It is apparent that neither of these models fit well with our description of the development of the duplicity theory. As regards Popper's model, none of the classical theories of Newton, Young, Schultze, Kühne and Hering was triggered by an attempt to falsify or refute current hypotheses or theories. Newton's starting point was the rectangular form of the prismatic solar spectrum he observed when he looked at its beautiful colours; Young based his theory on the old and well-known fact that three pigments were sufficient to produce every object colour; Schultze reflected on the fact that nocturnal and diurnal animals tended to have rod- and cone-dominated retinas, respectively, and that colours were absent in night vision; Kühne's theory was instigated by the great discovery of Boll that rhodopsin bleached in light and regenerated in the dark; and Hering was spurred on by Mach's psychophysical maxim. Thus, there is little evidence of long, fruitful falsification periods ending with falsification of the most successful hypothesis and the development of a new, better one triggered by this last falsification.

It is evident that Maxwell (1855, 1860) provided conclusive evidence in support of Young's trichromatic colour theory by demonstrating that three standard spectral lights were sufficient to produce all spectral colours, but his experiments represented a confirmation of an already existing theory and not a falsification. It should also be noted that Young's theory was immediately accepted by the relevant scientific community without much debate following the discoveries of Maxwell.

HECHT'S PHOTOCHEMICAL THEORY

The finding of Loeser (1904) that cones also had the ability to increase their sensitivity during early dark adaptation was confirmed by Hecht (1921/1922). He found that cones could increase their sensitivity markedly even during the first few seconds after bleaching.

More importantly, however, Hecht developed a photochemical theory for dark and light adaptation of rods and cones that had a strong influence on a whole generation of research workers. Certainly, he has played a central role in the developmental history of the duplicity theory. In a series of papers he provided an array of evidence supporting his photochemical theory (see Hecht, 1919/1920a, b, c, 1921/1922). In its essence and in its most simple version, the theory runs as follows: light acts on a photosensitive substance S and decomposes it into two precursors called P and A. The sensitivity of the eye, then, depends on the concentration of these precursors, not on the quantity of the photosensitive substance S. Thus, the model states that the amount of fresh precursors necessary for a threshold response is always a constant fraction of the amount of the precursors already present in the system. Hence, dark adaptation was thought to depend on the regular decrease in the concentration of the residual precursors present in the sensory system. This decrease was assumed to proceed according to the dynamics of a bimolecular reaction, to be independent of light stimulation, and, in accord with Kühne's (1879) ‘Optochemische’ hypothesis, to result in a reformation of the photosensitive substance S. Thus, the model states that the amount of fresh precursors necessary for a threshold response is always a constant fraction of the amount of the precursors already present in the system.

A breakthrough in our understanding of the underlying mechanisms of dark and light adaptation of rods and cones would, of course, have represented an important step forward in the development of the duplicity theory. Yet, due to the complexity of the task involved, the progress in our understanding has proceeded at a very slow pace. Indeed, no general agreement about basic sensitivity regulation mechanisms of rods and cones has yet been reached (see Cameron et al., 2006; Baehr et al., 2007).

In line with Parinaud's (1885) assumption that both light and dark adaptation were determined by a changes in the amount of rhodopsin in the rods, most of the leading research workers during the hundred years that followed tended to conceive of adaptation as controlled by photochemical processes in the outer segment of the receptors. Of course, the research workers knew that this presumption was an oversimplification. Thus, it had long been known that sensitivity regulation of the visual system could not be controlled by one single mechanism operating at one site only, but rather was the result of different mechanisms engaged at different sites in the visual pathway. Besides the obvious regulation of the incident photons by the variation of the pupil size (in humans the diameter of the pupil may vary from about 8 mm in scotopic to about 2 mm in photopic vision, reducing the light incident on the receptors by about 1.2 log units), evidence had been provided indicating that important light- and dark-adaptation mechanisms were located more centrally than the outer segment of the receptors (e.g. Kuffler, 1953 ; Lipetz, 1961 ; Rushton, 1965a, 1972 ; Barlow, 1972).

A first clue to an understanding of the sensitivity regulation mechanism in rods was given by the discovery by Boll in 1876 that the photopigment rhodopsin, situated in the outer segment of the rod receptors, was bleached by light and regenerated in the dark. He also showed that the bleaching effect of light depended on the wavelength used (Boll, 1878).

Influenced by this great discovery, Kühne, in an extensive research work, provided strong evidence in favour of the view that the sensitivity difference between rods and cones had a photochemical basis (Kühne, 1877a, b, 1877–1878, 1879). He investigated the bleaching and regeneration processes of rhodopsin in much more detail than Boll and made an important theoretical contribution with his influential ‘Optochemische’ hypothesis, where he presumed that the phototransduction in both rods and cones was photochemical in nature. Accordingly, he presumed that the apparent colourless cone receptors contained photochemical substances, and that these substances became involved in visual processing under daylight conditions. Indeed, in opposition to Schultze's (1866) duplicity theory, he presumed that even rods were activated by photochemical, colourless substances in daylight.

Furthermore, Kühne (1879) made an important distinction between the photochemical substances and their photoproducts, and argued that it was the photoproducts, not the photosensitive substances that generated the neural activity in the retina. Moreover, he discovered that rhodopsin may regenerate in two quite different ways: a rapid anagenese from photoproducts of rhodopsin and a slower neogenese from new substances formed after rhodopsin had been bleached.

DARK AND LIGHT ADAPTATION BASED ON SIMILAR MECHANISMS

In contrast to the assumption of Rushton that light and dark adaptation were determined by quite different mechanisms, Barlow (1964) put forward the hypothesis that their underlying mechanisms were very similar; that the sensitivity regulation during dark adaptation was mainly determined by so-called 'photon-like' events. Thus, he presumed that the presence of bleached photopigment caused the receptors to show effects similar to those induced by light. Both the bleached photopigments and light increased the intrinsic noise level of the receptors. In fact, the effect of bleached pigments was assumed to add to that of light so that the separate effects of the two factors could not be distinguished at the receptor output.

BOTH NOISE AND NEURAL MECHANISMS INVOLVED

There was, however, a serious obstacle to Barlow's noise theory, since the expected square root law was not fulfilled in incremental threshold measurements. Thus, if a light flooding the retina elevated the threshold solely by virtue of the intrinsic noise level of receptors (i.e. the statistical fluctuation in the number of quanta absorbed) the incremental threshold should rise in proportion to the square root of the intensity of the flooding light (see De Vries, 1943; Rose, 1953; Barlow, 1957). Opposed to this prediction, however, it had been found that, over a large range of background intensities, the incremental threshold for a large test field rose in direct proportion to the background intensity (the Weber law).

EACH RECEPTOR TYPE HAS A SEPARATE AND INDEPENDENT ADAPTATION POOL

Presupposing that the site of adaptation is located in summation pools situated centrally to the receptor level, one might expect that different kinds of receptor (rods, ‘red’, ‘green’ and ‘blue’ cones) would tend to interact in determining the adaptation process. In opposition to this view, however, Rushton (1965a) presented evidence suggesting that each class of receptor had a separate and independent AGC pool. Thus, he suggested that for rods and for each kind of cone the dependence of threshold on background and on bleaching were private, i.e. separate and independent.

Strong evidence in favour of independence between rod and cone adaptation pools was given by the well-known fact that the threshold level obtained during long-term dark adaptation may remain unchanged for several minutes during the cone-plateau period, while the sensitivity of the rods may increase by several log units. Indeed, with deep-red test light, the absolute threshold level of the cones may remain invariant for more than 20 min during the cone-plateau period before the rods eventually influence the threshold level. Furthermore, well-founded evidence in favour of the independence of the adaptation pools for all the different receptor types was found by Rushton (1965a) in the extensive research work of Stiles on incremental threshold (see Stiles, 1978). This research work may be illustrated by Stiles's well-known experiment together with Aguilar in 1954 on light adaptation of the rod mechanism.

At the end of the mid-twentieth century period (1930–1966), spectrophotometric measurements of spectral absorption in single photoreceptors had conclusively demonstrated that the retina contained three different types of cone, each cone with only one photopigment, together with the rods that contained only rhodopsin. Hence, it had been proved that the speculative idea of König (1894) and Willmer (1946, 1961) that rods may function as the primary ‘blue’ receptor, Willmer's and Granit's ideas of 'day rods', Polyak's idea that the trichromacy of colour vision was based on three types of bipolar cell, and the dominator-modulator theory of Granit, were all wrong. Moreover, Lie (1963), in his important and comprehensive study, had found that rods under scotopic test conditions mediated achromatic colour sensations only, and that under mesopic test conditions they contributed an achromatic component, desaturating the chromatic cone component.

Apparently, with the exception of the idea of rod-cone interaction, the duplicity theory had become strikingly similar to the old, orthodox conceptions formulated by von Kries (1929).

ELABORATION AND REVISION OF THE TWO MOST BASIC ASSUMPTIONS OF SCHULTZE'S DUPLICITY THEORY

As can be seen, the developmental history of the duplicity theory in the 100 years between 1866 and 1966 may be characterized by complex and comprehensive theory constructions. Outstanding examples are the theories provided by Schultze (1866), Parinaud (1881, 1885), König (1894), von Kries (1929), G. E. Müller (1930), Polyak (1941), Granit (1947, 1955) and Willmer (1946, 1961).

This hypothesis of Kühne and Hecht must be considered an important insight with regard to biochemical processes underlying dark adaptation. Yet, it only represented a first step towards an understanding of the highly complex processes involved in the photochemical sensitivity regulation mechanisms of rods and cones. Obviously, a deeper understanding would require more information on both the molecular structure of the rod and cone photopigments and the bleaching and regeneration processes generated by light.

MOLECULAR BASIS OF BLEACHING AND REGENERATION OF PHOTOPIGMENTS IN RODS AND CONES

Inspired by Hecht, Wald, in the early 1930s, set out to throw light on these largely unexplored research topics (see Wald, 1949a, 1958, 1968). His profound discoveries and insights earned him the Nobel Prize which he shared with Granit and Hartline in 1967.

Firstly, he discovered vitamin A in the retina (Wald, 1933, 1934/1935). Shortly thereafter, he concluded that the photopigment rhodopsin was a conjugated carotenoid-protein engaged in a bleaching-regeneration cycle when acted upon by light (Wald, 1934, 1935/1936). Thus, in line with the hypothesis of Kühne and Hecht that light decomposes rhodopsin into its two precursors, Wald presumed that the caroteoids, all-trans retinal (vitamin A aldehyde) and vitamin A represented both photoproducts and precursors of rhodopsin.

THE COLOUR THEORY OF TSCHERMAK

In developing his own colour theory, Tschermak (1902, 1929) made a critical evaluation of the colour theories of Young-Helmholtz, Schultze and Hering. With regard to Young-Helmholtz's colour theory, he was severely critical. Thus, he asserted that the basic assumption of three independent, primary colour-related processes postulated by the trichromatic theory could not be reconciled with the phenomenological analysis of colour sensation that revealed six qualitatively different unitary sensations: red, yellow, green, blue, white and black. It would, for example, be impossible to give an adequate explanation of the uncompounded yellow-related material process by green- and red-related processes, or the uncompounded white-related process by red-, green- and violet-related processes. Also, in opposition to the trichromatic colour theory, experiments on colour mixture, colour induction and colour contrast clearly revealed opponent interaction processes going on in the visual system.

Finally, in accord with von Kries (1911), Tschermak (1902, 1929) pointed out that the basic assumption of the trichromatic colour theory, that white sensation was generated when the three different types of cone receptors were activated to about the same degree, was seriously challenged by the fact that colourless sensation could also be observed in scotopic vision where only rod receptors were known to function.

With regard to Schultze's duplicity theory, on the other hand, he found the evidence strongly in favour of its basic assumptions that cones functioned in day vision giving rise to both achromatic and chromatic sensations, and that rods functioned in night vision, giving rise toachromatic sensation only.

The most basic assumption of the orthodox duplicity theory that only rods function in night vision and only cones in day vision has been generally agreed on during the whole developmental period from Schultze (1866) to the present (2009).

Schultze had provided two strong arguments in favour of this basic assumption. Firstly, rods and cones tended to dominate in nocturnal and diurnal species, respectively. In fact, some nocturnal species had a pure rod and some diurnal species a pure cone retina. Secondly, colour vision, a characteristic feature of cones of the central fovea, was absent in night vision.

Reasoning from this assumption, one would expect that a reduction in light intensity from that of daylight would lead to the functional appearance of rods at some particular intensity level. Schultze, however, did not state where and how this transition from cone to rod vision and vice versa happened. Furthermore, he did not provide any theory to explain the difference in sensitivity between the two receptor types. This reveals a serious limitation in his theory construction and shows that the most basic assumption of his theory was without any adequate explanation.

Several of the leading scientists within vision research have attempted to bridge this gap in Schultze's theory construction by developing theories of the sensitivity regulation mechanisms ofthe rod and cone receptor systems. Actually, the development of these theories may be considered a main thread in the history of the duplicity theory.

Newton (1675) had rightly presumed that light signals were transformed in the retina, and that the information about the visual world was conveyed to the brain by the optic nerve. His suggestion that visual information was transmitted by a vibration code, however, proved to be wrong. Thus, early in the twentieth century it had become generally accepted that the optic nerve fibres reacted to light by discharging a series of brief electrical action potentials. With increasing light intensity, the fibres tended to increase their firing rate, but the size of the discharge remained constant. (For a review of how our knowledge about the electrical nature of nerve impulses emerged, see Boring, 1957, pp. 30, 39–43 and Granit, 1947.)

THE ELECTRICAL RESPONSES TO LIGHT STIMULI IN SINGLE OPTIC NERVE FIBRES

H. K. Hartline was the first to make a thorough investigation of the electrical responses to light stimuli in single optic nerve fibres. A small bundle of fibres dissected from the optic nerve was split successively until only a single fibre remained. Thereafter, the electrical activity of the fibre, generated by a light stimulus, was recorded by means of an oscillograph capable of registering small, rapid voltage fluctuations. The recordings were made under conditions where the eye was illuminated by light of various intensities, durations and wavelengths (see Hartline, 1940, for a review).

In retrospect, it can be seen that all the classical theories in vision research with the exception of Hering's opponent colour theory were instigated by what may be termed observational facts. The theories of Newton (1671/1672) about light and colour, for instance, were triggered by his observation that the prismatic solar spectrum was rectangular in form.

Accepting the laws of refraction, he had expected the form to be circular. Yet, comparing the length of the spectrum with its breadth, he found to his surprise that it was about five times greater. He also found the two sides of the rectangle to be straight lines and the ends to be semicircular. It is important to note that Newton's observations did not result from an attempt to confirm or falsify the refraction laws. Instead, the observations were made quite accidentally.

To explain his observations, Newton initiated a series of experiments, testing various hypotheses and soon reached the famous conclusion that white sunlight was compounded of an innumerable number of different rays and that the colours of the prismatic spectrum were original and connate properties of these rays – all in sharp contrast to the generally held view that the colours of the spectrum were qualifications of the white homogeneous sunlight caused by prismatic refraction.

Schultze's duplicity theory was also based on observational facts. Thus, he arrived at his theory by combining two different sets of data. (1) Diurnal and nocturnal animals tended to have retinas dominated, respectively, by cone and rod receptors.

The new discoveries of Polyak, Hartline, Kuffler and Granit clearly indicated that the old orthodox formulation of the duplicity theory had serious flaws. In particular, the seminal discovery of Polyak and Granit that rods shared neural pathways with cones appeared to contradict the basic assumption of the theory that rods and cones functioned independently of each other. Obviously, the time was ripe for a reformulation of the theory to make its statements consistent with the new discoveries. In addition to the reformulations proposed by Polyak and Granit (see above), important changes to the theory were made both by Willmer (1946, 1961), Saugstad and Saugstad (1959) and Lie (1963).

Willmer's reformulation represented a quite new version of the duplicity theory. Indeed, he held that the rods under photopic conditions played an important role in trichromatic colour vision.

Saugstad and Saugstad called for a more moderate revision of the theory. They held that its statements should be reformulated making the meaning more explicit. Also, they pointed to well-founded evidence against the theory and concluded that its statements should be qualified by the development of a more comprehensive theory, taking into account the structure and function of the nervous system.

The contributions of Willmer and Saugstad and Saugstad were primarily theoretical. Lie, on the other hand, made a thorough empirical, psychophysical investigation focused on rod-cone interactions in colour vision under mesopic test conditions.

NEWTON'S UNIVERSAL COLOUR THEORY

One major root of the duplicity theory as formulated by Schultze, von Kries and G. E. Müller is represented by the Newton tradition. Within this line of research a rudimentary understanding of the cone mechanisms developed, ending up with the formulation of the famous Young-Helmholtz trichromatic colour theory (Helmholtz, 1867). This theory profoundly influenced Schultze, von Kries and G. E. Müller in their attempt to construct their theories. In fact, the theory forms an integrated part of the duplicity theory and its development may therefore be seen to represent the starting point of the development of the duplicity theory.

Certainly, the development of the trichromatic theory was in many ways initiated by Newton's ingenious experiments and theories on light and colour. In fact, his contribution deserves to be ranked as the first major paradigm shift within vision research in modern times. Surprisingly, however, Newton's revolutionary ideas about light and colour are, today, not generally well known. In the following, therefore, we present his theories in some detail.

Newton's theories were first published 19 February 1672 as a letter in Philosophical Transactions of the Royal Society of London (1671/1672). His most important ideas about light and colour are stated below in his own words. They are given in his propositions 1, 2, 3, 7 and 13 (see also Cohen, 1978, pp. 53–57).

The spikes recorded by Hartline and Kuffler with the microelectrode technique represented end products in a series of consecutive events in the retina, starting with absorption of photons. To gain information about these extremely complex events that preceded the discharges of the optic nerve fibres, the research workers had to rely on the measurements of the electroretinogram (ERG). Indeed, it was generally held that mechanisms underlying the ERG response directly determined the impulse pattern of the optic nerve.

The ERG technique was first employed in 1865 by Frithiof Holmgren, a Swedish physiologist. He applied a pair of electrodes to an eye and found that the galvanometer connected to the electrodes gave a marked deflection both when the eye was illuminated and when the light was turned off. (For a description of the development of this technique, see introduction section of Granit, 1963.)

SUPPORTING EVIDENCE FOR THE DUPLICITY THEORY FROM THE ERG TECHNIQUE

A very extensive review of the research literature on ERG was made by Granit (1947). Presuming that the ERG response represented an average reaction, reflecting the processes of activated photoreceptors, the evidence reviewed was found to support the duplicity theory, suggesting that there were two quite different ERG response patterns of the retina, the so-called E- and I-ERG responses – the former characteristic of rod-dominant and the latter of cone-dominant retinas.