a white patch, then the observer would see two monochromatic patches of the same luminance. If the white sample was then replaced with a colored sample, the Reference White patch would appear brighter than the sample patch, which absorbs more of the monochromatic light.

The observer then modulated the intensity of the spectrum light falling on the white object by adjusting the rotating sectors. For example, the observer would adjust the intensity of a green band falling on the white card so it matched the intensity of the green band falling on the sample. The observer would then progress through the remaining spectral bands, the luminance for each band calculated using Talbot’s Law. This may be the first instance in which “spectral reflectance,” as compared to a reference white, was measured.

As with the color photometer spectral sensitivity measurements, two methods were used to determine the precise reflected luminance: 1) equalize intensity by moving the slit with constant sectors, or 2) adjust the sectors with a constant slit. Method 1 was used when measuring bands in spectral regions where there was a large change in luminance. Method 2 was used when measuring bands in spectral regions where there was not much change in luminance.

Another effect observed by Abney and Festing was the change in perceived hue that occurred when white light was added to a pure color. They wrote, “It was curious to note the change in colour produced on the coloured half of the screen when illuminated partially by a portion of the spectrum weak in luminosity, and partially by weak white light. It was absolutely impossible to match the colours, when even a very small percentage of white light fell on the screen.” This observation became known as the “Abney Effect” and supported the need to eliminate excess ambient light from their measurements. It’s unclear why Festing’s name was not also attributed to the “Abney Effect,” or to “Abney’s Law,” for that matter.

Spectral reflectance is defined as the proportion of light reflected from an object at each wavelength relative to a white object. The corollary is that a colored object is a spectrum with different weightings for each wavelength. Abney and Festing wrote, “If the rays of the spectrum itself could be taken in the proportions in which they are reflected from any pigment, say emerald green, and recombined, the resulting light should be of the same color as that reflected from the pigment.” Using this theory as a foundation, they tested their spectral measurements by engineering a physical weighting of the spectrum based on Talbot’s Law. One section of a rotating disk was marked with a curve corresponding to the reflectance curve of a given color sample. Reflectance was translated to arc length and market out at radii corresponding to different wavelengths. An example of such a plot for Prussian Blue from Abney’s “Colour Mixture” book is shown below. The longest radius corresponds to 400 nm and the shortest to 700 nm.

A section was then cut from the disk tracing the x-axis and the line marked along the different radial positions. The rotating disk was placed in front of the color photometer ground glass, so the spectrum overlapped the radius cut away from the disk. As the disk spun, more light would pass from spectral regions in which a greater area was cut out. The light was then recombined into the object color using a lens. This experiment could be thought of an extension to Abney’s Law, showing the additivity of not just a few primaries, but the entire spectrum. The luminance of a given color is equal to the integrated spectral luminance of component spectral bands.

In their final work together, published in 1893, Abney and Festing explored the impact of visual field on color perception^​29​, foreshadowing the research that went into the development of the 1931 2-degree standard observer by the CIE. Two patches, using the setup from their first experiment, were viewed side-by-side at three feet, at which distance the patches subtended a field-of-view of about 5 degrees. Abney and Festing understood that the area in the direct center of the visual field was different from the surrounding vision. They tested luminance measurements using a modified apparatus with patches subtending different angles of view: 1.2 deg., 1.5 deg., 1.8 deg., and 2.4 deg (by changing the size of the patches and viewing distance). No variation in measurements was found from 1-1.8 deg., but the 2.4 deg condition showed some difference. Abney and Festing attributed this effect to the presence of the macula, though they were not aware of other factors that played an important role, such as the difference in cone distribution on the retina.^​†​ They also tested a similar macular theory by placing the white reference spot six inches from the colored spot, outside the macular field of view, while the colored spot remained within the macular field of view. The observer focused on the colored spot while trying to view the white spot in their periphery. They observed “an increase in luminosity to the outer part of the retina to the portion of the spectrum from about E [528 nm] to the violet end, over that to the central part of the retina,” but that “the reverse is the case with respect to the portion from the green to the red. Evidently, therefore, the outer part of the retina is less sensitive than the central part to the less refrangible rays of the spectrum.”^​29​

At this point we take a brief pause in our series on Orthochromatic Photography. In the Part 3 we will explore how mid-19th century sensitometry late 19th century photometric practices led to developments in plate technology and tone reproduction that would impact important color, photography, and printing developments in the 20th century.

To be continued…

---

Disclaimer

This article was written by Brian Gamm in his personal capacity. The views, thoughts, and opinions expressed in this article belong solely to the author, and not necessarily to the author’s employer, organization, committee or other group or individual with which the author has been, is currently, or will be affiliated.

---

References

---

1. ^​*​
   Presidential inaugural address, to the General Meeting of the British Association, Edinburgh (2 Aug 1871). In Report of the Forty-First Meeting of the British Association for the Advancement of Science (1872), xci. https://todayinsci.com/QuotationsCategories/M_Cat/Measurement-Quotations.htm
2. ^​†​
   Most of the cells responsive to color are located in the fovea, a small section of the retina roughly in the center of the visual field. The fovea region has smaller, more tightly packed light-sensing cells, which is why our acuity is also higher in that region (think about reading this small text on your phone or computer).

1. 1.

   Johnston SF. A History of Light and COlour Measurement: Science in the Shadows. Institute of Physics Publishing; 2001.
2. 2.

   Wall EJ. Elementary Sensitometry. American Photography. 1923;17:298-309. https://www.google.com/books/edition/The_Photographic_Journal/_xw_AQAAMAAJ?hl=en&gbpv=0
3. 3.

   Warnerke L. On Sensitometers. The Journal and Transactions of the Photographic Society of Great Britain. 1879;5(5):75-92. https://www.google.com/books/edition/THE_JOURNAL_AND_TRANSACTIONS_OF_THE_PHOT/_CsFAAAAQAAJ?hl=en&gbpv=0
4. 4.

   Woodbury WE. Sensitometer. In: The Encyclopædic Dictionary of Photography. The Scovill & Adams Company of New York; 1896:1. https://www.google.com/books/edition/The_Encyclop%C3%A6dic_Dictionary_of_Photogra/cIAvAAAAMAAJ?hl=en&gbpv=0
5. 5.

   Dr. Vogel on Sensitometers. The Photographic News. 1883;27(1305):567. https://www.google.com/books/edition/The_Photographic_News/MJ7xuwZQUo0C?hl=en&gbpv=0
6. 6.

   Warnerke L. On Actinometers. The Journal and Transactions of the Photographic Society of Great Brittain. 1879;4(2):19-31. https://www.google.com/books/edition/THE_JOURNAL_AND_TRANSACTIONS_OF_THE_PHOT/_CsFAAAAQAAJ?hl=en&gbpv=0
7. 7.

   Abney W. Cantor Lectures: Recent Advances in Photography. Journal of the Society of Arts. 1882;30(1549):897-905. https://www.google.com/books/edition/Journal_of_the_Royal_Society_of_Arts/UexFAQAAMAAJ?hl=en&gbpv=0
8. 8.

   Fatet J. Recreating Edmond Becquerel’s electrochemical actinometer. Archives Des Sciences Journel. 2005;58:147-156. https://www.researchgate.net/publication/281697105_Recreating_Edmond_Becquerel’s_electrochemical_actinometer
9. 9.

   Spence JB. MANUFACTURE OF METALLIC COMPOUNDS FROM SULPHUR AND SULPHIDES,. Published online 1881. https://www.google.com/books/edition/Official_Gazette_of_the_United_States_Pa/-vx8AAAAMAAJ?hl=en&gbpv=0
10. 10.

    Horman W. A MEDAL MADE FROM SPENCE’S METAL. The E-Sylum. Published 2010. https://www.coinbooks.org/esylum_v13n47a12.html
11. 11.

    Pritchard HB. Mr. Leon Warnerke At Champion Hill. In: The Photographic Studios of Europe. Piper and Carter; 1882.
12. 12.

    Stolze F. Upon A Standard Light For Sensitometers. The British Journal of Photography. 1881;28:447-448, 459-460. https://www.google.com/books/edition/The_British_Journal_of_Photography/T54OAAAAQAAJ?hl=en&gbpv=0
13. 13.

    England Births and Christenings, 1538-1975. FamilySearch. Published March 21, 2020. Accessed October 9, 2021. https://www.familysearch.org/tree/person/sources/KH3Y-QN1
14. 14.

    Hannavy J, ed. William de Wiveleslie Abney. In: Encyclopedia of Nineteenth-Century Photography. Routledge, Taylor & Francis; 2013:103.
15. 15.

    Abney W. The Speed of Plates. Anthony’s Photographic Bulletin. 1897;28(11):340-342. https://www.google.com/books/edition/Anthony_s_Photographic_Bulletin/65cPAQAAIAAJ?hl=en&gbpv=0
16. 16.

    Major-General E.R. Festing, C.B., F.R.S. Nature. 1912;89:299. https://www.google.com/books/edition/Nature/9I5FAAAAYAAJ?hl=en&gbpv=0
17. 17.

    Abney W, Festing ER. Colour Photometry – Part 1. Philosophical Transactions of the Royal Society of London. 1886;177(2):423-456. https://www.google.com/books/edition/Philosophical_Transactions_of_the_Royal/qhE2AQAAMAAJ?hl=en&gbpv=0
18. 18.

    Abney W. Colour Vision Being the Tyndall Lectures Delivered in 1894 at the Royal Institution. William Wood & Company; 1895. https://archive.org/details/colourvisionbei02abnegoog/page/n12/mode/2up
19. 19.

    Carolina Sparavigna A. Carbon-Arc Light as the Electric Light of 1870. ijSciences. Published online 2014:1-7. doi:10.18483/ijsci.581
20. 20.

    Abbott J. Light. Harber & Brothers; 1871. https://www.google.com/books/edition/Light/QEEIAAAAIAAJ?hl=en
21. 21.

    Shepherd ES. The Scientific Translation of Colour into Monochrome. The Photographic Journal. 1898;22(11):247-359. https://www.google.com/books/edition/The_Photographic_Journal/_xs_AQAAMAAJ?hl=en&gbpv=0
22. 22.

    Grassmann H. Zur Theorie der Farbenmischung. Annalen der Physik und Chemie. 1853;165(5):69-84. http://iitp.ru/upload/content/1512/GrassmannTheorie%20der%20Farbenmischung1853.pdf
23. 23.

    Cohen J. Visual Color and Color Mixture. University of Illunois Press; 2001.
24. 24.

    Abney W. Colour Measurement and Mixture. Society For Promoting Christian Knowledge; 1891.
25. 25.

    Hyde EP. Bulletin of the Bureau of Standards. 1906;2(1):1-32. https://www.google.com/books/edition/Talbot_s_Law_as_Applied_to_the_Rotating/SfPOL_Y5LysC?hl=en&gbpv=0
26. 26.

    Lide DR. National Institute of Standards and Technology; 2001:1. https://www.govinfo.gov/content/pkg/GOVPUB-C13-310bc7b3121ed82b8f13fc15a5c9e639/pdf/GOVPUB-C13-310bc7b3121ed82b8f13fc15a5c9e639.pdf
27. 27.

    Abney W, Festing RE. Colour Photometry – Part II. The Measurement of Reflected Colours. Philosophical Transactions of the Royal Society of London. 1889;179:547-570. https://www.google.com/books/edition/Philosophical_Transactions/13k7AQAAMAAJ?hl=en&gbpv=0
28. 28.

    Gorham J. J. E. Adlard; 1859. https://www.google.com/books/edition/The_Rotation_of_Coloured_Discs_Applied_t/ZrdbAAAAcAAJ?hl=en
29. 29.

    Abney W, Festing RE. Colour Photometry – Part III. Philosophical Transactions of the Royal Society of London. 1893;183:531-565. https://www.google.com/books/edition/Philosophical_Transactions_of_the_Royal/Ss6RL1wYh28C?hl=en&gbpv=0