r The Old Corner Book Stcre, Inc, ^"'0"' - Mass. The Cambridge Psychological Library AN INTRODUCTION TO THE STUDY OF COLOUR VISION CAMBRIDGE UNIVERSITY PRESS C. F. CLAY, Manager Hotliloti: FETTER LANE, E.G. CBlJmtiutgf) : loo, PRINCES STREET ILontfon: H. K. LEWIS, 136 GOWER STREET, W.C. Bmnba^ aiiti ffalruftn : MACMILLAN AND CO., Ltd. fforonta: J. M. DENT AND SONS, Ltd. STokoo: THE MARUZEN-KABUSHIKI-KAISHA All rights reserved ■ff o- — l^ ^ -^ o lO _ s? 8 1. S SfeSs)**:^ D QC o w P. (f) u I — I < 00 a w AN INTRODUCTION TO THE STUDY OF COLOUR VISION BY J. HERBERT PARSONS, D.Sc, F.R.C.S., OPHTHALMIC SURGEON, UNIVERSITY COLLEGE HOSPITAL; SURGEON, ROYAL LONDON (mOORFIELDS) OPHTHALMIC HOSPITAL Cambridge : at the University Press New York : G. P. Putnam's Sons 1915 Cambritgr : PRINTED BY JOHN CLAY, M.A. AT THE UNIVERSITY PRESS .'> \ PREFACE rpHE vast literature on colour vision consists almost entirely of -^ papers written in support of some particular theory. It is peculiarly difficult to obtain a general and unbiassed view of the subject. I have here endeavoured to separate the best established facts of colour vision from the theories, and have then discussed the chief theories in the light of these facts. I wish to express my great indebtedness to Sir WilUam Abney, K.C.B., F.R.S., Professor W. Watson, F.R.S., and Dr Myers for their invaluable assistance in a very difficult task, which demands a not inconsiderable knowledge of such diverse subjects as physics, physiology, and psychology. The chief references to the literature have been given in footnotes. The frequent references to " Abney " are to his Researches in Colour Vision (London, 1913), and to " v. Helmholtz" to his Hand- buch der physiologischen Optik. In the latter case each of the three editions contains material which is absent from the others, and the exact reference is given in the footnote. Note. Fig. 12, p. 47. The discrepancy in these curves has been found to be due to a technical defect in the apparatus. More recent observations show the identity of the curves obtained by the two methods, thus confirming Ives's results (Fig. 38). J. H. P. November, 1914 / G 0 zt CONTENTS Section Chap. Chap. Chap. Section Chap. Chap. Chap. Chap. Section Chap. Chap. Section Chap. Chap. Section Chap. Chap. Chap. Section Chap. Chap. Section Chap. Chap. Chap. Chap. PART 1 THE CHIEF FACTS OF NORMAL COLOUR VISION I. The Bases of Colour Vision. page I The Physical Basis 1 II. The Anatomical Basis 7 The Psychological Basis . . . . . . . 17 The Spectrum as seen by the Light-adapted (Photopio) Eye. The Spectrum: Hue, Luminosity, Saturation ... 27 The Discrimination of Hue in the Spectrum ... 30 The Mixture of Pure-colour Stimuli 33 Tlie Luminosity of the Spectrum 42 The Spectrum as seen by the Dark-adapted (Scotopic) Eye. III. IL I. II. Ill, IV. III. I. IL IV. I. II. V. I. II. Ill VI. I. II. VII. I. II. Ill IV. Adaptation or Temporal Induction Scotopia or TwiUght Vision Regional Effects. The Field of Vision for Colours The Macula lutea and Fovea centralis .... Temporal Effects. Recurrent Vision; the Talbot-Plateau Law; the Flicker Phenomenon ........ Successive Induction or After-images .... , The Effects of "Fatigue" Areal Effects. The Local Quantitative Effect ...... Simultaneous Contrast or Spatial Induction The Evolution of Colour Vision. Introduction ......... The Comparative Psychology of Colour Vision . The Colour Vision of Primitive Races . . . . The Development of Colour Vision in the Child 49 52 67 81 85 101 112 117 125 130 131 145 152 PART II THE CHIEF FACTS OF COLOUR BLINDNESS Chap. I. Introduction: Colour Names 158 Chap. II. Dichromatic Vision 162 Chap. III. Anomalous Trichromatic Vision 182 Chap. IV. Monochromatic Vision 186 Vlll CONTENTS PART III THE CHIEF THEORIES OF COLOUR VISION Section I. General Review. Chap. I. Introduction ..... . 193 Chap. II. Historical Review of Modern Theories of Colour Vision 196 Section II. The Duplicity Theory . 203 Section III. The Three-Components Theory (Young-Helmholtz). Chap. I. Statement of the Theory . 213 Chap. II. Researches based upon the Theory I. Normal Colour Vision . 220 II. Dichromatic Vision • • 232 III. Anomalous Trichromatic Vision : Approximate Dichromatism > • 235 IV. Anomalous Trichromatic Vision : Shift of a Sensatior I Ctu-ve > 243 Section IV. The Opponent Colours Theory (Hering). Chap. I. Statement of the Theory , , 251 Chap. IT. Researches based upon the Theory . 263 Section V. Other Theories. I. Donders' Theory . 270 II. Ladd-Franklin's Theory 271 III. McDougall's Theory 274 IV. Schenck's Theory . 285 V. Wundt's Photochemical Theory 289 VI. G. E. Muller's Theory 290 VII. Edridge-Green's Theory 291 Index of Subjects . , 301 Index of Authors . . 305 The Prismatic Spectrum of Sunlight, showing the positions of the chief Fraunhofer lines and wave-lengths . . .... Frontispiece The author wishes to make grateful acknowledgment for permission to repro- duce the following iDustrations : to the Royal Society for Figs. 68 — 74; to the Royal Society and Messrs Longmans, Green & Co. for Figs. 11, 12, 14, 19—23, 28—33, 40, 41, 49, 61—65; to the Kaiserliche Akademie der Wissenschaft of Vienna for Figs. 5, 60, 66 and 67 ; to the proprietors of the Journal of Physiology for Figs. 2, 3 and 17, of the British Journal of Psychology for Figs. 34 — 37, 42 and 43, of the Philosophical Magazine for Figs. 38 and 39, of Mind for Fig. 75, and of the Zeitschrift fur Psychologic und Physiologie der Sinnesorgane for Figs. 15, 16, 25 — 27, 45, 46 and 50 — 55; and to Messrs Vieweg und Sohn, Braunschweig, for Figs. 1, 6, 7, 13 and 57. PART I THE CHIEF FACTS OF NORMAL COLOUR VISION SECTION I THE BASES OF COLOUR VISION CHAPTER I THE PHYSICAL BASIS What is generally understood by the term " light " is a composite congeries of allied manifestations of energy, comprising such apparently various phenomena as heat, Hght in the narrower sense of the word, and chemical action. Various as these phenomena are, they are physically identical in character, all consisting of radiant energy in the form of waves of identical character, differing only in the length and rapidity of the vibrations. Broadly speaking, the longest waves cause the sensation of heat, the shortest give rise to chemical action, whilst those of intermediate length cause the sensation of light. If we take ordinary sunlight as the basis of our investigations, it is possible to split it up by appropriate means into its component "rays," differing from each other in wave-length. Of these certain are visible, and constitute light in the narrower sense of the word, but instead of giving rise to the sensation of white light, they, to the majority of people, show certain pure colours, viz. red, orange, yellow, green, blue, and violet, in order, the red having the longest and the violet the shortest wave-length. The visible spectrum extends from about 723 ixfi at the red end to 397ju,/li at the violet end. v. Helmholtz under the most favourable conditions was able to see as far as about 8^5 [x/jl. The limitation of the spectrum at the violet end is less precise, because the rays in this neighbourhood are changed into rays of greater wave-length p. c. V. 1 2 COLOUR VISION by the media of the eye, particularly the lens and retina. This " fluor- escence " causes them to produce a lavender-hued sensation, which does not denote true visibility of the short wave-length rays. Beyond the red end are waves of greater length (extending to 60,000 /^/j), which when absorbed cause a rise in temperature ; beyond the violet end are waves of smaller wave-length, which are capable of causing chemical action. So striking is the physiological phenomenon of the visibility of the intermediate series that the heat rays are commonly spoken of as " infra-red," and the actinic or chemical rays as " ultra-violet." This custom is unfortunate, since it tends to obscure the importance of the physical uniformity of the series. For example, not every normal individual is able to see all the rays from 723 fifx to 397 /x/x ; for most people the range is less extensive, roughly from 700 /x/x to 400 /x/x. Again, though the ultra-violet rays are particularly potent in inducing chemical action, the visible rays are also, but in less degree, actinic, and the same is true, in still less degree, of the infra-red rays. Further, all rays when absorbed cause a rise in temperature. The most convenient and striking method of demonstrating actinicity is by the photographic film, so that we have come to regard a photograph of the spectrum as a complete analysis of the light under observation, too often forgetting that the photographic effect varies with the specific sensitiveness of the film to particular groups of waves. Thus it is only by specially sensitized films, invented by Sir William Abney, that it is possible to demonstrate infra-red rays photographically. It is further essential that the methods employed for analysis of the light be suitable for their purpose. For example, an ordinary spectro- scope, with glass prisms and lenses, suffices to demonstrate the visible spectrum, but is almost useless for showing the ultra-violet rays, since these are absorbed by the glass. In order to demonstrate the full extent of the spectrum it is necessary to use a train of lenses and prisms made of quartz or Iceland spar, which allows a maximum of rays to pass unimpeded. Probably more error has crept into the subject of colour vision from inexact description of experimental conditions and the nature of the stimuli employed than from any other cause. Two green lights may appear identical in colour to the eye, yet their physical characters may differ widely. Again, mixing a yellow and a blue pigment will produce a green pigment, yet the more general statement that green results from mixing yellow and blue is not accurate. The complete range of simple colours can be obtained in a pure state THE PHYSICAL BASIS by only two methods, dispersion and diffraction. When white light is passed through a glass prism, as in Sir Isaac Newton's original experi- ment, a spectrum is obtained. Only under certain, now well-defined conditions is such a spectrum pure, i.e. the colours do not overlap. It is commonly said that the white light is " split up " into its component parts, which are coloured. Lord Rayleigh has given sound reasons for the view that white light is not thus analysed into component parts, but that the periodicities characteristic of the several rays are in reality imposed by the prism and are not antecedently present in the white light. Be this as it may, dispersion of white light by prisms enables us to obtain coloured light in a pure state. By passing white light through a diffraction grating a pure spectrum can also be obtained. This method has the advantage that the deviation of the component rays varies within narrow limits directly with the wave-length, i.e. equal differences of wave-length are separated by equal distances in the spectrum. It suffers, however, from the disadvantage that the spectrum is less bright and less extended than the prismatic spectrum, and from the still greater objection that the interference spectrum is never free from scattered Kght. In the prismatic spectrum the dispersion increases as the wave-length diminishes, so that the violet end is much more extended than the red end and its intensity is diminished. Moreover the amount of dispersion depends upon the character of the prism or prisms employed. Hence it is necessary for accurate observations that each prismatic spectrum shall be calibrated. The Fraunhofer lines, being absolutely constant in situation, afford a series of fixed points from which the calibration curve of the given spectrum can be constructed^. The o A. U. A = 7606 in extreme red. B = 6869 in deep red. Lithium . . = 6707 in bright red. Hydrogen C = 6564 in bright red. Sodium . . .. A = 5897 in orange. Sodium . . D> = 5891 in orange. Thallium . . = 5351 in yellow green E = 5271 in green. Magnesium .. b, = 5184 in green. Magnesium b-2 = 5174 in green. Hydrogen F = 4862 in blue green. Strontium = 4609 iu blue. G = 4308 in violet. Calcium . . H = 3969 in extreme violet Calcium . . K = 3934 in extreme violet ^ Burch, Practical Exarcincs in Physiological Optics, p. 102, Oxford, 1912. 1—2 4 COLOUR VISION o ^ o table gives the principal lines m Angstrom units (1 A.U.— one ten millionth part of a millimetre = 0"1 ixfi). The most convenient method of calibration, however, is by the mercury lines as given by the " mercury arc^." In spite of the necessity for calibration the prismatic spectrum is more generally suited than the diffraction spectrum for physiological experiments on account of its greater brightness and relative freedom from scattered light. Whatever spectrum be employed the source of hght must be constant. Lights which we commonly regard as giving " white light," such as sunlight, the arc light, incandescent light, and so on, vary much in character and consequently in the constitution of their spectra. Sunlight varies so much that it is generally unsuitable for the purposes in view, the variations being not only in intensity but also in com- position, owing to the unequal absorption of different rays by the atmosphere, and this absorption again varies greatly according to the amount and nature of the matter suspended in the air. The arc light is the most satisfactory, and after this probably the Nernst lamp, though the latter has not yet been sufficiently investigated^. Less satisfactory are gas light, petroleum and so on, but as many of the experiments of earlier observers have been made with such sources they have to be considered if these researches are to receive due appreciation. Some sources of illumination, especially used for investigation of the ultra- violet rays, such as the Schott uviol mercury vapour lamps, are wholly unsuited, since they do not give continuous spectra. For experiments on colour vision many such details which cannot be discussed here must be attended to^. Suffice it to say that by taking proper precautions it is possible to obtain a spectrum which is practically constant during the time necessary to take a series of observations and which can be reproduced from time to time with a minimum of variation. If such a spectrum is viewed through the eyepiece of an ordinary spectroscope a direct spectrum is seen. This method has usually been adopted, as for example by Aubert, von Helmholtz, Clerk-Maxwell and others. By a slight change in the optical arrangements the spectrum 1 Watson, Practical Physics, p. 309, 1906. 2 Abney, Researches in Colour Vision, 1913, Chap. v. ; Golant, Ztsch. f. Sinnesphysiol. XLiii. 70, 1908. ^ E.g. Tigerstedt, Handb. d. physiol. Methodik, Bd. in. Abt. 2, Sinnesphysiologie n. Lpipzig, 1909. THE PHYSICAL BASIS 5 can be accurately focussed upon a screen. Such a projected spectrum can then be viewed by several observers at the same time, a very con- siderable advantage in testing colour vision. The use of a projected spectrum necessitates further care in detail, for the character of the spectrum will depend upon the optical properties of the screen^. A matt white surface must be used and that obtained with magnesium oxide is probably best. In order to obtain the most accurate information from the experi- ments the observations must be as far as possible quantitative and not merely qualitative. In many physiological experiments this counsel of perfection cannot be complied with and we are reduced to the informa- tion which can be obtained from merely qualitative observations. When, however, it is possible to obtain quantitative results it is generally necessary to have a constant light for purposes of comparison. Now, photometry is admittedly one of the most faulty of physical measure- ments, chiefly, to use a paradox, because it is, in most cases, not really physical but physiological. One of the most important and unique features of Sir William Abney's apparatus is that the intensity of the comparison light bears a constant physical relationship to that of the spectrum used, since it is obtained by the reflection of a portion of the original beam of Hght from the surface of the first prism. Hence any variation in the original beam will cause similar and simultaneous variations in both the spectrum and the comparison light^. Measurable changes in the intensity of the light are best obtained by the use of rotating sectors, sometimes called the episcotister (Aubert), or by the annulus, a gelatine wedge impregnated with ivory black^. Reduction of intensity by means of Nicol prisms as in v. Helmholtz' spectrophotometer, may not be free from error, since quite an appreciable amount of polarisa- tion of the light is produced by the prisms used to form the spectrum. Much of the German work has been done by this method, and care has by no means always been taken to calculate the corrections necessary owing to this cause. Pure spectral colours rarely occur in nature, and much of the litera- ture on colour vision is devoted to observations with pigments, coloured glasses and so on. It is necessary, therefore, to say a few words about these complex colours, chiefly with the object of putting the reader upon his guard. When white light passes through a red glass or trans- parent red fluid certain rays are absorbed. The red rays are transmitted 1 Abney, p. 46. - Ibid. Chap. iv. » Ibid. Chap. vi. 6 COLOUR VISION in greatest quantity, so that the dominant colour of the light reaching the eye is red ; but it is not pure red. Most blue substances, such as copper salts, allow the blue rays to pass, but also some of the green and violet, though few of the red and yellow. Yellow substances allow much red and green to pass as well as the yellow, but little blue and violet. The true composition of the transmitted light can only be determined with the spectroscope. The case of pigments is similar. Each speck of powder is a small transparent body which absorbs certain rays of light. When light falls upon such a powder a small portion is reflected from the upper surface ; this is white. The remainder passes deeper and is reflected from deeper layers. The deeper it passes the greater is the absorption and the more intense the colour. Hence a coarse powder appears more intensely coloured than one which is finely divided. Reflection varies with the number of surfaces, not with the thickness of the particles. The larger the latter the deeper the light must penetrate for the same number of surfaces to be met as when the particles are smaller. The absorption is therefore greater in a coarse than in a fine powder. The reflection at the surfaces is diminished when the intervals between the particles are filled with a fluid of refractive index nearer their own than that of air. Hence powders are generally whiter when dry than when mixed with water or oil. The amount of absorption of light by a transparent body can be measured and expressed in the form of a coefficient. If a spectrum is viewed through an orange glass very little red, orange and yellow, but much green and all the blue are absorbed, as shown by dark bands in the regions of absorption. In this case the coefficient of absorption increases as the blue is approached. By knowing the coefficients of absorption of different media the effects of combining them in various ways can be calculated. By empirical experiments colour screens or filters can be made which transmit certain portions only of the spectrum, and in some cases approximately monochromatic light can be obtained in this manner. These filters are much used for photographic purposes. The characters of the absorption by Jena glass filters and by various fluid media are described in Tigerstedt's Handbuch der pkysiologischen Metliodih^, which also gives an excellent resume of the methods which have been employed for the investigation of colour vision. No pigments accurately represent spectral colours, for the reflected 1 Bd. III. Abt. 2, Sinnespkijsiologie ii pp. 47 and 52. THE PHYSICAL BASIS 7 light is always more or less impure. The nearest approach is given by the following list : Red — vermilion (not scarlet vermilion) mixed with a small quantity of permanent violet. Orange — orange cadmium. Yellow — chrome yellow. Green — Prussian blue mixed with aurelin. Blue-green — viridian mixed with a small amount of cobalt blue. Blue — ultramarine. Violet — permanent violet mixed with a small amount of blue, (Abney.) CHAPTER II THE ANATOMICAL BASIS I do not propose to discuss fully the anatomy and physiology of the eye and visual paths, but it is necessary to draw attention to certain features of special importance in colour vision. This course will doubt- less emphasise the great complexity of the subject, which is too often wilfully ignored. The eye resembles a photographic camera, in which the cornea and crystalline lens represent the lens-system, the iris the diaphragm, and the retina the sensitive plate. The size of the pupillary aperture is not under voluntary control, but varies with the intensity of light entering the eye and other causes. This fact has to be taken into consideration in some experiments. (The reader is recommended to read the earlier chapters in the author's Manual of Diseases of the Eye, 2nd edition, J. and A. Churchill, London, 1912.) The optical system of the normal eye at rest is focussed for distant objects, i.e. parallel rays are brought to a focus upon the retina. Focussing for near objects is brought about by automatically altering the " strength " of the crystalline lens (accom- modation), not by altering the length of the eye as in the photographic camera. Some eyes are naturally too short or their optical system at rest is too weak, so that accommodation is required even for distant objects (hypermetropia). Others are too long or their optical system at rest is too strong so that it is impossible to focus distant objects with- out the aid of concave spectacles and near objects may be seen clearly without the aid of accommodation (myopia). Many eyes show slight differences in the radius of curvature of the cornea in meridians at right angles to each other and this is often associated with slight tilting of the lens (regular astigmatism). The lens always shows slight irregularities 8 COLOUR VISION causing irregular astigmatism. For these and other reasons the retinal image of a luminous point is never accurately punctate^. Further, the optical system of the eye is not achromatic. This fact might be regarded as of extreme importance in the consideration of colour vision, but in general the effect is negligible. One would also expect diffraction at the edge of the pupil, and it can be demonstrated to occur, especially when the pupil is small ; but this effect is also negligible under most conditions. There are, however, occasions when both these factors must be taken into account. The crystalline lens normally possesses a slightly amber-yellow hue, which is inappreciable in youth, but increases as age advances. In elderly people the colour of the lens causes an appreciable absorption of the more refrangible rays (green, blue and violet). This fact must be borne in mind in estimating the visual sensations of such people. The retina is a transparent membrane lining the back of the vitreous chamber. It is composed of several layers, the outer of which is a mosaic of rods and cones. The rods and cones are a neuro-epithelium, connected with bipolar cells which form the outer nuclear layer. A second set of cells forms the inner nuclear layer. The arborisations of these two sets of cells form an intermediate outer reticular layer. Arbori- sations from the inner nuclear layer and processes from the ganglion cell layer form the inner reticular layer. The ganglion cell layer is com- posed of larger cells which give rise to the axons of the optic nerve. These axons form a layer upon the inner surface of the ganglion cell layer, the nerve fibre layer. The optic disc or head of the optic nerve is situated about 3 mm. to the inner or nasal side of the posterior pole of the eye. Light falling on this area causes no sensory impression (Mariotte's blind spot). Outside the retina proper is the layer of retinal pigment epithelium. It consists of cells which are hexagonal in section and have processes passing forwards between the rods and cones. The cells contain minute needle-shaped crystals of pigment. This layer bears some resemblance to the " backing " of a photographic plate. External to the retina is the choroid, a highly vascular membrane having the function of nourishing the outer layers of the retina. The inner layers of the retina are nourished by the retinal blood vessels, which spring from the disc and radiate over the surface of the nerve fibre layer. The walls of the retinal blood vessels are transparent, but ^ Parsons, " The Perce^jtion of a Luminous Point," Roy. Lond. Ophth. Hosj). Reports, xvm. 239, 1012 ; xix. 104, 264, 274, 1913-4. THE ANATOMICAL BASIS 9 since the colouring matter of the blood is contained in highly refractile corpuscles the blood column is opaque. The part of the retina to which most attention must be directed for the present purpose is that comprising the rods and cones and the pigment epithelium. It is shown conclusively by Purkinje's experi- ment, which depends upon the position of shadows thrown by the retinal vessels upon the percipient layer of the retina, that the primary seat of the visual impulses is in the layer of rods and cones. Here the most sharply defined image is formed by the optical system. As already stated it is not perfectly defined physically, but as will be seen later physiological and psychological compensation tends to counteract the physical defects. The rods and cones, as their names imply, are minute cylindrical and conical structures. They project vertically — or more accurately radially — from the surface of the outer limiting membrane of the retina. Almost exactly at the posterior pole of the eye is situated a small area in which vision is most distinct. This area is impregnated with a yellow pigment and hence is called the macula lutea or yellow spot. In the centre of the yellow spot there is a conical pit, the fovea centralis, caused by thinning out of the retina. In this minute area the structures are reduced to little more than neuroepithelium and ganglion cells. Moreover the neuroepithelium in this region consists entirely of cones, though the cones are slender and elongated and are more rod-like here than elsewhere^. This change in structure of the cones may be evidence of some physiological combination of the functions of both rods and cones in this situation. Passing peripherally in every direction from the central fovea it is found that rods gradually make their appearance between the cones, and soon the number of rods in a given area becomes greater than that of the cones, so that at the extreme periphery of the retina only a few scattered cones are to be found. The following are some measurements of the diameter of the foveal cones : Max Schultze, 2-8 /m ; H. Miiller, not more than 3 ^ ; Merkel, 3/x ; Welcker, S'l— 3-5;Lt; Wadsworth, 2-5 /x; Kuhnt. 2—2-5^; Kolliker, 4"5 — 5'4ju,; Koster, 4-4: /x; Greeff, 2-5 /x; Dimmer, 3 — 3'5/x. The foveal region is an elliptical area with the long axis horizontal. The long axis measures about 0'3 mm., the vertical 0*2 mm., and the total area is 0"5 — 0*6 sq. mm. Taking the diameter of the inner hmbs ^ Greeff, in Graefe-Saemisch Handb. d. ges. Augenheilkunde, Theil i. Bd. i. Cap. v. 1900. 10 COLOUR VISION of the foveal cones at 0*3 yCt, there are about 300 in the long axis, and 60 in the short, 1300 — 1400 in 0"1 sq. mm. The diameter of the outer limbs is 0'6 — 0'75 ijl. The cones are arranged in curved lines (Max Schultze) or spirals (Fritsch), and are not quite regular. There are small spaces between them, measuring from 0*05 to 0'27 of the trans- verse section of the inner limb. Greeif says that the cones are very closely packed in the fovea, and in a specimen of Heine's were hexagonal in transverse section. Koster^ examined three normal children's eyes and found that the part completely free from rods occupied a circular area 0*44 — 0'55 mm. in diameter, the part relatively free from rods 0*88 mm. In the eye of a youth aged 20, the rod-free area was 0-901 mm. He concludes that in the adult the rod-free area measures about 0*8 mm. in diameter, subtending a visual angle of 3° 3'. This is probably a maximum, and there is physiological evidence to show that the rod-free area varies in size in different individuals. Three areas must be carefully distinguished : Fovea Centralis, measuring 0*24 — 0'3 mm. in diameter, subtending 55'— 70' ; Rod-free area, measuring 0'8 mm., subtending 3° 3' ; Macula, measuring 1 — 3 mm., subtending 4° — 12°. Dimmer- describes a fovea centralis, 1*5 mm. in diameter (the macula of Koster), containing in its centre a foveola (the fovea centralis of Koster). Gullstrand^ regards the yellow colouration of the macula lutea as a post-mortem change, a view which is scarcely consistent with its absorptive capacity for coloured lights during life. Fritsch^ describes the site of clearest vision as the area centralis, possessing a central depression, the fovea centralis, which may or may not contain a foveola. Comparative Anatomy. The distribution of rods and cones in the retinae of lower animals is of great theoretical importance. Many erro- neous statements have gained currency and have been used as arguments in favour of certain theories. The great variety in the forms of the neuroepithelial cells prevents any generalised classification. GreefE^ says that there are rods and cones in the retinae of most 1 Arch.f. Ophth. XLi. 4, 1, 1895 ; Arch. d'OpM. xv 428, 1895 2 Arch. f. Ophth. Lxv. 486, 1907. ■•» Arch.f. Ophth. lxii. 1, .378, 1905; lxvi. 141, 1907. * Ueber Bau u. Bedeututig d. Area centralis des Menschen, Berlin, 1908. ^ Graefe-Saemisch Handb. d ges. Augenheilkunde, Theil i. Bd. i. Cap. v. 1900. THE ANATOMICAL BASIS 11 mammals, amphibia and fishes, the number of rods much exceeding that of cones. In birds, on the other hand, cones are much in excess of rods. In most reptiles (lizards, snakes, tortoises), only cones are found. There are vertebrates possessing only rods, e.g. amongst fishes, rays and dog-fish ; amongst mammals, hedgehog, bat, mole and night-ape {Nyctipithecus felinus). There are also animals of nocturnal habits possessing only rods. Owls, mice, and rell-mice have only a few rudimentary cones ; rats also possess a few cones. Hess^ found rods in fowls and pigeons, though they are few in the posterior and superior parts of the retina, which are most used in pecking. As regards nocturnal birds Schultze (1866), Krause (1894), CTreeft' (1900) and Piper (1905) give contradictory statements. Hess found a "not inconsiderable number " of yellow or brown oil globules in the retinae of the owl and hawk, and these are present only in cones. They are fairly uniformly distributed. Coloured oil globules are found in the cones of birds and reptiles ; similar colourless bodies are found in fishes and amphibia. The globules are more deeply coloured, yellow or brown, in night birds and in tortoises than in day birds. They are absent in the crocodile. Research has shown that at any rate in lower animals stimulation of the retina by light is accompanied by structural, chemical, and electrical changes^. Structural Changes. The chief structural changes are the photo- tropic reaction of the pigment epithelium and the contraction of the cones. To these may be added changes in the Nissl granules of the ganglion cells^. When the frog's eye is exposed to light the pigment granules wander into the cell processes between the rods and cones. This light effect is complete after 5 to 10 minutes' exposure. The retreat of the granules to the complete dark position takes one to two hours. The light effect is limited to the area stimulated, so that an " epithelial optogram " can be produced (Klihne). The violet end of the spectrum is more strongly " retinomotor " than the red end (Angelucci^, Engelmann^), and red light causes little reaction. Light on one eye causes wandering 1 Vergleichende Physiol, d. Gesichtsinnes, Jena, 1912. 2 Garten, in Graefe-Saemisch Handb. d. ges. Augenheilkuiide, Theil i. Brl. iii. Cap. xii. Anhang, 1907-8. 3 Birch-Hirschfeld, Arch. f. Ophlh. L. 166, 1900; LXiii. 1, 8.5, 1906. « Arch. f. Anat. u. Physiol. 353, 1878. '^ Arch J. d. ges. Physiol, xxxv. 498, 1885. 12 COLOUR VISION of the pigment in both (Engelmann). Light on the skin produces the effect in frogs ; so too electrical stimulation of the optic nerve, probably through the centrifugal nerve fibres described by Ramon y CajaF. The phototropic reaction of the pigment epithelium has not yet been proved to occur in mammals^. In the frog's eye in the state of rest (darkness) the cones are extended ; on exposure to light they become contracted. This reaction occurs in all animals that have been examined, including man (van Genderen Stort^) ; but Garten (in Graefe-Saemisch, loc. cit.) found the reaction doubtful in monkeys. The reaction is slow, taking two or more minutes even with intense light. The violet end of the spectrum acts most strongly (Engelmann), but the reaction to the red end is greater than that of the pigment. Light on one eye causes reaction in both, as also light on the skin, so long as the brain is intact (Engelmann, Nahmacher'*). Chemical Changes. Light on the retina causes it to become acid (Angelucci, Lodato^, Dittler^), and its staining reactions are said to alter. The most important chemical change, however, has to do with the visual purple or rhodopsin. This remarkable substance was discovered in the rods of the frog's retina by H. Muller in 1851. Boll in 1876 discovered that it was bleached by exposure to light. Kiihne in 1878 first studied it exhaustively. It occurs in all animals which possess rods, and is present in the rods only (Kiihne). Hence it is absent from the fovea. This statement is denied by Edridge-Green and Devereux Marshall' for monkeys, but their observations have not been confirmed. Hering, however, points out that visual purple may possibly not be wholly absent from the cones, and the peculiar rod-like character of the human foveal cones makes it not improbable that they contain some of the substance. The question merits further investigation. Kiihne investigated the chemical characteristics of rhodopsin, the most noteworthy facts being its solubility in bile acids and their salts, and its resistance to strong oxidising and reducing agents. It is not the cause of the fluorescence of the retina. It can only be seen ophthalmoscopically in fishes which possess a white tapetum (AbelsdorfE^). Tait^ and Boll ^ Die Retina der Wirbeltiere, Wiesbaden, 1894. 2 Hess, Vergleichende Physiol, d. Gemchtsinnes, Jena, 1912. 3 Arch. f. Ophth. xxxni. 3, 229, 1887. « Arch. f. d. ges. Physiol, liii. 375, 1893. 5 Arch, di Oft. ix. 267, 1902. * Arch. f. d. ges. Physiol, cxx. 44, 1907. ' Trans. Ophth. Sac. xxn. 300, 1902. ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xiv. 77, 1897 ; Sitz. d. Berliner Akad. 1895. 9 Proc. R. S. ofEdin. vii. 605, 1869, THE ANATOMICAL BASIS 13 state that it can be seen entoptically on waking in the morning as a rose-red ring round the fixation point, projected against the white ceiling. Haab attributes this phenomenon, not to the visual purple, but to the pigment of the yellow spot. Edridge-Green^ confirms Tait's and Boll's observations, and states that the purple can be seen to flow in waves from the periphery of the yellow spot towards the point of fixation. The colour of the visual purple varies in different animals and under different circumstances^, and gives different spectroscopic absorption bands (Kiihne ; Kottgen and Abelsdorff^). The maximum absorption in fishes is in the yellow (540 ft/u,), in mammals in the blue-green (500 M/i) ; hence it is reddish violet in fishes and purple in mammals. The occur- rence of " visual yellow " as an intermediate stage in the bleaching of visual purple (Kiihne) seems to have been conclusively disproved by Kottgen and Abelsdorff. They found that as the visual purple was bleached the relative absorption remained unchanged. The substance therefore becomes gradually less concentrated, without passing through a yellow stage. The bleaching of visual purple is limited to the area exposed to light, so that an optogram or image of the luminous object, such as a window, can be obtained. Such an optogram can be partially preserved by alum solution, somewhat as a photographic negative is fixed, though the processes are entirely dissimilar. Two to seven minutes' exposure to light suffice to obtain a good optogram in the frog's retina. Light on one eye does not cause any bleaching of the visual purple of the other. The bleaching of visual purple by monochromatic light has proved to be of great theoretical interest. Observations have been made by Konig^ Kottgen and Abelsdorff, and Trendelenburg^. Trendelenburg took two specimens of frog's visual purple and exposed one to Hght of the sodium line (D, 589 /x/x) and the other to another Hght from the same dispersion spectrum, the diminution of absorption being measured by the spectrophotometer. The human achromatic scotopic luminosity 1 J. of Physiol. XLi. 263, 1910 ; xlii 428, 1!)11 : xlv. 70, 1913. '^ Garten, in Graefe-Saemisch Handb. d. ges. Augenheilkuride, Teil i. Bd. ni. Kap. xii, Plate VII, 1908. ' Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xn. 161, 1896. * Konig, p. 388. '" Centralhl. f. Physiol, xvii. 1904 ; Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxxvn. 1, 1904. 14 COLOUR VISION 40 3-0 2-0 10 00 ' 1 — ■ — -^^V ^^-^^ 1 ^^^ - ^ 'n^^- t ^^- J^ \^ t \ I 55^ f _S^ t - SS.^ _r - ^^N T _ ^s"!^^ ^^^N X; 1 __ 1 — 1 — L_--_J — , — 1 — ^1 — 1 — ^1 — L OS Ol CJi cn 03 o O CD CD ^ Fig. 1. • • — the bleaching values of spectral lights for frog's visual purple, and X — X — X the human achromatic scotopic luminosity curve. Abscissae, wave- lengths of the prismatic spectrum of the Nernst light ; ordinates, an arbitrary scale. (Trendelenburg.) curve will be referred to later (pp. 53, 55). Actual readings are as follows : Wave-length 589 542 530 519 509 491 474 459 Bleaching value 1 3-40 3-62 3-45 3-09 1-69 0-975 0-299 Scotopic luminosity 1 3-62 3-91 3-19 2-67 1-42 0-621 0-346 Visual purple is regenerated after bleaching in the living and " surviv- ing " retina, but only on contact with the pigment epithelium. Regene- ration commences in the frog after 29 minutes and is complete in one to two hours. The process is more rapid in the rabbit, commencing in seven minutes and being complete in 33 to 38 minutes. (It may be noted here for convenience that increase of sensibility in dark adaptation in man after good light adaptation commences in 7 to 8 minutes and reaches nearly the maximum in 40 to 45 minutes (Piper^).) It has already been stated that the visual purple occurs in nearly if not all animals possessing rods from petromyzon to man. It is present in albinotic animals, e.g. rabbit, and has been found in the 9- and 7- months' human foetus (Klihne, Fuchs and Welponer'^). Kiihne said that it was absent in the bat {Rliinoloj^lius hijjposiderus), but this statement has been disproved by Trendelenburg. Kiihne found no visual purple ^ Ztsch. f. Psychol, m. Physiol, d. Siimesorg. xxxi. 161, 1903> 2 Wiener med. Woch. 221, 1877. THE ANATOMICAL BASIS 15 in fowls and pigeons, but Boll, Angelucci, and van Genderen Stort found it in pigeons. Hess dissolved out the oil globules with benzol, which does not affect the purple in frog's retina, and found traces of rhodopsin in both fowls and pigeons, but in far less amount than in man, ox, frog, etc. There was more present in the hawk, buzzard, goose and duck. Amongst reptiles the crocodile possesses many rods and is rich in visual purple. Electrical Changes. The retina, connected through non-polarisable electrodes with a galvanometer or capillary electrometer, shows a " current of rest " (du Bois-Reymond). On exposure to light there is usually a short negative variation followed by a longer positive variation, ooos — ^ /^ imn^ / "^"^- • 40I Utiu / otr / ^^ 000? 0001 luea. ] r \ ^1 9 r 2' »■ ♦• r ■ 6' 7'. «• »■ 1 0^' ± r- 0 I" *■ 3' •« y 0' r ? r Kf Fig. 2. Electrical changes in the frog's eye caused by light. Plotted curves from the analyses of three capillary electrometer records. The illumination in the case of the upper record was white light, in that of the middle red light, and in that of the lowest violet light. Abscissae, time after the commencement of the illumination in seconds ; ordinates, the electromotive force in ten thousandths of a volt. (Gotch.) VoU •0007 0 01 02 &3 &4 &5 06 07 08 09 10 II l« Fig. 3. Curves constructed from typical electrometer records of eyeball responses to the light from the red, green, and violet regions of the interference spectrum of the arc light. (Gotch.) 16 COLOUR VISION and cessation of the light stimulus causes a further positive variation (Holmgren), a fact of considerable theoretical interest. The electrical changes have been studied by Dewar and McKendrick^, Klihne and Steiner^, Fuchs^, Waller^ Himstedt and NageP, and others, but Gotch's experiments with the capillary electrometer are the most conclusive^. He found that spectral red light gave a latent period of nearly -fn second and a difference of potential of 0*0004 volt ; green, -fy" and 0-0005 volt; violet, ^" and 0-00024 volt. Himstedt and Nagel obtained a slight reaction with ultra-violet and Rontgen rays. They also found that in the dark-adapted eye the maximum effect was obtained at 544jLt/x, the site of maximum luminosity in dark adaptation in man (vide infra) ; in the light-adapted eye at about the D line (589 /x/x), the site of maximum luminosity in light adaptation ; but though the light sensibility of the dark-adapted eye is more than a thousandfold that of the light-adapted for some colours, CI. E. Mliller^ obtained no appreciable difference in the electrical reaction. Engelmann^ found that stimulation of one eye caused a reaction in the other also, but the positive variation on removal of the stimulus was absent. The relation of intensity of stimulus to strength of response does not accurately follow the Weber-Fechner law {vide infra, p. 20), but Talbot's law (v. p. 92) is followed more accurately (de Haas^). When an object is looked at directly a sharp image is formed on the fovea and the immediately surrounding area. An object therefore which subtends less than 3° at the nodal point of the eye will form its image entirely upon the rod-free area of the retina. Larger objects subtending 4° — 12° will form their images on the macular region, in which only a few rods are present in the peripheral parts. Objects surrounding that fixated form images on the peripheral regions of the retina which are richly supplied with rods. The acuteness of form vision falls off rapidly in passing from the fovea to the periphery, but movements of objects having their images in the periphery are very readily observed. Not only is form vision different according as the image is at the 1 Trans. R. S. Edin. xxvii. 141, 1873. 2 Heidelb. Unters. in. 327, 1880. ^ ^^j-ch. f. d. ges. Physiol. LVi. 408, 1894. * Phil. Trans. Roy. Soc. Lond. cxciii. 123, 1900. * A7m. d. Physik, iv. 1901 ; Ber. d. Naturf. Ges. Freiburg, 1901. « J. of Physiol. XXIX. 388, 1903 ; xxx. 10 ; xxxi. 1, 1904. ' Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xiv. 329, 1897. 8 Helmholtz' Festschrift, 197, 1891. " Inaug. Dissert., Leiden, 1903; ref. in Nagel's Jahresbericht f. Ophth., 73. 1903. THE ANATOMICAL BASIS 17 fovea or peripheral, colour vision is also affected. Foveal or central vision must therefore be clearly distinguished from peripheral or eccentric vision. Vision is profoundly affected by the condition of the retina at the moment of stimulation. The condition of the retina at any given part is determined by two factors, temporal and spatial. The temporal factor is the nature of the stimulation to which the retina has been previously submitted (temporal induction). If the eye has been exposed to bright light it is said to be light-adapted. I shall speak of vision under these circumstances as photopia, and the light-adapted eye as a photopic eye. If the eye has been kept completely free from light for a considerable period it is said to be dark-adapted. I shall speak of vision under these circumstances as scotopia (as an equivalent for the German Ddmmerungssehen), and the dark-adapted eye as a scotopic eye. The spatial factor in retinal sensibility is the reciprocal action of different areas upon each other. The excitability of a given area is affected by the condition of sensibility and stimulation of the surround- ing areas (spatial induction). CHAPTER III THE PSYCHOLOGICAL BASIS The physiology of the senses may be regarded as the intermediate link between the outer world and the consciousness of the individual. The physical stimuli come into contact with the sense organs and set up sensory neural physiological changes, which may be transformed into, or at all events accompanied by, mental processes leading to sensation, perception, and comparison. In this long and complicated path there are two points of contact, the physico-physiological and the physiologico-psychological. Of the former we know a Httle, and the elements of our knowledge have been discussed in the last chapter. Of the latter we know nothing, but of the final perfected perceptions we have some knowledge, the elements of which we must now discuss. The awakened perceptions, their qualities and attributes, and their inter-relations belong to the domain of psychology. They depend in some occult manner upon the sensations aroused by the physiological processes which are set in activity by the physical stimuli. We have good reason to think that there are many gradients and junctions upon p. o. \ . 2 18 COLOUR VISION the physiological path, so that the resultant nervous activities set up in the higher centres of the brain are very different from those which started in the peripheral mechanism. Try as we may to dissociate the physiological from the attendant psychological phenomena we are unable to do so, for the act of analysis itself is a psychological process. We are practically limited to the consideration of the relationships which can be made out with greater or less certainty to exist between the stimuli and the resulting sensations. With regard to these relationships there are certain generalisations which have been thought to hold good for all the so-called " senses." One of these generalisations, enunciated in 1785 by Bonnet and independently by Johannes Midler in 1826^, is called Miiller's Law of the Specific Energies of the Senses. It is to be noted that the word energy is not used here in its ordinary physical sense. Miiller stated that we can experience no kinds of sensation through the inter- mediation of external causes which we cannot also experience through the sensation derived from the condition existing in the nervous organs without external agency. These internal causes call up different sensations in the different senses according to the nature of each sense. Further, the same external stimuli arouse in the different sense mechanisms different sensations according to the particular sense ; and different external or internal stimuli, acting upon the same sense mechanism, ahvays arouse the same sensation. The essential features of Miiller's law are contained in the sentences in italics. He amplified the law by saying that a sensation is not the transference of a quality or condition of external bodies to the consciousness, but the transference of a quality or condition of a sensory mechanism to the consciousness, though that quality or condition is occasioned by external stimuli. These qualities or " energies " of the senses are specific to each sensory mechanism. Miiller left it open as fo the site of the specific energies, whether in the brain or elsewhere in the nervous paths, but he considered it certain that the central endings of the sensory nerve-paths in the brain were capable of arousing the specific sensations independently of the conducting paths. It is now generally held that the seat of the specific sensations is in the brain. If this assumption be true Miiller's law relieves us of the necessity of predicating the con- duction of different kinds of impulses by the peripheral nerve-paths. It is customary to divide the stimuli which can arouse a specific ^ Zur vergleichendcn Physiologlt dcs Gesichtssinnes, Leipzig, 1826 ; Ha/idb. d. Physiol, d. Menschen, Coblenz, 1840. THE PSYCHOLOGICAL BASIS 19 sensation into adequate and inadequate. The former are those which arouse the sensation under ordinary circumstances, the latter are all other effectual stimuli. Thus light is an adequate stimulus for the eye, pressure on the eyeball an inadequate. The next great step in the advance of our knowledge of the relation- ships subsisting between physical stimuli and sensations was made by E. H. Weber, the founder of modern psycho-physical methods. In general terms it may be stated that a stimulus must attain a certain intensity in order to excite a sensation and that stimuli of greater intensity excite stronger sensations. There is therefore a quantitative relationship between the stimulus and the sensation. The minimum effectual intensity of stimulus is called the general threshold or the general liminal value. A higher value may arouse a sensation differing in quality from the other ; this value is called the specific threshold or the specific liminal value. Thus a coloured light of low intensity may excite a colourless sensation ; when of a higher intensity it may excite a sensation of colour. The intensity of the stimulus has to be increased by a definite amount before a difference in the amount of the sensation becomes appreciable. This amount may be called the differential threshold or the liminal discrimination value. Equal increases of physical intensity do not give equal increases of sensation. There are three chief psycho-physical methods whereby estimates can be obtained. (1) The method of least perceptible differences or the limiting method : one of two identical stimuli is regularly increased by small increments until a just perceptible difference between the constant and the variable stimulus is experienced. (2) The method of right and wrong answers or the constant method : the observer is asked which is the greater in the case of such pairs of stimuli (constant and variable), the variables here being fewer in number and presented with 'the constant in irregular order, and the average of a large number of such estimates being taken. (3) The method of mean error or the method of production : the observer picks out what he regards as just appreciably different stimuli ; the sum of the differences from the standard divided by the number of trials gives a lower limit of the threshold value. These results are accurate only when submitted to strict statistical processes^. Weber stated his conclusions in the form of a law : The just appreci- able increase of stimulus hears a constant ratio to the original stimulus ; i.e. two stimuli in order to be discriminated must be in a constant ratio, ^ See Myers, Text-hook oj Exp Psychology, Chaps, x. and xv. London, 1911. 2—2 20 COLOUR VISION which is independent of the absolute magnitudes of the stimuh. For white hght Fechner could distinguish a difference of jljy, v. Helmholtz y^y, of the light intensity. Fechner went further and attempted to express sensations in terms of quantitative units. His most important assumption was that all just noticeable differences of sensation contain an equal number of sensation-units. Fechner's law states that the sensation varies as the logarithm of the stimulus ; i.e. the sensation changes in arithmetical proportion as the stimulus increases in geometrical proportion. Stated algebraically, if E is the measure of a sensation and hE the just appreciable difference, S the measure of the stimulus and hS a small increment, then 8^=0^ (Weber's Law) where C is a constant : therefore, on the questionable assumption that it is permissible to integrate BiwaW finite quantities {hE, etc.) = C log S + C (Fechner's Law) where C" is another constant. Weber's law does not hold good for very low or very high intensities of stimuli, and is only approximate at the best. An immense super- structure has been built up upon these psycho-physical foundations^. The bases are insecure on mathematical as well as on physiological grounds. So far as the latter are concerned we have no unit of sensation (cf. p. 61), and the variations, though quantitative, are only relative. The chief difficulty, however, is to be found in the ever-changing condition of the sensory apparatus. The deductions are not without value, for some quantitative relationship certainly exists, even if it be not so simple as Fechner's law implies. The problem is still more complicated when we come to comparison of specific qualities in sensations. Thus we recognise brightness as a quality of coloured lights and we may say with some degree of certainty that a given red light has the same brightness or luminosity as a given blue light. We can thus estimate qualitatively equivalent stimuli and it is also true that we can attempt to estimate qualitatively equivalent ^ Lipps, Grundriss d. Psychophysik, Leipzig, 1903 ; G. E. Miiller, Ergeb. d. Physiol. II. 2, 267, 1903 ; v. Kries„ in Nagel's Ha7idb. d. Physiol d. Menschen, m. 16, 1904. THE PSYCHOLOGICAL BASIS 21 discrimination values. The question therefore arises whether sensations can be divided up into elementary qualitative parts. The simplest theory of such psychological analysis is that elaborated by Mach and others that each psychological element has a physiological counterpart, which is itself the expression of a physical counterpart. Each is indeed an accurate image of the other, or, to use Fechner's simile, the psychical and the physical are the concave and convex sides of the same curve. V. Kries, McDougall, and many others are of the opinion that this view is untenable. If two perceptions differ, yet possess a certain similar quality, that quality may be regarded either as made up of the sum of similar preformed parts in the various constituents of the two percep- tions, or as a totally new psychological rearrangement of the under- lying factors. Of these alternatives the latter is the more likely, and we must at present remain content to regard certain psychological similarities as not capable of analysis. The psychological analysis of our sensations brings out other funda- mental facts of importance. We not only see a light or a colour, but we see it at a definite time and in a definite place. Leaving aside the temporal element for the moment, we find that our orientation in space is largely dependent on vision. Hence it arises that our visual impres- sions are projected outwards to definite positions in the outer world. In this respect visual " sensations " differ from such sensations as pain, heat, cold, etc.^ This unconscious projection of impressions is responsible for the fact that we associate our sensations with certain properties of external objects. We speak of objects as being round, bright, red, and so on, — an inaccuracy which is responsible for much confusion. A luminous object sets the ether in vibration ; when these vibrations stimulate the retina they give rise to sensations, which we describe as bright, red, and so on. These qualities are therefore subjective and must be strictly dissociated from the physical stimuli which give rise to them. This aspect of the subject has been lucidly treated by Hering^. As he says, " our visual world {Sehivelt) consists essentially of dift'erently presented colours, and objects, as seen, that is visual objects {Sehdinge), are nothing but colours of different nature and form^." The whole of 1 Cf. Hering, Grundziige der Lehre vom Lichtsinn, in Graefe-Saemisch Handb. Th. i Bd. m. Kap. xii. 1905. - Loc. cit. ^ " The eye sees no form, inasmuch as light, shade, and colour together constitute that which to our vision distinguishes object from object, and the parts of an object from each other." Goethe (1810). " AU vision is colour vision, for it is only by observing difference^! of colour that we distinguish the forms of objects." Clerk-Maxwell (1871). 22 COLOUR VISION the nervous apparatus of vision constitutes an " inner eye," which builds up a new visual world under the compulsion of the stimuli derived from the real objects of the outer world. Things seen, visual objects, or colour forms must therefore be clearly distinguished from the real objects. The untutored regard the green of a leaf as an attribute of the leaf. The physicist, however, knows that colour depends upon the light reflected from the leaf and calls the reflected light green. The physiologist knows that the leaf, which appears green when looked at directly may appear yellow or grey when its image falls upon the peripheral part of the retina. He is therefore inclined to regard the colour as an attribute of the eye itself. Finally, to the psychologist the green is neither an attribute of the leaf, nor of the light, nor of the eye, but a psychical phenomenon, a definite qualitative entity in con- sciousness. Colours therefore are visual qualities, and we are only justified in speaking of red or green objects, red or green rays, and so on in the broad sense that the objects or rays appear red or green respectively under the ordinary conditions of vision. Brightness and darkness, again, are not attributes of the objects or light rays, but of the colours as visual qualities. White, grey, and black must be included amongst the colours as visual qualities, but may be distinguished from the variegated or toned colours as untoned colours. The common attribution of colours to the objects themselves, thus implying that the colours are properties of the objects, is largely a matter of memory. We say that snow is white, soot black, blood red, because under the ordinary conditions of life these objects appear to be of those hues. In this sense the colours may be well termed " memory colours " {Geddchtnisfarheyi, Hering^). The appearance of a given object at a given moment is by no means determined solely by the nature and intensity of the rays falling upon the eye and the condition of the nervous appara- tus of vision at the time. These are but the primary and fundamental exciting factors. They awaken unconscious reproductions or memories of bygone experiences, which act as secondary but potent factors in the subconscious sphere, modifying and in many cases determining the ultimate conception. Thus it is that, though we are firmly convinced that snow is white and blood is red, the pink glow of a snow-clad mountain and the pallid hue of a face seen by the light of the mercury arc are regarded as accidental colours in no wise modifying our impressions of the actual colours of the objects. Every object with which we are > Cf. Katz, Centralhl. /. Physiol., xx. 1906. THE PSYCHOLOGICAL BASIS 23 familiar awakens a memory picture in our minds : "we see it through memory-coloured spectacles." Thus we often see it quite different from what it is, and our capacity to dissociate accidental colours from the so-called real colours of objects is very highly developed. Thus, the shadows on the surface of a body, which largely influence our perception of its shape, relief and distance, we instinctively regard as an epiphe- nomenon, and we think that we see the actual or real colours through the darkness of the shadows. A shadow on white paper appears quite different to us from a grey spot on the paper, even though both reflect exactly the same amount of light. Similarly a patch of cigarette ash on a black coat conveys a different impression from a patch of bright sunlight. Indeed, the difference manifests itself in words ; the one we usually call white or grey or black, the other bright or dark. Hering has devised some simple, but very instructive experiments to illustrate these facts. One example will suffice. In a room lighted by a window on one side, the opposite wall being white, standing with the back to the window and holding up a grey sheet of paper, the paper looks grey and the wall white. The wall however reflects only a portion of the light into the eye and " is " therefore grey. By looking through a tube it is possible to select a grey paper which exactly matches the greyness of the wall ; yet directly the tube is removed the wall at once appears white, whilst the paper still remains grey. If, however, the edge of the paper is fixed with one eye only, the wall appears to be on the same plane and of the same tint as the paper. In this case different localisation produces diff'erent colour impressions, and the ultimate perceptions depend, not upon dift'erences of physical light intensity, but upon other impressions which simultaneously enter into conscious- ness and modify judgment. One does not generally pay special attention to the colours of objects, but uses them merely as indicators, specially associated with the objects; hence when the object is seen again the colour impression is immediately revived. Some dresses look blue by daylight, bluish-green by electric light, and the wearers often think it strange when in artificial light people say that they are bluish-green. Such people may even correct themselves and say that they are certainly blue, when they are told that they are blue. Hering points out that these facts have nothing to do with simul- taneous contrast {v. Section VI, Chap, ii) as has sometimes been thought. They are indeed examples of the association of ideas or sympsychosis. They show, however, the necessity for eliminating as far as possible all 24 COLOUR VISION subsidiary impressions, such as localisation, shape, dimensions, etc., when comparing colours. Further, it is easy to show that the apparent brightness and colour of objects can be altered within a wide range without disabusing our minds of the opinion that the colours are inherent properties of the objects. Thus, the paper of a book appears white and the print black, whether we read it in the morning or at mid- day or in the evening. Yet Hering has shown by accurate measurements that the print may actually reflect more light at mid-day than the paper did in the morning. Similarly the paper remains " white " and the print " black " whether the book be read by daylight or gas light or electric light or in the shadow of green trees. " The approximate constancy of the colours of visual objects in spite of gross quantitative and qualitative variations of the general illumination of the visual field is one of the most remarkable and weightiest facts in the domain of physiological optics " (Hering). We shall see that these gross variations are compensated for by processes of physiological adaptation of the visual nervous structures as a whole and that stimulation of retinal areas by light arouses reciprocal activities in neighbouring areas. These complex processes should deter us from drawing too dogmatic conclusions from the psychological analysis of colour sensations. As McDougall^ says, mental activity consists in the process of establishing in the mind relations between one thing and another. This process in its best-defined form is apperception, " the process by which a mental system appropriates a new element or otherwise receives a fresh determination^." Each mental system is gradually built up by a series of apperceptive processes, each such process con- sisting in the presentation of some one aspect or feature of the whole object through some sense-organ, and the bringing of this feature into mental relation with various other aspects and features previously apperceived and incorporated into the mental system. " In almost every moment of waking life an apperceptive process is taking place ; whenever an object is attended to the presentation of it is apperceived^." Mental activity then consists essentially in the perpetual succession of apperceptive processes, and the essence of apperception is the appro- priation of the relatively novel presentation by the mental system built up by previous apperceptions. At each apperception of any given presentation of an object the appropriation of it by the mental system is more ready and more complete, while the consciousness excited by 1 Brain, xxiv. 605, 1901. ^ stout. Analytic Psychology, 1896. * Stout, loc. cil. II. p. 113. THE PSYCHOLOGICAL BASIS 25 this aspect of the object becomes less and less vivid, until finally, when the appropriation of it by the mental system becomes complete, it is implicitly apprehended, or, in terms of conation, the stimulus applied by this aspect of the object is responded to automatically, while some other aspect occupies the focus of consciousness. If, however, bearing in mind the underlying complex factors, we attempt to make a psychological analysis of visual qualities it will be generally agreed that they can be divided into two groups of colour sensations, the untoned and the toned. The untoned or colourless form a continuous series from the blackest black through all gradations of grey to the whitest white. The toned or coloured include four, red, yellow, green and blue, together with all the gradations between them. So far as the insecure foundations of psychological analysis go red, yellow, green, and blue are simple or pure visual qualities. All other hues are psychologically mixtures of these qualities. Thus, orange obviously partakes of the nature of both red and yellow, purple of both red and blue, and so on. Yet that psychological analysis is necessary is shown by the acceptance of green as a simple colour sensation ; for without analysis most people would say that green is a mixture of yellow and blue. That such is the case is doubtless due to familiarity with the behaviour of pigments, yet it may have a deeper significance, since the distinction between green and blue is vague amongst many primitive races and frequently amongst the uneducated classes. Moreover, there are difficulties associated with black and white and two toned colours, brown and olive-green. " White " is particularly variable, chiefly owing to complexities arising from adaptation. Hering regards difiuse sunlight as distinctly yellow, and the " whites " of sun- light, arc light, incandescent gas light and so on, even when reflected from a surface of compressed magnesium oxide, show gross variations which are not submissible to psychological analysis. Opinion differs as to whether " black " is the negation of all sensation, as generally accepted by physicists, or an active sensation, as accepted by many psychologists. It is certain that there is a blacker blackness than that experienced when the eyes are carefully shaded from the light. Simi- larly the sensation of brown cannot be elicited by merely reducing the intensity of a yellowish-red or any other spectral light or mixture of spectral lights, and the same applies to olive-green. Under these circumstances the spectral colours approximate more and more nearly to black. In order that a brown sensation may be experienced the stimulus effect of the yellowish-red light must be " blackened " by 26 COLOUR VISION simultaneous or successive contrast, or the appropriate pigment must be mixed with black. We shall have to discuss these anomalies more fully in the sequel, but enough has been said to show that psychological analysis can afford no infallible criterion. SECTION TI THE SPECTRUM AS SEEN BY THE LIGHT-ADAPTED (PHOTOPIC) EYE CHAPTER I THE SPECTRUM: HUE, LUMINOSITY, SATURATION If a pure spectrum, e.g. that of the arc light, of moderate intensity is observed a band of colours is seen. Of these, four are clearly defined as separate and distinct from each other, viz. red, yellow, green and blue, the red region consisting of the least refracted rays, the blue of the most refracted. Between the red and yellow we distinguish a region which is called orange. The gradation from red to yellow is gradual and it will be generally admitted that orange partakes of the natures of both red and yellow psychologically, the red element diminishing as we pass from red to yellow and the yellow element correspondingly increasing. Between yellow and green a somewhat similar gradation occurs, the yellow gradually becoming more and more tinged with green until we fail to recognise any yellow at all and the colour gives the impression of pure green. Passing further towards the blue an inter- mediate green-blue region is met with, showing the same gradual transition until the blue no longer gives any impression of green. Pass- ing beyond the blue we gradually come to a region in which the pre- dominant sensation is still of the order " blue," but it is not pure blue. It is called violet. Now violet is a colour which occurs rarely in nature. There is, however, a colour in nature which is often called violet, but which is really purple. True purple does not occur in the spectrum, but it can be obtained by mixing pure red light with pure blue light, and we can pass from blue to red through violet and the mixtures of blue and red which are called purple and carmine. We have thus travelled in a circle and returned to the original 28 COLOUR VISION starting place, red. This is a very important fact, for it can be proved that with the help of the colours thus obtained, either pure or mixed with each other or with black in various proportions, all known colours and tints can be reproduced. We can map out the spectrum into its separate colours, using the Fraunhofer lines as convenient fixed points, but as the colours pass gradually into each other the limits are more or less arbitrary, v. Helmholtz gives the following names to the different regions of the spectrum : ine Wave length in ixn Colour A 760-40 Extreme red. B 686-853 Red. C 656-314 Junction of red and orange. D f589-625 \ 589 -024 Golden yellow. E 526-990 Green. F 486-164 Cyan blue. G 430-825 Jiniction of indigo blue and violet H 396-879 Limit of violet. Speaking generally then, change of wave-length causes a change in colour, or in the hue or tone of a colour. The tone changes most rapidly on both sides of the yellow, most slowly near the ends of the spectrum. For a certain distance at each end change of wave-length is no longer accompanied by change of tone, at the red end beyond 655 /x/u, at the violet end beyond 430 ixfi. Apart from the change in colour the most striking feature of the spectrum is the difference in brightness or luminosity of different parts. The brightest part is in the yellow at about the D line, the luminosity diminishing continuously on both sides to the extreme ends. The brightness varies with the intensity of the light, but if the intensity is increased beyond a certain point the colours also change in tone. The colours on each side of about 500 fx/n behave differently ; the red, orange, yellow and green approximate to yellow, the blue-green, blue and violet approximate to blue. Though the brightness of the colours increases with the intensity of the light it does not follow the curve of energy of the spectrum. Whereas the brightest part of the spectrum is in the yellow the curve of energy rises continuously from the violet to the red end^. ^ Nichols, Phys.Rev. xxi. 147, 1905; Knirmp, Physisch-ophthalmologische Grenzprobleme. Leipzig, 1906, p. 4. THE SPECTRUM: HUE, LUMINOSITY, SATURATION 29 If now the pure spectral colours be successively mixed with gradually increasing quantities of white light they become paler until eventually no colour can be distinguished. They are said to become less saturated. A given colour may therefore be defined by its hue, its luminosity, and its degree of saturation. In regard to the hue or tone the matter is relatively simple, so long as we adhere to the term hue or tone in this definite sense. The terms tint, nuance, shade and so on should be avoided. With regard to luminosity and saturation the matter is by no means so simple. An unsaturated colour is also an impure colour in the physical sense of the word, for it no longer consists solely of rays from a single small region of the spectrum. But we are confronted with another fact, less easy of explanation, viz. that great increase of in- tensity of the light not only alters the hue, but also alters the saturation, so that eventually it produces only the sensation of white light. It would seem therefore that luminosity is in some recondite sense an inherent " whiteness " in the colour itself, differing in degree in different spectral colours and varying with the intensity of those colours. Clearly we are here, at the outset, face to face with a physiological fact of immense importance, and much of the difficulty of colour vision is concerned with this fact. The terms hue or tone {Farbenton), brightness or luminosity {Hellig- keit), and saturation or purity {Sdttigung) are now generally used in the well-defined senses given above. One has to be careful, however, in reading the older and some modern works. Thus Aubert^ uses Farben- ton for Clerk-Maxwell's " hue " (" one may be more blue or more red than the other, that is, they may differ in hue ") ; Farbenniiance for Helmholtz's " Sdttigungsgrad " and Grassmann's Intensitdt des bei- gemischten Weiss and Clerk-Maxwell's " tint " (" one may be more or less decided in its colour ; it may vary from purity on the one hand to neutrality on the other. This is sometimes expressed by saying that they may dift'er in tint") ; Farbenmtensitdt for Helmholtz's LicktstdrJce and Clerk-Maxwell's " shade " (" one may be lighter or darker than the other ; that is, the tints may differ in shade "). Edridge-Green uses the terms hue, luminosity and purity. Hue is employed in the usual sense. Of luminosity, however, he says : " No coloured object can have the luminosity of a white object reflecting practically the whole of the light impinging upon it. Therefore if we take absolute reflection as 100, a fraction of 100 will give the relative luminosity of any body." 1 Physiologic der Netzhaut, Breslau, 1865. 30 COLOUR VISION Purity is " the freedom of the colour from admixture with white light," but he says " when I speak of a colour being mixed with white light, I have a different meaning from that which is signified by most writers on colour." His explanation should be read carefully^, but he appears to use the term " white," which denotes a physiological sensation, in the sense of a physical property of the light. CHAPTER II THE DISCRIMINATION OF HUE IN THE SPECTRUM Observations on the discrimination of hue in the spectrum have been made by Mandelstamm^, Dobrowolski^, Peirce*, Konig and Dieterici^ UhthofE^, Brodhun', F. Exner^, Steindler^ Edridge-Green" and others. Of these, the most accurate and complete are those of Steindler. The maximum discrimination sensibility for hue, i.e. the smallest difference in wave-length (8x) which gives rise to appreciable difference in colour-tone, occurs in the yellow and the blue-green. It has already been remarked (p. 28) that there are regions at each end of the spectrum in which differences of hue are no longer appreciated. Two neighbouring wave-lengths in these regions may appear to differ slightly in hue, but only owing to differences of intensity. When the intensities are suitably modified the differences disappear. Konig and Dieterici found three maxima, I at 440 /it/', II at 490 /x//, and III at 570 /^/x (Dieterici) or 590 fxu (Konig). Of these the greatest discrimination sensibility was II. The minima were at 450 /u./z between I and II, and at 540 /xju between II and III. These results were obtained by the method of mean error (i'. p. 19). (Fig. 4.) 1 Colour-blindness and Colour PcrccjjfAon, 2nd ed. London, 1909, p. 60. ' Arch. f. Ophth. XIII. 2, 399, 1867. » Arch. f. Ophth. xviii. 1, 66, 1872. * Amer. J. of Sc. xxvi. 299, 1883. 5 A7171. d. Physik, XXII. 579, 1884; Arch. f. Opihth. xxx. 2, 158, 1884; Konig, Ztsch. f. Psychol, u. Physiol, d. Sinncsorg. viii 375, 1895 : in Konig. pp. 23, 105, 367. « Arch.f Ophth. xxxiv. 4, 1, 1888. ' Ztsch. f. Psychol, u. Physiol, d. Sinncsorg. in. 89, 1892. * Sitz. d. Wiener Akad. cxi. ii a, 857, 1902 * Sitz. d. Wiener Akad. cxv. ii a, 115, 1906. i" Vide infra. Part iii THE DISCRIMINATION OF HUE IN THE SPECTRUM 31 Fig. 4. Curves ot discrimination sensibility for hues of the spectrum of gas light. A, A, A, for normal trichromat ; B, B, B, for a deuteranope. Abscissae, wave-lengths. (Konig.) Uhthoff used the method of least perceptible differences {v. p. 19), and obtained the following results. X (550 640 630 620 610 600 590 580 570 550 530 510 o\ 4-70 2-97 1-68 1-24 1-08 102 0-91 0-88 1-10 166 1-88 1-29 \ 490 480 470 460 450 5\ 0-72 0-95 1-57 1-95 2-15 His maxima are therefore II at 490 /x/x, and III at 580 /xju,. V. Kries^ points out that Konig's and UhthofE's methods are not strictly comparable, since the former measures the general discrimina- tion sensibility rather than the true specific or hue discrimination sensibility, which is measured by the latter. Steindler used the interference spectrum of the arc-light, and examined twelve persons with normal colour vision. She obtained four maxima for her own right eye (Fig. 5) : I at 435 /x/x ; II at 497 /x/x ; III at 585 /x/x ; IV at 636 /x/x. The corresponding minima were between I and II at 454 /x/x ; II and III, 535 /x/x ; III and IV, 624 /x/x. Comparing different individuals, II or III may show the greater discrimination sensibility. Thus her own III was greater than II whereas we have seen that Konig's and Dieterici's II was greater than III. The exact positions vary — for I, 15 — 20 /x/x; for II, 30 /x/x; for III, 15/i/x; for IV, 15 — 20 /x/x. The average positions were — II, 492 /x/x; 1 Nagel's Huudh. d. Physiol, d. Mcnschen, ni. p. 252. 32 COLOUR VISION III, bSlfjifx; IV, 635-5 /xju. : and of the corresponding minima — between I and II, 458 /jifx ; II and III, 533 /x/x ; III and IV, 627 |U./x. / 60 \ 4V \ V ^ a/ 20 I 1 IT 1 1 ) f 1 ^ 573 -'J 99 482-1 1-190 ., . . 567 1 Indigo- blue 464-5 1-221 J5 • • . . 5$4-4 99 461-8 1-222 Green-yellow 563-6 Violet 433 and beyond 1301 Observations have also been published by v. Frey and v. Kries^, Konig and Dieterici^, Angier and Trendelenburg^ and others^. If these results are plotted as curves with wave-lengths from 400-500 /x^u, as abscissae and wave-lengths from 560-680 [ifx as ordinates the curves nearly resemble hyperbolae, but differ slightly from each other. The differences have been attributed in part to macular pigmentation. Krarup"^ has re-investigated the subject and finds that the comple- mentary colours change somewhat as the intensity of the illumination is altered. There is no change due to this cause from 460 to 480 i^fM, but a gradual increase from that point up to 512 /x/x. With a suitable, relatively low, illumination the curve is a rectangular hyperbola. The ratio of the quantities of Aj to X.,, where Xj and A.^ are the wave- lengths of two complementary colours, is approximately constant and independent of the intensities of illumination. Glan^ came to the con- clusion that the energies of \ and X. at the percipient retinal structures ^ Arch. f. Anat. u. Physiol. 336, 1881. ^ Wied. Ann. xxxiii. 1887; in Konig, p. 2()1. i ^ Ztsch. f. Psychol, u. Physiol, d. Sinncsorg. xxxix. 284, 1905. i 1 Hclmholtz, 3rd ed., ii. p. 107. ^ Physisch-ophlhalmologische Grenzprvbleme, p. 100, Leipzig, 1906. i " Arch. J. d. ges. Physiol, xxix. 53, 1886; Wied. Ann. xlviii. 1893. ! 3—2 36 COLOUR VISION must be equal in order that white may be perceived. This law is not true, but Krarup points out that the ratio of the energies of \i to \o at the retina is independent of the intensities of illumination. The ratio is not constant, still less equal to unity, but if the energy ratios calculated from Angier and Trendelenburg's quantities (slit-widths) be plotted as ordinates, against wave-lengths as abscissae a symmetrical curve re- sembling a parabola and having its apex at about 608 //./x results. Ebbinghaus^ states that the brightness of the resultant white is equal to the sum of the brightnesses of the constituent complementary colours. Most observers have found that the white is brighter. Complementary spectral colours have seldom the same luminosity, i.e. they do not look equally bright. The nearest approach to equal luminosity is orange, 607'7 ^/x, and blue, 489"7 fifji. Colour mixing shows that the order of saturation of spectral colours diminishes from violet successively through indigo-blue, red and cyan-blue, orange and green, yellow (v. Helmholtz). It will at once be observed that there is a range from about 560 to 492 ju/i, i.e. green, which possesses no spectral complementary. White can only be obtained from green by mixing it with both red and violet, i.e. purple. There are hues with which we are familiar in pigments, etc., that do not at first sight fall into any of the categories mentioned. Of these the most striking is brown. Brown, olive green, and greys possessing some coloured hue are obtained by mixing black with a spectral colour or mixture of colours. Further evidence will accumulate in the course of our discussions in favour of the view that black is an actual and effective stimulus. We can now return to the unfinished colour table or diagram, of which only two rectilinear portions have as yet been mapped out. It has been pointed out that the graph must be a closed figure in one plane and that the various points upon it follow the law of the centre of inertia of masses. If three colours, neither of which can be obtained from a mixture of the other two, are represented by three points on a plane, then assigning to them values in terms of any unit, the situations and quantitative values of their mixtures can be ascertained. The problem is well stated by Greenwood^. " In order to establish the correctness of this method it is necessary to prove that, given the experimental ^ Ztsch.f. Psychol, u. Physiol, d. Sinnesorg v. 176. 1893. ^ Physiology of the Special Senses, London, 1910, p. 131. THE MIXTURE OF PURE-COLOUR STIMULI 37 laws of colour mixing, this construction is valid in all possible cases, i.e. that the situation of the mixed colour coincides with that of the mass centre of two equivalent masses (1) when the two constituents can be mixed from the three chosen colours ; (2) when one can and the other cannot so be mixed ; (3) when neither can be so mixed." v. Helm- holtz^ has supplied the mathematical proof. The diagram will vary in form according to the source of light and according to the choice of units and fixed points. It is best represented in such a form as in Fig. 7. Green Yellow Red Orange Pig. Purple Diagram of a colour table. Violet The position of the pure white sensation is obtained by dividing the line joining any two complementary colours according to the relative amounts of those colours required to produce white when mixed to- gether. The deviation of the curve from a straight line beyond 540 /x/z. indicates the unsaturated nature of the mixtures. Since, for example, mixtures of green and violet are less saturated than spectral cyan-blue, the curve must deviate further from the white point than the straight line joining green and violet. From experi- mental results it is found that the curvature is sharpest in the green. These facts are quite independent of any theory of colour vision whatever, and their importance is absolutely fundamental. " The colour table merely expresses in systematic manner directly observed physiological equivalents of objectively different lights " (v. Kries). The facts show that the totality of physiological activities comprises far fewer elements than the objective stimuli. The varieties of stimuli of all possible mixtures derived from a source of constant intensity can be reproduced on a co-ordinate system as points in a plane. If we introduce variations of intensity the law can be stated in general terms: " The entire physiological valency of every conceivable light and 1 Helmholtz, 3rd ed., n. p. 112. 38 COLOUR VISION light mixture can he comprehensively represented as the function of three variables " (v. Kries). Hence a colour diagram representing varying intensities of colours and colour mixtures must be in three dimensions, as in Lambert's colour pyramid^ or Runge's sphere^. Within a certain range, which includes all ordinary conditions of colour vision except those mentioned on p. 36, this law shows that every conceivable light or light mixture gives rise to a sensation which can be accurately matched by the sensation produced by a suitable mixture of only three lights. In other words, from the point of view of stimuli normal colour vision is trichromatic. It is to be noted carefully that the colour table does not express the change in physiological valency which corresponds to variation in absolute intensity. The unit of intensity is fixed for the given table, as is also the choice of the three variables. Theoretically the choice of unit intensity of the three variables is arbitrary. The choice of variables merely involves a change in the co-ordinate axes. If the variables are selected too close together the table involves negative stimuli and the stimulus values cannot be reproduced experimentally. This is, however, of no theoretical importance. We have here the basis of colour equations. For example, if spectral green-blue {Bg) is mixed with red {R) in certain proportions it matches a mixture of green {Gr) and violet (F), which may be expressed thus : aBg h^R = jGr + eV. Hence we can obtain a value for Bg aBg = jGr + eV-^R, which is strictly accurate though incapable of objective interpretation. As Greenwood well puts it^ — in colour equations " addition is uniform, the same result being always obtained when the same quantities are summed ; it is commutative, the order of operations does not affect the result ; it is associative and homogeneous. . . .If we define subtraction, in terms of arithmetical quantity as uniform, non-commutative and non- associative, similar analogies can be observed ; but this is of little importance, since a justification of the use of the symbol of addition will suffice for our purposes." 1 Beschreibung einer mit dem Calnu'scJien Wachse ausgemahlten Farhenpyr amide, Berlin, 1772. - Die. Farhenhujcl. Hamburs;, 1810. Cf. Clicvroul, Expose d^nn moyen de definir ef de nommer Ics coidcurs, Paris, ISfil ; liofler, Ziscli. f. Psychol, n. Physiol, d. Sinnesorg. LViii. 356, 1911. 3 Loc. cit. p. 133. THE MIXTURE OF PURE-COLOUR STIMULI 39 From the experimental point of view a constant spectrum is chosen and three colours are selected, e.g. a red, a green and a blue. Each part of the spectrum is then matched by mixing different quantities of the three together. This process is called " gauging the spectrum " [Aiclinng cles Spektrums, v. Kries). There are many technical difficulties and the results are not wholly free from objection. The best published results are those of Konig and Dieterici^ and Abney and Watson. The colour tables deduced from these results are seen in Figs. 8 and 9. Green Red G,n Blue Fig. 8. Colour triangle. A, B, C, etc., Fraunhofer lines. (Konig.) In Konig's curve the sudden bend at the extreme end of the violet is probably due to fluorescence. As already mentioned colour mixtures often produce unsaturated colour sensations. Consequently the match has then to be made by a suitable addition of white light to the comparison spectral colour. No three spectral colours can be chosen which when mixed will accurately match in hue and saturation all spectral colours. There is good physiological evidence of colour-sensations of much greater saturation than the spectral colours {vide infra, Section V, Chap. ii). Such colours, being less mixed with white, must lie outside the colour table, in some such positions as shown at the angles of the circum- scribed triangles in Figs. 8 and 9. 1 Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. iv. 241, 1893. 40 COLOUR VISION B 49 50 51 52 53 54 55 56 57 58 Fig. 9. Colour triangle. W, white ; B, red ; G, green ; B, blue. The numbers are those of an arbitrary scale of the spectrum of the arc light. (Abney and Watson.) TJie Influence of Macular Pigmentation. The macula liitea, as its name implies, is permeated with a yellow pigment, and therefore absorbs certain spectral rays more than others. The variations in colour matches and in the estimation of complementary colours by various normal-sighted individuals were attributed by MaxwelP to this cause. Glan^ and Sachs^ examined the absorption of the yellow pigment from the macula of human eyes. Sachs found that absorption is inappreciable in the red and orange, commences in the yellow-green and gradually increases towards the violet end. Moreover, the amount of pigment varies considerably in different individuals. Sachs examined nine specimens : the mean coefficients of transmission of monochromatic lights of these cases are as follows (Krarup)*: 1 Bhil. Trans. Boy. Soc. Lond. CL. 57, 1860. 2 Arch. f. d. ges. Physiol, xxxix. 53, 1886. ^ Ibid. L 574, 1891. * Loc. cit. p. 18. THE MIXTURE OF PURE-COLOUR STIMULI 41 X = 670— 590 580 575 570 560 555 550 540 535 530 1 0-991 0-986 0-981 0-971 0-966 0-962 0-951 0-946 0-941 X= 520 510 505 500 490 480 474 470 464 454—420 0-905 0-800 0-770 0-740 0-700 0-680 0-677 0-675 0-672 0-670 Owing to this absorption one would expect variations in colour mixtures, and it is indeed found that, for example, in matching a homogeneous yellow with a mixture of red and green, some people require slightly more green than others, so that a given match appears too green to some, too red to others. If such people make the match eccentrically, so that the images fall just outside the macula, where there is no yellow pigment, their matches agree^. v. Frey and v. Kries^ exhaustively investigated the matches of numerous normal-sighted individuals (students) and found that the deviations were such as would be expected from pig- mentary absorption. This variation in individuals is therefore purely physical, and must not be confounded with allied variations which are due to physiological abnormalities. (See Part II, Chap, iii.) Since the physiological abnormalities very rarely affect the blue end of the spectrum, a ready method of distinguishing the effects of macular pigmentation from them is by measuring their match between spectral blue-green and a mixture of blue and green, v. Kries^ found that reduction of blue by minimum and maximum macular pigmentation was in the ratio 1 : 0"31, the corresponding ratio for green (517 /x/x) being 1 : 0-5. (See also Section IV, Chap, ii.) The Influetice of Lenticular Pigmentation. With advancing years the crystalline lens becomes tinged with a yellow pigment. The effect is similar to that of macular pigmentation, but does not manifest itself until late adult life. Konig* estimated the coefficients of transmission of monochromatic lights by the lens of a man, aet. 55. 1 Cf. Hering, Arch. /. d. ges. Physiol, uv. 277, 1893; Breuer, Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xin. 464, 1897. ~ Arch. f. Aimt. u. Physiol. 336, 1881. 3 Ztsch. /. Psychol, u. Physiol, d. Sinnesorg. xm. 284, 1896. " Konig, p. 346. 42 COLOUR VISION CHAPTER IV THE LUMINOSITY OF THE SPECTRUM " It is to be noted that the most luminous of the prismatic colours are the yellow and orange. These affect the senses more strongly than all the rest together ; and next to these in strength are the red and green. The blue compared to these is a faint and dark colour, and the indigo and violet are much darker and fainter, so that these compared with the stronger colours are little to be regarded^." Fraunhofer^ first published in 1817 measurements of the brightness of the various parts of the spectrum. Vierordt^ published a very similar curve of the " strengths of coloured light " of the sun's spectrum. He measured the amounts of the white light which must be added to various parts of the spectrum in order to produce a just appreciable diminution in saturation. Fraunhofer's and Vierordt's curves agree well except towards the violet end, a fact which is of considerable theoretical importance^. Draper^ however found all parts of the dif- fraction spectrum equal by this method, and the red brightest in the prismatic spectrum. Accurate investigations of more importance for our purpose date from those of Abney and Festing^ and Konig and Ritter'^, but before entering upon these and other researches we must attack the difficult problem of realising what exactly is being measured. It may be stated at once that we shall be unable to give a completely satisfactory solution to the problem at this stage, but as we proceed in our discussion the profound importance and significance of luminosity in colour vision will become more apparent. Two white lights of different intensity impress the senses with stimuli which are interpreted as differing in brightness or luminosity. The measurement of differences of luminosity is the function of photo- metry. Weber's and Fechner's laws (p. 20) suffice to show that even with white light the relationship of the sensations to the stimuli is by no means simple. When lights of different colours are compared the difficulties 1 Sir Isaac Newton, Optics, Bk. i. Part I, Prop, vii, Theor. vi. 1704. ^ Gesammelte Schriften, p. 1, Miinchen, 1888. 3 Pogcj. Ann. cxxxvii. 200, 1869. * Cf. Helmlioltz, Zfsch. f. Psychol, u. Physiol, d. Skmcsorg. ii. 1, 1S91. 5 Loncl., Edin. nnd Dublin Philos. Mag. viii. 75, 1879. 6 Phil Trails. Roy. Sac. Land. 1886, 1888, 1892, 1899, etc. ~' 1891, in Konig, p. 144. THE LUMINOSITY OF THE SPECTRUM 43 are greatly increased. The comparison of brightness or luminosity of coloured lights is known as " heterochromic photometry " (Briicke). In conformity with Newton's remarks, quoted at the beginning of this chapter, and with the general consensus of opinion, brightness must be regarded as a quantitative conception. When one colour is added to another on the same field we are conscious that the brightness of the field is greater than with either colour separately. An exact match between two colour mixtures means that they have the same hue and the same luminosity, and v. Helmholtz draws the conclusion that Grassmann's second law (p. 33) can be extended to include brightness^. He does not admit, however, that " equally bright lights added give equally bright mixtures," and there is no doubt that this statement is incorrect under certain circumstances (Purkinje's phenomenon, v. p. 57). Apparently this is not the only exception V. Helmholtz would admit, for as he says " equally bright red or blue doubled give the red brighter than the blue." Briickner^ also holds that the brightness of a mixture is not always equal to the sum of the luminosities of its components, blue and yellow especially giving a brighter mixture. V. Helmholtz indeed expresses quite candidly his doubts as to the accuracy of direct comparison of the luminosities of different coloured areas. " I scarcely trust my judgement iipon the equivalence of the heterochromic brightnesses, at any rate upon greater and smaller in extreme cases. I admit, however, that one can gradually so darken one of two coloured fields that no doubt remains as to the other being now the brighter.". .." As far as my own senses are concerned I have the impression that in heterochromic luminosity ecjuations it is not a question of the comparison of one magnitude, but of the combination of two, brightness and colour-glow {Farbengluth), for which I do not know how to form any simple sum, and which too I cannot further define in scientific terms." V. Kries^ says that " it must not be considered as an obvious fact that e.g. a given blue and a given red sensation are necessarily in the strictest sense equally bright. For we could only state this definitely if we were certain that the equality of the luminosity impression corresponded with a particular physiological condition so that we could substitute a well defined physiological entity for the subjective conception of brightness ; but this is by no means the case." » Helmholtz, 2nd cd., p. 440. 2 ^^c/i. /. d. gi\^. Physiol, xcviii. 90. 1904. * Nagel's Handb. d. Physiol, d. Menschen, m. 259. 44 COLOUR VISION Whether or not we agree with these words, that a high degree of accuracy in such estimations can be attained is shown by Abney's results, notably the following^ : " To test the illuminating value of colour mixtures, three slits were placed in the spectrum, in the red, green and violet. The luminosities of the rays coming through each were measured — (1) separately ; (2) in pairs ; (3) the whole combined. R 20.3 (R+G) .. 242 G 38-5 (G+V) .. 45 V 8-5 (R+V) .. 214 (R+G+V) 250 Combining these together we get : R+ G+ F=250 (R+G)+ F =250-5 (R+ V)+ G =252-5 {G+ V)+ 72=248 {R+ G+ F)=250 =250-25 by least squares. Within the limits of error of observation the luminosity of the combined spectrum measured as white is equal to the luminosity of spectrum colours measured separately." The very fairly coordinated results of different observers and different methods show that in these luminosity measurements there is something measured which appeals to their senses as of the same order. Most will agree that it is a complex " something " which is expressed by V. Helmholtz as a combination of " brightness and colour glow," and which we shall see later is an integral part of Hering's theory and is expressed as the " white valency " of the colours. Further analysis of this " something " must be left till a later stage. In the meantime we will designate it " luminosity." Direct comparison or "equality of brightness 2" is not the only method of measuring the luminosity of spectral colours and mixtures. So early as 1735 Celsius employed the visual acuity to determine the relative brightness of various regions of the spectrum, and this method was also employed by Buffon and Sir Wm. HerscheP. The same principle has been applied to white light in the discrimination photometers of Houston 1 Abney, p. 105. ^ This is the term usually api^lied to this method by physicists. 3 Parsons, Roy. Lond. Ophth. Hosp. Rep. xix. 277, 1914. THE LUMINOSITY OF THE SPECTRUM 45 and Kennelly, Fleming and others^. This method, however, is too inaccurate for the purpose in view, though Crova's curves^ are in- teresting^. Charpentier^ and Haycraft^ used the method of finding the dis- tance at which small areas of different colours become invisible. Another method, based on the fact that the size of the pupil depends, inter alia, upon the intensity of the light, has been advocated by Gorham^, and has also given good results with coloured papers. Rivers', Martins^, and Briickner^ have suggested other methods. Heterochromic photometry has become of importance technically since about 1880 on the introduction of the arc light for commercial purposes. Hence a considerable amount of attention has been devoted to it^°. Apart from the equality of brightness method adopted by Abney and Festing, and Konig and Ritter, the method which provides most accurate data for physiological purposes is flicker photometry, suggested by Talbot (1834), but first used by Schafhautl (1855), and subsequently by Rood^^, Polimanti^^ and others. When before two lights of different luminosities an episcotister (a metal disc containing alternate closed and open sectors) is rapidly rotated, an unpleasant flickering of each light is seen at certain rates of rotation. Now it is found that the flickering ceases at the same speed of rotation of the episcotister for the two lights when their luminosities are the same. The causes of flicker are complex and will require further attention later (Section V, Chap. i). The results obtained are not free from criticism, but if due precautions are taken they agree well with those yielded by the equality of brightness method. Konig's luminosity curves for the gas spectrum with different light intensities are given in Fig. 10. They were obtained by the equality of brightness method. We are concerned here only with the higher intensities {F, G, H). The curves have the same ordinate at 535 /x/z, because that ordinate was taken as unity. It will be noted, however, that they all cross and reverse their relative positions. This is due to the Purkinje 1 Trotter, Illumination, London, 1911. ■^ Ann. de Chim. et de Phys. vi. 528, 1885 : La Lumiere elec. xvm. 549, 1885. * Trotter's Illumination, p. 171. * La Lumiere et les Couleurs, Paris, 1888. * J. of Physiol. XXI. 126, 1897. « Proc. Roy. Soc. Lond. xxxvn 425, 1884. ' J. of Physiol, xxn. 137, 1897. ^ Beitrdge zur Psychol, u. Philos. i. 95, 1896. * Arch. f. d. ges. Physiol, xcviii. 90, 1904. lo Trotter, he. cit. 101. " Amer. J. of 8c. xlvi. 173, 189:5 ; Science, vu. 757 ; viii. 11, 1898. 12 Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. xix. 263, 1899. 46 COLOUR VISION 3 8- 3-6 ^ \ 3-4- / Light. int.p.nRitip!iH 3-2 G 3-0 / / / ( / ( / / 1 1 i \\ F E ?R / / t \ \\ 2-G / / / ' 1 . D ;_ , c 2-4- ' , V I _ „ B 2-2 f i 'n v\ A 20 1-8 If 1 1 1 1 1G ji 1 1 K 1 1 1 1-4 I'l III i ' \ 1-2 ' 1 i / -■■ --'- --.^ ..\ 10 II " 1 / / /'/ r- ".'■'■■ \S^^^ 0 8 '/• ■■''/.■■' ^ ^ x."^ V -v. 00- 0 4 0-2- 1 ■ / / f ... .■ 1 870 650 625 605 590 575 555 535 520 505 490 470 BCD E F 450 430 G Fig. 10. Normal trichromat's luminosity curves lor different intensities of light {A, B, C, cto. ; A being the lowest, and // the highest intensity). Abscissae, wave-lengths of the prismatic spectrum of gas light ; ordinates, an arbitrary scale. (Konig.) r— — ~\ 0 'AN too / / \ 90 7 \ \ ^l / \ \ Rn .* \j/ f \ \\ ^// \ 70 / \ 4 ' 1 5 -A V fin / / 7 y f f 1 "^ rA \ •in J 1 § n ^ 7 1 1 \ f \ nn / ' / {i u \ A v/ / \ A i 10 / ^i> ' / / ^ i \» J / ^y / '2 \^ i 7ft / /. / \ ^ ^ 7^ / / \ \ 10 r1 ^ ^ / V s^ 1 {±1 ^ "^ ^ === =^ I 7 £ 1 1 1 f^ ^»~ _. .0 6 8 10 12 H 16 18 Id ZZ Zt, U Z8 50 32 34 36 18 4o « 44 4€ 48 50 5Z 54 56 5B 60 62 64 66 m 3839« 506 S8-ISS'7 6l-3 Fig. 11. Normal trichromat's photopic luminosity curves for the fovea, macular region, and 10° from the fovea. Abscissae, arbitrary scale of the prismatic spectrum of the arc light, the Fraunhofer lines and some important wave-lengths bemg indicated ; ordinates, arbitrary scale, the maximum luminosity being 100. (Abney.) THE LUMINOSITY OF THE SPECTRUM 47 phenomenon (p. 57). Maximum luminosity is at about (JlO/x/x with these intensities. Krarup^ has calculated the curves from Konig's results in terms of the energy values of the spectrum, making allowance for absorption by the macular pigment, so that the results are inde- pendent of the particular source of light. Abney's luminosity curves for the spectrum of the arc light with inclined carbons are shown in Fig. 11. They were also obtained by the equality of brightness method. The following are the macular luminosities for the chief Fraunhofer lines, the source of light being the crater of the positive pole of the electric arc, with sloping carbons : Line \ Lumincsity B OSfi-G /u/x 4 Li (red) 670-5 8 G 656-2 17 D 589-2 09-5 E 526-0 48 h (Mff) 518-3 36 F 486-0 6 Li (blue) 460-3 2 G 430-7 0-() The difference between the luminosity curve by the equality of brightness method and the flicker luminosity curve is shown in Fig. 12. /oo 90 80 70 60 50 -fO 30 20 /O SS/^ 26 30 55 'fO -fS SO S5 60 65 Fig. 12. Normal trichromat's photopic luminosity curves with the equality of brightness and the flicker methods. Abscissae, arbitrarj^ scale of the prismatic spectrum of the arc light ; ordmates, arbitrary scale, the maximum luminosity by the equality of brightness method being 100. (Abney.) (Cf. Fig. 38.) 1 — A "^ \ h / N \ . 7 (/ vv // /n V 7 M LI mine 5ity ( urve > N Fl icker curve (max) 925) i ^ t r ^ // f \ ^ ^ ^ b \ E 1 D 1 I-' Li lJ ^ ^ Loc. cit. p. 46. 48 COLOUR VISION The flicker curve was taken with Prof. W. Watson's flicker apparatus used in conjunction with Sir Wm. Abney's apparatus. The conditions were — arc light with positive carbon horizontal, current 18 amps., D line ^ 1 metre candle, patch 1 sq. inch observed at a distance of 3 ft. It will be noticed that the flicker maximum is less than the equality of brightness maximum, and in the proportion 92 '5 : 100, but Ives's results {v. p. 96, Figs. 38, 39) show almost complete coincidence. The most striking confirmation of the general accuracy of these methods is that the luminosity values thus found agree with their periphery values. We shall see later (p. 67) that at the periphery of the retina colours of moderate intensity cease to convey the impression of colour and appear white, so that the periphery values are not com- plicated by the intrusion of the colour factor. Polimanti^ confirmed this fact for the flicker method. 1 Loc. cit. SECTION III THE SPECTRUM AS SEEN BY THE DARK-ADAPTED (SCOTOPIC) EYE CHAPTER I ADAPTATION OR TEMPORAL INDUCTION When we pass suddenly from bright sunlight into a dimly lighted room we can see nothing for a time until we become " accustomed to the darkness." When we pass from the dark into bright light vision is also difficult and may be painful. We therefore infer that the sensi- bility of the retina becomes increased at low illuminations. This automatic process of levelling the sensibility of the retina to the require- ments of the moment is called dark or light " adaptation " (Aubert) or " temporal induction." Dark adaptation is a relatively slow process. It is characterised by a rise in the sensitiveness of the retina to light, which is slow during the first ten minutes of exclusion of light from the eyes, rapid during the following twenty or thirty minutes, and again slow or almost negligible after that period. The general character of the curve of retinal sensibility is the same in all cases, but there are marked individual variations in the rapidity and amount of the rise, thus explaining the fact that some people see very much better in a dull light than others, though variations in the size of the pupils and other factors {vide infra) are not without importance in this respect. In night-blind people there may be only a very slow rise, the ultimate sensibility after an hour being near the normal limit. In severe cases there is very little rise even after several hours. Such adaptation is normal in the colour-blind, even the totally colour- blind^. Strychnin and brucin cause increase in the amount and rapidity 1 Tschermak, Ergebnisse d. Physiol, i. 1, 700, 1902. p. c. V. 4 50 COLOUR VISION of the rise of sensibility : santonin has no effect. Mydriatics have an indirect effect ; the first slow rise is prolonged from ten to twenty minutes but is followed by the normal rapid rise to the normal height. Very short exposure to bright light, e.g. striking a match, causes a very temporary fall without materially altering the course of the curve. The increase in sensibility after very prolonged dark adaptation is more transient than the increase during the first hour, i.e. it is more quickly and completely abolished by exposure to light. Dark adaptation of one eye has no effect upon the other^. Besides this temporal variation in the sensitiveness of the retina there is a well-marked regional variation. In the condition of light adaptation the fovea is the most sensitive part of the retina, though little attention has been paid to the degree of adaptation in the researches published on this subject. (The light sensitiveness of the various parts of the retina must be carefully distinguished from their visual acuity for form.) The regional sensibility for colours of the retina of the light- adapted eye has been worked out by Vaughan and Boltunow^, v. Kries^, and Guillery*. Vaughan and Boltunow found the sensitiveness at 10° from the fovea to be \, at 20° to be y^, and at 35° to be ^f'^^ of that of the fovea itself. In dark adaptation the fovea is the least sensitive part of the retina^. In other words the fovea is a region of physiological night-blindness (v. Kries). The relative central scotoma in dark adaptation was long ago recog- nised by astronomers, who noticed that stars of small magnitude were seen better if viewed somewhat eccentrically. " Pour apercevoir un objet tres pen lumineux, il faut ne pas le regarder " (Arago). It is strikingly illustrated in viewing the Pleiades : by direct fixation four or at most five stars are seen ; by indirect fixation a number of weaker stars become visible. Different observers use different parafoveal^ spots for clearest vision in dark adaptation' and the spots vary with the degree of dark adaptation. The nearer the intensity of the stimulus is to the threshold of the dark-adapted fovea the nearer is the spot to the fovea : the feebler the light the more eccentric is fixation. With a given sub-minimal foveal stimulus Simon found that he fixed 2° from ^ Cliarpentier, La Lumiere et les Couleurs, p. 175, Paris, 1888. 2 Ztsch. f. Sinnesphysiol. xlii. 1, 1907. ^ Zl-sch. f. Psychol, ii. Physiol, d. Sinnesorrj. ix. 81, 189fi. * Ibid. xii. 261, 189() , xiii. 189, 1897. ^ Donders, Brii. Med. J. 1880. * I.e. in the region near the fovea. ' Christine Ladd-Franklin, in Konig. p. 353 ; Simon, Ztsch. f. Psychol, u. Physiol, d. Sinmsorg. xxxvi. ]8fi, 1904. ADAPTATION OR TEMPORAL INDUCTION 51 the fovea after ten minutes dark adaptation, 1|° after twenty minutes, and 1° after an hour. The direction is constant for the same eye and varies with different eyes ; it depends upon muscular balance and refraction rather than upon the specific sensibility of the parts of the parafoveal region (Simon). Although the fovea is night-blind relatively to the periphery it is capable of a slight degree of dark adaptation^, but the small rise in sensitiveness of the fovea is only appreciable after previous very strong light adaptation, such as looking at the open sky. Breuer and Pertz^ showed that the peripheral rise in retinal sensi- bility in dark adaptation is rapid from 1° to 4° around the fovea, then slower to a maximum between 10° and 20° beyond which it falls. This is seen graphically in Fig. 13, where the sensibility of the fovea and the Fig. 13. Sensibility of the fovea and parafoveal region for mixed bluish-white light. Abscissae, to the left of zero degrees to the temporal, to the right, degrees to the nasal side of the fovea (0) ; abscissae, arbitrary scale. (Breuer and Pertz.) parafoveal region to bluish-white light are shown. The abscissae to the left represent the teniporal side, to the right the nasal. Pertz's experiments show that the scotopic fovea is more sensitive to red light than the periphery, though the difference is slight. Blue light gives a curve resembling that in Fig. 13. Yellow shows 'a slight rise in the paracentral area. The alterations in sensibility differ according to the size of the area 1 Charpcntier, Arch. iTOpht. iv. 291, 1884 ; xvi. 87, 1 896 ; Fick, Arch.f. d. yes. Physiol xmi. 481, 1888; Treitel, Arch.f. Ophth. xxxv. 1, 50, 1889; v. Kries, Arch.f. Ophth. xlii. 3, 95, 189G; Tschermak, Arch. f.d. fjea. Physiol, lxx. 320, 1898; Bloom and Garten, ibid. Lxxii. 1898. ^ V. Kries, Ztsch.f. Psychol, u. Physiol, d. Sinncsorfj. xv. 327, 1897. 4—2 52 COLOUR VISION stimulated, and the relations between sensibility and the area stimulated are different in the light- and dark-adapted eye, and also in the foveal region and the periphery. For foveal vision the sensibility is pro- portional to the area stimulated (Ricco, Loeser^). In the dark-adapted periphery the sensibility is proportional to the square root of the area stimulated'^, but only for composite- white light and objects subtending a visual angle of 1° to 10°. Above 10° the sensibility rises more slowly. The rise is still less even for smaller angles with red light. The rise of the curve of sensibility in dark adaptation therefore varies with the size of the area of retina stimulated and with the nature of the light. In the light-adapted eye there is no definite relationship between the rise of sensibility and the size of the visual angle. In dark-adapted eyes binocular summation of stimuli occurs, so that the sensibility is about twice as great with both eyes as with only one (Piper), though individual variations occur. In light-adapted eyes no such binocular summation occurs (Fechner^), but care must be taken that there is good light adaptation, and one eye must be covered for only a very short period, otherwise partial dark adaptation occurs. In this respect there is a noteworthy analogy to the effects of the size of the area of the retina stimulated : with complete light adaptation the stimuli to different parts of the retina are not summated, given that the visual angle exceeds a certain (small) size, nor are the stimuli to the two eyes summated, whereas in the condition of dark adaptation both summations occur. CHAPTER II SCOTOPIA OR TWILIGHT VISION It will be readily appreciated that complete dark adaptation rarely occurs under normal conditions of life. Scotopia is the condition of vision in which there is a relatively high degree of dark adaptation. It will be best to consider the conditions of vision after prolonged stay in a feebly lighted room. If now coloured objects are viewed under feeble illumination the colours cannot be distinguished, but all appear to be of various shades of grey. The eye is totally colour-blind. ^ Parsons, Roy. Loncl. Ophth. Hosp. Rep. xix. 114, 1913. 2 Piper, Ztsch. /. Psychol, u. Physiol, d. Sinnesorg. xxxn. 98, 1904. » V. Helmholtz, 3rd ed. ii. 287. SCOTOPIA OR TWILIGHT VISION 53 A spectrum of low intensity appears as a colourless bright streak, varying, however, in brightness in different parts. Consequently accurate matches can be made between any two parts of the spectrum by merely modifying the intensity of one light. If the intensity of the spectrum is slightly raised the colours become evident in a definite order and the relative brightness of the different parts becomes altered. As the intensity is still further raised the eye becomes light adapted and the spectrum shows all its hues with the relative brightnesses described in Section II. Scotopic vision at very low intensities is therefore achromatic ; with slightly raised intensities of light it becomes chromatic. We may distinguish the two conditions as achromatic and chromatic scotopia respectively. The achromatic scotopic values of different parts of the spectrum were first investigated by Hering and Hillebrand^ for the dispersion spectrum of daylight. Abney and Festing's curve^ is shown in Fig. 14. The striking feature is that the brightest part, instead of being in the neighbourhood of the D line (yellow), is moved further towards the violet end, and is at about 530 /Z/u- instead of at about 580 /^t/u,. The luminosity curve falls slowly towards the violet end, and sharply towards the red ; the red end is shortened. It may be stated at once that all types of colour blind give almost identical achromatic scotopic curves^. Hence we can consider here results obtained both by normal observers and by colour-blind observers like Nagel* (a deuteranope). Very accurate observations were made by Schaternikoff^ in v. Kries's laboratory. Fig. 15 shows the similarity between Nagel's (deuter- anopic) and Schaternikolf's (normal) curves. Fig. 16 shows the sunlight, Nernst light, and gas light curves^. The summits of the gas light curve (537"2/x/x) and the sunlight curve (529'3;u./x) differ slightly, and light from cloudless sky gives a rather higher value in the green-blue and blue than direct sunlight. An Auer lamp, which is rich in green rays, will give a slightly different curve from a carbon filament electric lamp, which is rich in red rays. The curve will depend upon the energy distribution in the given spectrum, and of course the diffraction or interference spectrum curve 1 Siizungsher. d. Wiener AkacL, math.-naturw. Kl. xcviii. 70, 1889. 2 Phil. Trans. Roy. Soc. Lond. clxxxiii. A, 531, 1892. ^ Raehlmann, Zlsch. f. Augcnlilk. ii. 31.5, 403, 1899, for a possible exception ; also Tschermak, Ergeh. d. Physiol, i. 2, 703, 747, 1902. * V. Kries and Nagel, Ztsch.f. Psychol, u. Physiol, d. tiinnesorg. xii. i, 1896. 5 Ibid. XXIX. 2fi5, 1902. * Trendelenburg, Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. xxxvn. 1, 1904 54 COLOUR VISION ■ t ^ ps f1 ,N ^ ^/ V' / .* \ ^^ \ /,^ i\ V ^ ^^ .^ / "U f / aA ^ \ < r/ /. \'^ / ,/ \ *^ 11 ^y /^ o Vs u k\ ^ y /^ l" \ 3 / / i \ *i ^ ^/ / (^ 1 5( \ ^ y ' A l^^ / ^ v^ \ h A 1 \ ?t. \ ■ x: V / ^ r*^ . r^'V i> / /yi V* A i i.i' „iL > / '\ ^- \ ■C ^ t . X r-' -\ tr> \ ^ r «• H , ' — -r ■^ t 1 ■ U 1 -. \ -. it. i 1 K '00 90 30 70 60 50 40 30 20 - 10 6 8 10 \Z 14 16 18 ZQ Zt 24 26 2& 50 52 JS4 56 58 40 42 44 46 48 50 52 54 56 58 60 62 64 66 22-8 34 39 8 50 6 591597613 Fig. 14. Normal trichromat's photopic and scotopic luminosity curves. (One amy] acetate or Hefner unit is equal to 0-9 international candle power. ) Abscissae, arbitrary scale of the prismatic spectrum of the arc light ; ordinates, arbitrary scale, the maxi- mum of luminosity being in each case 100. (Abney and Festing.) 300 280 • /A // ft f \\ 260 240 22 0 ■ 20 0 ■ 180 160 * 1 ll y 140 120 ■ 100 80 60 40 20 0 J 01 2345678 9 10 1112 13 14 15 oioi olOioicnoicnoioicn Ol 05 Ol O) O) ^ Oi CO -* O 0-» -t».(DCOCD^ 01CnO>-«l CONOCnCD K3 CD CD CO -* W-* 4kO*>.IO COOJ J^-J fO 17 19 21 o 03 CD o CTI O 23 25 27 4^ CO ID Fig. 15. Scotopic luminosity curves. trichromat). — Nagel's curve (deuteranope) Schaternikoff's curve (normal X X X X X and o o o o o. other observations by Nagel. Abscissae, wave-lengths of the prismatic spectrum of gas light ; ordinates, arbitrary scale. (Schaternikoff.) SCOTOPIA OR TWILIGHT VISION 55 will differ from the prismatic curve. Krarup^, using the energy values for the Hefner light calculated by Angstrom, and allowing for absorption by the macular pigment on the basis of Sachs's researches, has calculated the achromatic scotopic luminosity curve from Konig's results. This curve is independent of the source of light employed and shows a remarkable agreement with (1) the curve of absorption of visual purple ; (2) the luminosity curve of the totally colour-blind {vide infra) ; and (3) the similarly corrected achromatic scotopic luminosity curve of a protanope investigated by Konig. 3000 ^ \ 2500 - {/ \ ^: \ £000 '// 1 > k^ \, // '/ ^ \ \ 1500 . // 1 ^ \ \ 1000 1, ' N \ \. J 1 \ k. ^ 1500 ../ ^N, ^:^ *v 1 1 1 11 1 1 =^ Sunlight Nernstlight Gaslight ^2 S Iff If III I Fig. 10. Scotopic luminosity curves for direct sunlight (Schaternikofi'), gas light (Schater- nikoff), and Nernst light (Trendelenburg), reflected from a magnesium oxide-coated surface. Abscissae, wave-lengths of the prismatic spectrum ; ordinates, arbitrary scale, the maximum of luminosity being in each case 3000. (Trendelenburg.) The flicker method gives concordant results. Fig. 17 shows Haycraft's^ results for colour-discs with three intensities. The ordinates show the number of revolutions of a semi-disc per second which were found necessary to abolish flicker : the abscissae are wave-lengths. The extraordinary similarity between the scotopic luminosity curve and the curve of values of different monochromatic lights for bleaching frog's visual purple (Trendelenburg) (Fig. 1), seems to point to some underlying physiological fact of great significance. We shall return to this point later. ^ Loc. cit. p. 21. 2 J. of Physiol, XXI. 12(». 1897. 56 COLOUR VISION As may be easily imagined there are many practical difficulties about the observations. Before examining these and other such experiments more in detail it will be well to consider briefly some of the outstanding peculiarities of scotopia or twilight vision as compared with photopia or daylight vision. 400 460 620 680 640 700 760 Fig. 17. Liiminosity curves determined by the flicker method (critical frequency readings). The upper curve i-epreseuts a briglit spectrum, the lower a dark spectrum, and the middle curve one of intermediate intensity. (Haycraft.) " At nightfall in the summer the order of disappearance of colour may often be seen ; orange flowers may be plainly visible, yet a red geranium may appear black as night ; the green grass will be grey when the colour of the yellow flowers may yet be just visible. An early morning start in the autumn before daybreak will give an ample oppor- tunity of satisfying oneself as to the order in which colours gradually reappear as daybreak approaches. Red flowers will be at the outset black, whilst other colours will be visible as grey. As more light comes from the sky the pale yellow and blue flowers will next be distinguished, though the grass may still be a nondescript grey. Then, as the light still increases, every colour will burst out, if not in their full brilliance, yet into their own undoubted hue^." It is instructive with good dark adaptation in a dull light, to arrange different coloured cards, about ^ Abney.. Colour Vision (Tyndall Lectures), p. 107, 1895. SCOTOPIA OR TWILIGHT VISION 57 the size of a postcard, according to their apparent brightness. Red and green will occupy the dark and light ends of the series and orange will match a dark blue. Similar experiments can be made to give quantitative estimates with the colour top. The explanation of these facts is found in the scotopic luminosity curves. In the bright prismatic spectrum of gas light the red (670 /x/x) appears about ten times as bright as the blue (480 ixfju). On the other hand the achromatic scotopic value of the red is less than one-sixteenth that of the blue. Hence with failing light the brightness of different coloured objects alters, the colours towards the red end of the spectrum becoming relatively darker, those towards the violet end brighter, so that finally the reds appear almost black and the blues bright. This fact was first investigated by Purkinje (1825) and is known as Purkinje's phenomenon. Hering^ drew attention to the fact that the brightness of the blues increases much more rapidly than that of the reds diminishes. This depends upon the condition of dark adaptation of the retina. K5nig'2 {v. p. 45) paid more attention to variations in the intensity of the light. The dift'erences of these two authors on the subject of Purkinje's phenomenon depend upon the fact that two factors have to be considered — dark adaptation and stimulus-intensity {vide infra). It is to be noted that in every-day experience we are not dealing with complete dark adaptation. The scotopic luminosity curves given above apply only to very thorough dark adaptation. In such experiments as Konig's (Fig. 10) the adaptation was certainly changing, though there must have been a fair degree of dark adaptation, and the main alteration in the form of the curves is rightly attributed to intensity of stimulus. It will be seen from these and other such curves that the shift in sensation intensity with diminished stimulus intensity is gradual. I wish to avoid as far as possible introducing theoretical considerations into this part of my book, but it will simplify comprehension of the subject if it is pointed out at once that the easiest explanation of the gradual shift of the curves is the simultaneous action of two processes, one especially related to vision with relatively high stimulus values and light adaptation, the other to vision with low stimulus values and dark adaptation. Upon this explanation scotopic vision is a relative con- dition, its nature depending upon low^ stimulus values and a com- paratively high degree of dark adaptation. Whether the processes are subserved by independent physiological mechanisms is a question to be ^ Arch.f. d. ges. Physiol, i.x. 51G, 1895. 2 Konig, p. 114. 58 COLOUR VISION deferred, though what has been said in the last sentence is hardly in- telligible on any other assumption. Complete dark adaptation is only reached after prolonged exclusion of light. Hence a moderate degree of scotopia is generally present in every-day life, sufficient indeed to elicit the Purkinje phenomenon merely by sudden diminution of the intensity of the stimvilus. If after remaining for a considerable time in a moderately lighted room the illumination is suddenly diminished reds at once appear much darker and blues much brighter^. The relative roles played by dark adaptation and by intensity of stimulus may be stated somewhat as follows. Dark adaptation deter- mines the degree of scotopia. This is shown by the fact that the colour threshold or threshold value of photopic vision {Schivellenivert des Tagessehens, v. Kries) remains constant whilst the general threshold becomes lowered. Hence with good dark adaptation scotopic luminosity values are high. As light adaptation increases the photopic values increase and ultimately preponderate: hence Purkinje's phenomenon. With still stronger light adaptation intensities of stimuli below the photopic values are sub-liminal. As regards intensity of stiimdus, as soon as it reaches the chromatic threshold the luminosity values increase throughout the spectrum, but the increase in the red is very slight as compared with that in the blue. The older authors (Dove, Grailich, Helmholtz) directed their attention almost entirely to the intensity, so that Konig lays stress upon the alteration in colour matches as " deviations from Newton's law of colour mixtures^." Hering^ and his pupil Tschermak^, on the other hand, regard adaptation as the sole cause. Hering^ used two rooms separated by a partition which had round holes in it filled with coloured glass. Alteration in the illumination in one room varied the adaptation and in that of the other the intensity of the light observed. He found that with continued light adaptation, diminution of intensity of red and blue caused equal change in the luminosity of the two colours. Red, however, retained its colour slightly longer and appeared slightly brighter. With dark adaptation the blue became whiter and brighter ^ Bering, loc. cit. 2 Konig, pp. 108, 4] 6 ; cf . Tonn, Ztach. f. Psychol, u. Physiol, d. Sinnesorg. vii. 279, 1894. 3 Lotos, VI. 1885 ; vii. 1886. « Arch. f. d. ge.'i. Physiol, lxx. 297, 1898. * Arch.f d. ges. Physiol, lx. 519, 1895. SCOTOPIA OR TWILIGHT VISION 59 than the red without bringing about any change in intensity. Even momentary adaptation on lowering the illumination elicits the Pur- kinje phenomenon. These experiments cannot be regarded as con- clusive, partly on account of the different amounts of white light transmitted by the coloured glasses, but undoubtedly adaptation plays the preponderant role^. As might be expected from our previous experiences of luminosity phenomena the scotopic relative increase in brightness is associated with a diminution in saturation. Those red rays which have little (620 — 670jUjLt) or no (beyond QTO/jl/jl) scotopic stimulus value, show no change. The eft'ect of Purkinje's phenomenon is seen graphically in Fig. 18, where the abscissae represent intensities of red lights (670 /x^) and the ordinates those of blue light (450 fifx) ; points on the curve represent equivalent luminosities^. The deviation from a straight line at low intensities shows the relative increase in brightness of the blue light. 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 Fig. 18. Curve of equivalent luminosities of red (670 /xfj.) and blue (450 fxfx) lights at difiFerent intensities (slit-widths). (Konig.) Although different colours alter their appearance on diminution of their intensity it might be thought that accurate photopic matches would remain valid for the scotopic eye. This, however, is not the case. Albert^ noticed that spectral yellow when matched with a suitable mixture of red and green ceased to match on diminishing the intensity, the mixture becoming brighter and less saturated than the homogeneous ^ Cf. Feilchenfeld, Ztsch. f. Sinnesphysiol. XLiv. 51, 1909. ^ Konig and Brodhun, in Konig, p. 144. ^ Wiedemann's Ann. xvi. 129, 1882. 60 COLOUR VISION colour. Ebbinghaus^ and Christine Ladd-Franklin^ almost simul- taneously drew attention to the significance of the facts. They showed that three whites made by mixture of red and blue-green, yellow and blue, and green-yellow and violet darken unequally with proportionally diminishing intensity, the first least, the second more, and the third most. Konig's experiments^ with complementary colours are confirma- tory. Still more so are observations on the colour-blind. On the other hand as soon as dark adaptation is sufficiently great to abolish the sensations of colour any alteration of the intensity of the light which does not exceed the colour threshold fails to alter the various matches. It is true that Stegmann^ found slight differences, but they were too slight to be of much importance, and such as they were were in a direction opposite to that taken in the Purkinje phenomenon. As Lummer^ pointed out, the peculiarities of scotopic vision explain an observation of Williams^, H. F. Weber'^, and Aubert*^ that a body heated to redness in the dark first shows a grey glow. This occurs at 400° C. (H. F. Weber ; 379° C, Gray^) ; as the temperature rises the yellow-green rays increase and cause a yellowish-grey glow. At about 525° C. (Draper) the red glow commences, but the temperature varies with the conditions of the experiment. If the observer is light-adapted these preliminary stages are invisible. Abney^° had previously recog- nised the explanation of these facts. From what has already been said we see that there are two thresholds of vision — a general threshold, the minimal stimulus producing the sensation of light ; and a specific or colour threshold, the minimal stimulus producing the sensation of colour. The interval between them is known as the colourless or photochromatic interval. It depends upon the scotopic visibility of .the given light below the threshold of photopic vision, and varies with the condition of adaptation and the nature of the light stimulus^^ The colourless interval increases with in- creasing dark adaptation, and this is due to lowering of the general ^ Ztsch. f. Psycliol. u. Physiol, d. Sinnesorg. v. 145, 1893. - Nature, XLvm. 517, 1893. » Kcinig, p. 373. * Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxv. 226, 1901. * Wiedemann's Ann. LXii. 14, 1897. '^ Pogg. Ann. xxxvi. 494, 1835. "> Wied. Ann. xxxii. 25G, 1887. ** Physiologic der Netzhnuf, p. 41, 1865. 9 Proc. Phys. Sac. xiii. 122, 1894. ^° Colour Vision, Tyndall Lectures, p. 35, 1895 1^ Nagel and Schiifer, Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxxiv. 271, 1904; Loeser, Ibid, xxxvi. 1, 1904. SCOTOPIA OR TWILIGHT VISION 61 threshold, the specific threshold remaining almost or quite constant. As regards the nature of the light stimulus the colourless interval is greatest for light of short wave-length and least for light of long wave- length. In the orange it is small even with good dark adaptation. In the red beyond 670 /x/x it is almost completely abolished. In fact even with very good dark adaptation such a red light excites the sensation of red^, and the only evidence of a colourless interval is the alteration in the character of the sensation as dark adaptation becomes more complete, the red disappearing or becoming paler and brighter according to whether its wave-length is greater or less than about 670 jxfji. A minimal colourless interval can, however, be elicited under suitable conditions^degree of dark adaptation, size of field, paracentral or peripheral stimulation (Charpentier)^ {v. p. 81). The general light-threshold has been held to meet the requirements of a physiological unit of luminosity, for it has been accepted as an axiom that at their achromatic scotopic thresholds all lights are of equal brightness. " The light which can just be perceived has always the same brightness, no matter what be the light which acts as the stimulus" (Konig^). Abney and Festing* early recognised the importance of investi- gating what are here called the chromatic and general thresholds for different parts of the spectrum. Light from a monochromatic beam was matched with that from a white beam, both being altered in intensity by the use of rotating sectors (episcotister), or in Abney's later experi- ments by the annulus (p. 5). Accurate photometric observations of the comparison light were made and the experimental variations were calculated from the aperture of the sectors or the scale of the annulus. The " extinction of colour " was measured thus. The light from the D line was taken as a standard and was arranged so that the illumina- tion on the screen was 1 foot-candle. The intensity of the monochromatic light was then reduced by the annulus or episcotister until it appeared colourless and matched the comparison " white." The amounts of reduction necessary for various monochromatic lights throughout the spectrum were thus obtained. From these results the amounts of ^ Pai'inaud, Compt. rend, dc VAcad. Jr., 286, 1881 ; Konig, p. 144 ; v. Krics, Ztscli. f. Psychol, u. Physiol, d. Sinnesorg. ix. 86, 1896. 2 Arch, de Physiol. 1877 ; Arch. d'Opht. xvi. 337, 1896 ; Hering, Arch. f. d. ges. Physiol. LX. 535, 1895; Koster, .4rc^./. Ophth. xli. 4, 13, 1895; Tschermak, Arch. f.d. ges. Physiol. Lxx. 320, 1898 ; v. Krics and Nagcl, Ztsch. f. Psychol, u. Physiol, d. Sinntsorg. xii. 15, 1896. ' Konig, p. 190. « Phil. Trans. Roy. Soc. Lond. CLXXxm. 537, 1892. 62 COLOUR VISION reduction necessary when the illumination on the screen by each monochromatic light was 1 foot-candle were calculated. These results give the actual illumination at the point of extinction of colour in terms of foot-candles. Thus, when the illumination with SSN'^ 50 {558[x[x) was 0"0016 foot-candle at the screen the colour just dis- appeared, and so on. Fig. 19 shows one of the curves obtained. S (0 15 20 25 30 3S 40 'iS 50 SS 60 Fig- 19. " Extinction of colour " curve. Abscissae, wave-lengths of the prismatic spectrum ; ordinates, intensity of the illumination in candle-feet on the screen when the colour just vanishes. (Abney.) In the " extinction of light " a curve was plotted showing what was the fraction of the beam from each part of the spectrum which was just invisible. The absolute luminosity of each part of the spectrum was determined, and from these values a second curve was plotted with ordinates representing the absolute luminosities at the points of extinc- tion. The second curve shows the illumination in fractions of a foot- candle at the screen by each monochromatic light which would be just invisible. The branching of the curves (Fig. 20) beyond the green towards the violet end is due to macular pigmentation, the " whole eye " curve showing the intensities when the eye was allowed to wander. If a curve is plotted from the reciprocals of the extinctions the " per- sistency curve " (Abney and Festing) is obtained.. This is obviously the luminosity curve of the spectrum at the points of extinction. It is 1 SSN is the scale-number on Abney's arbitrary scale of the spectrum. SCOTOPIA OR TWILIGHT VISION 63 identical with the luminosity curve of a feeble spectrum when the maxima are made equal. Fig. 21 shows such a curve. Such are the conditions of the chromatic and general thresholds in the dark-adapted eye. What are the conditions of extinction of colour and light in the photopic eye ? This difficult problem has been 150 140 \ ' 1 ^ ^ ^ >» I i? ^ l^ ■ ^ 1 1 s\ Qj , i5i .1 \c ■> \\ 1 s^l.'^ -^ \ •J Il^ > \ c ^ >- \ 'tJ 1 Kf I '^ f^, v V \ > ^ \ ^ ; ,^ ( V \ ^J 1 ^ I N V \ \ t^ (i^ , ^ 's \ :- V \ \ i<<: '/'' f ; ^ '^ ^J ^ \ \ f t J • • / 'N ■ j / V ^ \ o s V \ / • ^ / / <: \ ^ \ V N s. V t^ ^'? r i / / — :\ t 'h< yk IV %', 7t itn >> % >e J ^' ' / 'A ;■'■' y / is -- -- X ^^ S^j \. V, \i- y / / "v V >> T X / / C 1 1 1 _J t^ G 1 \Y 90 80 70 60 50 40 30 20 10 0 2 +■ 6 8 10 12 14 16 18 20 ?i 24 26 28 30 32 54 36 38 40 42 44-^ 4& 50 52 54 56 58 60 62 64 66 Scalp of Spectrum Fig. 20. " Extinction of light "curve. Abscissae, wave-lengths of the prismatic spectrum ; ordinates, the fractions of the beams which are just invisible, the unreduced illumina- tion of tlie D light at the screen being one Hefner-foot (0-9 candle-foot). Tiie dotted lines show the proportion on the assumption that each beam has the same intensity as that of the D light. attacked for the general threshold by iVbuey^ in the following manner. A disc of matt white paper, 4 inches in diameter, was illuminated by white light. Through a hole in the centre, .| inch in diameter, a coloured light was viewed. The fovea was thus stimulated with monochromatic, ^ Proc. Roy. Soc Lond. A, lxxxvii. 1912. 64 COLOUR VISION and the parafoveal zone with white light, of known intensities. A small spot of shadow was thrown upon the white disc from which the para- foveal zone was illuminated. The blackness of this shadow is the same, irrespective of the intensity of the white light, though it appears to be different according to that intensity. The monochromatic light was then reduced until the spot illuminated by it matched the shadow spot in blackness. , f \ u K HS ^ 'Ul ■i^e ^ \^ ■ \ i / \ / ^ \ \ -J / r > y \ \ ,5, / \ \ P t \ ^ / = \ / \ ^ \ / '"' \ \ —J / V f 1 , \ / / 1 '<^ 6 \ \ \ ^' \ J \x \ h' ' / , / // 1 0 1 \ 1 \ 1 '■^} A S 1 \ 1^ ^"7 •A 7^ \ M \ "^ .( .^ V \, h ^'^ / 'i ■ k \ \ 1 ^ 7 v^ \ / 1 u \ k. { ^ / / •\r/^ - - A \ .^ >^ '^l\ '^t' rF % >\ J 1 \ y <^^ \ ^^ / / - \^' 5*- ■ V ^ \ %, y r \ / '1 a"' 7 N V. A >' J k > / / iS ib^ -J J \ s. i; \d - -^ '^ Li -^ ^^■ -f" \.\f^ 1 u s V -,c 1 -ll 1 ■^ 6 i 1 0 1 2 1 * 1 5 1 3 2 0 ? I 2 4 1 6 I 9 3( ) 3 2 3 4 3 6 3 S 4 0 4 2 4 1 4 0 4 S 5 3 5 2 5- <- 5" S 5 8 6 D 6 2 6 4 %i 110 100 90 &0 70 60 50 40 30 20 10 0 V 14 Ij" 13 £0 (I 24 i-b i6 M iZ 34 jb 38 40 42 41 40 43 50 b2 M- :3b 5» fcU bi? b4 bb Fig. 21. " Persistency curve " of a deuteranope. The values are the reciprocals of the " extmction of light " values, which are also shown. The " normal luminosity curve " is the curve for higher intensities of light (the photopic luminosity curve). The persistency curve is the achromatic scotopic luminosity curve (compare Fig. 14). (Abney and Festing.) The illumination of the white area was modified by the annulus so that intensities from 0*2 to 0'00078 foot-candle were obtained. Fig. 22 shows the logarithms of the intensities of the light for different wave- lengths for extinction with different strengths of illumination of the parafoveal zone with white light. Fig. 23 shows the corresponding persistency curves. These, as before pointed out, give the luminosity curves of the spectrum at the points of extinction, and we see that they correspond with Konig's SCOTOPIA OR TWILIGHT VISION 65 and other luminosity curves. With 0'2 foot-candle of parafoveal illumination the curve is the luminosity curve of the photopic eye. Fig. 22. " Extinction of light " curves for different degrees of light adaptation. Abscis- sae, wave-lengths ; ordinates, logarithms of the intensities of the light. (Abney.) 100 Fig. 23. " Persistency of light " curves corresponding to the " extinction of light " curves for different degrees of light adaptation. Abscissae, wave-lengths ; ordinates, reciprocals of extinction values. (Abney.) As the parafoveal illumination diminishes the curve shifts towards the violet end, so that the scotopic curve for complete dark adaptation is finally reached. p. 0. V. 5 66 COLOUR VISION It has been found in experiments on the colour threshokl that if the colour is just extinguished very slight illumination of the retina with white light will cause it to reappear without making any alteration in the intensity of the coloured light. Boswell^, working under Nagel's direction, showed that the same phenomenon occurred on the addition of any light of considerable scotopic value, such as green. Since the eye is dark adapted the summation of the colourless scotopic values of the coloured lights throws the value of the stronger above the chromatic threshold. It is not a question of contrast, as is shown by the action of white light, nor of complementary colours. The same fact is borne out by Abney's experiments, where the extinction intensity for retinal illumination of 0*2 foot-candle is 2 '6 times that of illumination of 0-00078 foot-candle. ^ ZLsch.f Sinnesphysiol xli 364, 1907. SECTION IV REGIONAL EFFECTS CHAPTER I THE FIELD OF VISION FOR COLOURS When we look at or fixate a particular object we are conscious of the presence and movement of other objects situated around the object fixated. The image of the object fixated is formed at the fovea, those of surrounding objects at various points in the peripheral parts of the retina. The area over which such outlying objects can be perceived is called the field of vision, which is therefore the projection outwards of all the points upon the retina which can initiate visual sensations. Its limits are usually plotted upon a chart by means of a perimeter^. The normal photopic field of vision for a well-illuminated white object, 10 mm. square at a distance of 45 cm., is a horizontally oval area extending upwards about 50°, outwards 90°, downwards 70°, and inwards 60°. It varies with the illumination, the size of the test object, the contrast of the test object with the background, and the state of adaptation of the eye^. The fields for colours are smaller (Purkinje) : by the ordinary clinical methods those for blue and yellow pigmented objects are about 10° smaller than that for white ; those for red and green 20° smaller, the red field being usually smaller than the green, and violet smallest of all. ^ See Parsons, Diseases of the Eye, p. 1(50, 2iid ed. 1912. - See Landolt, in GraeJeSuemisch Hand. d. (jes. Augenheilkunde, iv. 1, 0-48, 1904 (witli Bibliography). 5—2 68 COLOUR VISION Charpentier^ obtained the following results with coloured paper of 20 sq. mm. Violet Blue- Yellow- Red Orange Yellow Blue green green Inwards 35 38 45 55 60 67 70 Downwards 33 43 49 58 59 62 60 Outwards 45 52 60 65 70 78 78 Upwards 34 39 44 53 58 59 59 Hess's field for invariable red and green {vide infra, p. 70) with an object subtending 3° was 21° inwards, 43° outwards, 17° upwards, and 14° downwards. The fields for colours are approximately concentric with that for white. The colours change in appearance in passing from the point of fixation towards the periphery. Those at the red end of the spectrum pass through yellow to grey ; those at the violet end through blue to grey. Blue-green becomes green, then yellow-green, then yellowish white (Hering). Baird'^, who has paid particular attention to this point, finds that the changes in colour in passing from the periphery towards the centre are as follow. Red first appeared yellowish, then passed through yellow, orange and orange-red before it finally appeared red. Orange first appeared yellow, becoming more and more orange. Yellow appeared yellowish and gradually increased in saturation. Green appeared yellowish at first, gradually increased in saturation, assuming a greenish and finally a green tint. Blue became more and more saturated. Violet appeared bluish, then blue, and finally violet. Purple gave the longest and richest series of transitions : beginning with yellowish, its tone gradually moved down the spectrum, passing through orange-yellow, orange, red, and purplish red, before the pure tone finally appeared. Aubert^ found the limits of the coloured fields determined by the intensity of the light, and this result was confirmed by Landolt, Char- pentier and others. The periphery is therefore dyschromatopic rather than achromatopic (Charpentier). The diminution in intensity, as measured by the distance of visibility of coloured spots, is given in Fig. 24. The sudden curve in the blue at the macula is due to macular pigmentation and also probably to dark adaptation. 1 La Lumiere d les Couleurs, p. 193. 2 The Color Sensitivity of the Peripheral Retina, AVasliington, 1905. ^ Physiologie der Netzhaut, p. 116, 1865. THE FIELD OF VISION FOR COLOURS 69 The limits of the colour fields vary not only with the intensity of the light, but also with the saturation of the colour, and, above all, the size of the object. If these are sufficiently great colours may be recognised, almost, if not quite, at the extreme periphery (Bonders, Landolt). As already mentioned only the foveal region gives the unadulterated photopic reactions, unless the eye is very completely adapted to light, so that all traces of scotopia are eliminated from the peripheral field. 70 Fig. 24. Perception of colours in different parts of the field of vision. Abscissae, de- grees to the nasal side of the fovea {O) ; ordinates, maximum distance at which the colour is perceptible in a coloured paper 2 mm. square. (Charpentier.) Ordinary observations with the perimeter do not afford accurate details for comparison. If light adaptation is rendered as complete as possible by exposure to bright sunlight many points of interest are elicited. Under these circumstances it appears — within the limits of experimental error — that colour matches, spectral or composite, which hold good for the fovea remain good when viewed eccentrically, but though the matches remain matches the values alter, the colours changing in the raid-peripheral region, and becoming colourless in the extreme periphery. It may therefore be concluded that in the phofopic eye peripheral vision differs from central vision only in the direction of a diminution in sensi- bility, and not in the direction of a change in character of sensation. Further, all colour mixtures which appear colourless by central vision remain colourless by peripheral vision. Allowance must, however 70 COLOUR VISION ^ be made for macular pigmentation {vide infra, p. 124). The deduction is therefore more accurate if paracentral and peripheral regions are compared^. Care must be taken that light adaptation is complete and has been induced by colourless light, and that long exposure to coloured lights is avoided. Ole Bull^ was the first to employ stimuli the saturations and lumi- nosities of which were equalised and to discover colours which do not I change in hue in peripheral vision. His " physiologically pure " j'?^ colours were a purplish red, a bluish green, a yellow, and a blue. He found that the limits of the purple-red zone coincide with those of the blue-green zone, and that those of the yellow and blue approxi- mately coincide. The yellow-blue zone is considerably wider than the red-green zone. He employed colour papers. Hess^, also using coloured papers, similarly showed that with light adaptation colour pairs of fixed size, intensity, and saturation can be selected which lose their colour simultaneously and form a grey match when viewed peripherally, viz. green (495 fjifju) and red mixed with a small amount of blue; and yellow (574*5 /a/a) and blue (471 /u,/x). These colours are complementary. All colours of greater wave-length than 549 fjLfjL approximate yellow, all of less wave-length blue, in passing from central to peripheral vision. Thus orange and green-yellow become yellow ; blue-green, violet and purple become blue. Hence there are only four colours which gradually become paler without altering their colour tone, ultimately becoming colourless, as one passes from central to peripheral vision. Hess calls these colours " invariable " red, yellow, green and blue. The limits of the zones at which the colours become colourless vary with conditions already mentioned, but the limits for invariable red and green are the same, as also for invariable yellow and blue. Hegg^, Baird, and Dreher^ confirmed Hess's results. Baird found the " stable " or invariable colours to be a purplish-red, a yellow (about 570ju./i), a bluish-green (about 490 fxf^), and a blue (about 460 fiix). Dreher's values for the three last were 568 /z/a, 483 /^/x, and 461 /x/x. Baird found that the coincidence of the zones of each pair of stable colours holds good for the dark-adapted as well as for the light-adapted 1 Hess, Arch. f. Ophth. xxxv. 4, 1, 1889 : v Kries, Ztsch. f. Psychol, u. Physiol, d. Sinnesonj. xv. 26.3, 1897. 2 Arch.f. Ophth. xxvn. 1, .54, 1881 ; xxix. 3, 71, 1883. 3 Ihirl. xxxv. 4, ]. 1889. * Ibid, xxxvm. 3, 14.5, 1892; Ann. d'ocul. rix. 321, 1893; cxi. 122, 1894. * Ztsch. f. Sinnesphysiol. xlvi. 1. 19]]. THE FIELD OF VISION FOR COLOURS 71 eye. He used gelatine filters which transmitted approximately mono- chromatic lights. The yellow-blue zone is " dichromatic." Beyond this zone there is an extreme peripheral zone which is " monochromatic " or totally colour-blind. It is best demonstrated in the nasal and upper and lower portions of the field, but only with very small test objects in the peripheral portion. If the luminosity curve for different colours is worked out for this zone it is found to be quite different from the achro- matic scotopic luminosity curve, v. Kries^ determined the peripheral luminosity values and compared them with Nagel's corresponding values for the achromatic scotopic luminosity curve : Wavelengths 680 651 629 608 589 573 558 Periphery values 0-6 37-5 77-5 101 100 796 52-2 Scotopic values ? 3-4 14 0 355 100 256 351 The curves are shown in Fig. 25. 530 512 28-5 14-6 321 198 / V 3bU — J 32 5 < 275 250 ' ^ -. 22 5 ,' \ 20 0 17 5 IbO 125 -r — 100 75 50 ^ ^ 7 *-^ — ■ — / ^ / ■^ — 25 y — ..- ■■'' — 680 651 629 608 589 573 558 530 513 Fig. 2a. Photopic luminosity curve for the totally colour-blind peripheral zone of the retina Scotopic luminosity curve. Abscissae, wave-lengths of the prismatic spectrum of gas light; ordinates, arbitrary scale, (v. Kries.) The peripheral luminosity curve is highest at about 608 /m/x instead of 544 ju.^, thus nearly approximating the luminosity curve for the fovea. We have here a further proof that peripheral vision is to be regarded as central vision diminished in sensibility, whereas scotopia is a difTerent form of vision ; the former is a quantitative variation, the latter qualitative. ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesortj. xv. 247, 1897. 72 COLOUR VISION The dissimilarity of the peripheral and achromatic scotopic luminosity curves shows that Hess's experiments do not prove all that he attributed to them, for he chose colours of equivalent " white value " and stated that they remained good matches both in peripheral photopic and in scotopic vision. The matches are valid only for peripheral photopic or scotopic vision. On comparison of the peripheral photopic luminosity curve with Konig's and Abney's luminosity curves, making allowance for the fact that the two former are with gas light, the latter with arc light, we see that they agree, so that the statement on p. 71 is correct. Polimanti^ has confirmed the statement on the same individuals using the same apparatus and source of light with the flicker method (Fig. 26), 95 T\ ^N 85 / / i/ s N 75 1 1 ! ^^^ \ 65 / ■ii / I r \ [\ 55 / / / \ V. 45 ^'' / X^ V 35 J 0 V -\ iil 25 ^1 1 \ **i^ ^^ 15 t 1 5 687 1 — 664 642 624 1 606 1 589 565 ' 1 — 1 543 ' 1 1 534 . 509 A'd-2-5 /Va-2A'4-T5/l'a-1 A'A-0-5 Na. Na*^ Na-il /Vd+3 Na.*', Fig. 26 Photopic periphery luminosity curve (flicker metliod). Pliotopic central luminosity curve (flicker metliod). Photopic periphery luminosity curve (equality of brightness method). Abscissae, arbitrary scale of wave-lengths of prismatic spectrum of gas light, measured from the D line {No), with certain absolute wave-lengths indicated ; ordinates, arbitrary scale. (Polimanti.) Siebeck^, using minimal fields, has adduced further confirmation. He found that with spectral lights of moderate intensity it was impossible with accurate foveal fixation to abolish the sensation of colour without that of light by reduction of the field, but that this was readily effected 1°'5 from the centre. He compared the luminosities of the lights under 1 Ztsch.f. Psychol, n. Pkyftiol. d. Siiinesorcj. xix. 272, 1890. Ibid. XLi. 89. 1907. THE FIELD OF VISION FOR COLOUES 73 these circumstances and found that these " minimal field luminosities " agree with both the ordinary and the peripheral photopic luminosity curves, and therefore disagree with the achromatic scotopic luminosity curve. He found the maximum luminosity by this method at 601*3 fifj. : Wave-lengths 642-4 620-8 607-8 601-3 595-3 589 Minimal-field luminosities 66-() 91-5 112-8 128 110-9 100 Wave-lengths 5791 574-1 564-8 551-4 Minimal-field luminosities 81-6 79-5 68-6 52 90 The minimal -field and luminosity curves are shown in Fig. 27. IJU I 20 / \ / \ 10 / 1 .,\ /' \ 100 r s 11 / } •^ I h / / \ 1 QO /; ■ / / \ V ^. ^ 80 1 1 it / x yi V s \ / f ,^ k \ 70 f / 7 f / V \, \ / I'l/ v> S, \, / 1 !jl V, \ N i 60 / // \ •^ \ ^. 50 / / 1 1 s s. / f i * > N > I*. "♦O / '•! \ ^^ \ N / / V s \ 30 /■/ ^• -x^ N \ f •>, ^ ^ ^ N ■-*, 20 "-. N ^ ~^ =^ - •■^ ^ ^■■4 10 ^;-- ^^ ■—I — I — I — I — I — I — I — I — I — I — I — I — I — I — I I 1 — I — I — I — I — 1 — I — I — I — I — I — I . ~2 1,8 1.6 1,1 U 1 0,8 Q6 0.1 0.2 0 0,2 0.<. 0,6 0,8 t 1.2 XH 1,6 1.8 2 Z'J. 2.t 2,6 Zs 3 3.2 i.t i.6 i& ^ h o co" I 5_ ID I: J: in CM iT) 3. o Si Fig. 27. "Minimal-field luminosity curve." Photopic periphery luminosity curve. — Photopic central luminosity curve (flicker method) " Il'-curve " of a deuteranope (see Fig. 45). (Siebeck.) 74 COLOUE VISION Zalin^ has investigated the " minimal-time kirainosities.'' In these experiments the coloured light is exposed momentarily, when it appears colourless, and is compared in brightness with a surrounding white surface. The minimal-time luminosities completely confirm the results obtained by minimal fields and periphery values, as is shown by the following values : Wave-lengths 659 621 601 589 564 542 523 506 Minimal-time luminosities 22 5 79-0 105-0 100 76-6 59-4 380 15-2 Minimal -field luminosities 24-4 702 1048 100 74-4 58-3 37-5 20-2 Periphery values 21-6 73-9 99-6 100 79*9 54-7 36-7 15-2 Hess^, Hering^, and Tschermak* have examined the relative lumin- osity of different lights in different regions of the retina. Hess found that red and green pigments on a grey background, when viewed by dull daylight with moderate dark adaptation, became darker and brighter respectively in peripheral as compared with central vision. Hering confirmed this result with spectral lights. Tschermak used Hering's " double room " {v. p. 58) with spectral lights from Auer-gas or arc light against a daylit background. He found a relative increase in luminosity from 516 to 466 /x/x, a relative decrease from 693 to 525 ju/x, and no change from 525 to 516 /x/x, on indirect fixation. In the dark-adapted eye there is similarly a change in relative luminosity, and it occurs in the same sense as the Purkinje phenomenon. Hering^ has shown that for the dark-adapted eye, even on momentary dark adaptation, lights of equal brightness by direct fixation appear very unlike in luminosity on indirect fixation. He describes these alterations in luminosity as Purkinje's phenomenon produced by change of position in the visual field without change of light intensity. The limits of the colour fields of the partially dark-adapted eye for spectral colours have been worked out very thoroughly by Abney*'. 1 Ztsch. f. Sinnefsphysiol. XLVT. 287, 1911. ^ Arch.f. OpJdJi. xxv. 4, 1, 1889. 3 Arch.f. d. gcs. Physiol, xlvii. 417, 1890. 4 Ibid. Lxxxii. 559, 1900. "• Ihid. Lx. .'".19. 1895 « Abney, p. 190; Phil. Trans. Roy. Soc. cxr. 155, 1897. THE FIELD OF VISION FOR COLOURS 75 Fig. 28 shows the fields with red . . . . 670-5 /t/x 0-27 foot-candle yellow .. 589-2 „ 3-95 green .. 508-5 „ 1-89 blue .. .. 460-3 „ 0-36 RIGHT. 0 15 30^VvvV ,WV u^'^^^^t''^^^^'^^ 6s- ;{"7""/'m|Nifm,in.i< ,<^ 120 ,al. 1^****.,,^^ ^ ^^"•H^^^ ~^ §^ Blue Yellow Red. Green . Blue. Ynlow. Red. Gi-een. Intensity Fig. 32. The temporal and nasal limits of the fields of vision for four spectral colours at different intensities of illumination. (Abney.) For colours Abney^ found that between apertures subtending 4° 28' and 10' the fields decrease in extent in such a manner that for each diminution in aperture to half the diameter the diminution in field is 5° on the temporal and 4° on the nasal side. 1 Arch. d'Opht. vu. 13, 18S7. '^ Abney, p. 208. 80 COLOUR VISION This is the same ratio, 5 : 4, as for diminution of intensity of light by quarters (p. 79). The diminution by each quarter of the area is thus equivalent to a quarter of the intensity of light. Any a'perture subtending more than 5° will give the same field. Guilleryi confirmed Hess's results by using the size of the area stimulated as a guide to the size of the colour fields. For each colour a gradual increase is necessary in passing towards the periphery. The increase for yellow and blue and for red and green agree respectively, but of course the red-green increase differs from the yellow-blue increase. 50r^ 0 I 2 Z '}■ 5 6 Aperture in powers of - Z Fig. 33. The temporal ( ) and nasal (- ) limits of the fields of vision for four spectral colours for different areas of stimulation. (Abne}'.) The general threshold of light at the periphery is difficult to measure. Abney used a small spot of luminous paint for fixation. At 10° from the fovea he found that the extinction of all light from red light takes place when the light is about one-third brighter than is required for the fovea. With white paraffin light it is somewhat less. With green light at about the E line and with blue at the Lithium line the necessary reduction of the light is greater than for the centre of the eye. The photochromatic interval is greater for the peripheral than for the central part of the retina. ^ Ztfich. f. Psychol, u. Physiol, d. Sinnesorg. xii. 267, 189fi THE MACULA LUTEA AND FOVEA CENTRALIS 81 CHAPTER II THE MACULA LUTEA AND FOVEA CENTRALIS The Macula Lutea. The chief feature of macular vision is the unequal absorption of different rays by the pigment of the yellow spot. It has been necessary to refer to this disturbing feature at an earlier stage (pp. 40, 70). The Fovea Centralis. The photopic luminosity curve for the fovea is shown in Fig, 11, and calculations show that it is about one-sixth more sensitive to the D light than the macula (Abney), To the green and blue it appears to be less sensitive, and Konig^ even says it is blue-blind. This observation is certainly untrue and is probably due to the low scotopic luminosity value of the fovea combined with the absorption by the pigment which probably extends to this region. We have already referred incidentally to many features in which foveal vision differs from paracentral and peripheral vision. With its exceedingly high sensibility for form we have little to do here, but it should be borne in mind. More important for the present purpose are its high photopic chromatic sensibility, and its very low scotopic sensi- bility. It has already been mentioned that for white light achromatic scotopic sensibility is proportional to the area stimulated (p. 52 ; see also Section VI, Chap. i). For the fovea, in order to produce a luminous sensation, the total quantity of light, i.e., the product of the area and the illumination, must attain a certain value, and this value is constant for a given condition of adaptation. Since the colourless interval increases in dark adaptation owing to lowering of the general threshold, the chromatic threshold remaining almost or quite constant {v. p. 60), and since dark adaptation increases the foveal sensibility little or not at all, one might expect a very great diminution or even absence of the photochromatic interval at the fovea. Charpentier^ demonstrated a photochromatic interval for colours at the fovea, and Koster^ and Tschermak^ confirmed the results, thus differing from A. E. Fick^, Parinaud^, and Konig and v. Kries. If it exists for 1 Konig, p. 353. ^ ^rch. cCOpht. xvi. 337, 1896 ; La Lumiere d /(.*■ Couleur«, p. 2U(i. 3 Arch. J. Ophth. XLi. 4, 13, 1895. ^ Arch.f. d. ges. Physiol, lxx. 320. 1898. =^ Ibid, xuii., 48L 1888. « Ann. d'ocul. cxii. 228, 1894 ; Arch. d'Opht. xvi. 87, 1896. P. 0. V. 6 82 COLOUR VISION the fovea it is much less than for the periphery. The point is one having a very practical bearing and some experiments made by Gotch^ may be quoted. " If the eye is fixed in the dark room on one small spot of light (A) and a second small coloured light (B) is flashed out at some little distance, then the sensation produced by A varies with the position and hue of B. "KB is spectral green or blue, and is some degrees away from A (A being fixed by the visual gaze and thus in the centre of the field of vision), B gives rise to a sensation of rather dazzling white. This is especially the case when the light itself is of very small angular area. The sensation continues when B is moved nearer to A, and tends to mask any recognition of green or blue colour. In my own case such colour recognition may not occur with the Board of Trade light green light until this is focussed on the retina within 1|° to 2° from the centre when the dazzle disappears and the colour becomes quite plain. " If B is red, then neither the colour nor the light itself is recognised when it is so situated as to be focussed on the retina over 8° or 10° outside A. In my own case with a dim Board of Trade light red, the limit of failure of colour recognition was a little under 6|°. Beyond such recognition limiting distance there was no sensation of white dazzle, but a barely perceptible dull grey spot was sometimes seen instead of the red light. Further out this light became quite invisible. When brought as near as 6° it always showed as a distinct red spot, and continued so however close it was brought. There was no simul- taneous white dazzle to mask the hue ; thus the light, if visible was easily recognised. " A number of experiments on individuals ascertained to possess normal colour vision support the view that, in the dark-adapted eye red light is recognised as red over an area whose radius is three or four times that observed with green light ; yet the red light is not seen at all outside this larger area. On the other hand green (or blue) light, whilst it is only recognisable as green over the much more restricted central area, is seen as a bright light of a dazzling white type over a very extensive area. " As illustrations of such visual phenomena in connection with what is termed ' dark adaptation perimetry,' I append a few experimental results obtained from ten different observers, all with normal colour vision. ^ Report of Departmental Committee on Sight Tests, Appendix 3, 152. 1912. Red Light (X G563) Green Light (\ 51 G9) With in 6° or 7° Within U° or 2° JJ 7J° » sr ir »2r a 9° ., 10° 2h° „Sh° 6° „7° 2° „2h° ,, r „8° 2° „r 39 8° „y° 2° „2r J? 8° »n° 2J° „r »? 8i° „ 91° 2i° .,3° J> 6° „7r 2° ,,2^ ?J 8° „9° 2° „2i° THE MACULA LUTEA AND FOVEA CENTKALIS 83 Number of Angular Distances from Centre within which the Colour Observer of a Spectral Light 30" Diameter was recognised 1 2 3 4 5 6 7 8 9 10 " It may be added that with the red spot of light any recognition of light immediately beyond the sensitive limit given above was very doubtful or non-existent, although with practice a dull grey spot can be perceived for a short distance. On the other hand, with the green spot of light there was, from 3° outwards, most definite recognition of light over a very large area, the appearance being a white or bright dazzle. This was also present with blue light, but the hue was unrecog- nisable as regards both blue and green. Yellow light also gave the same bright dazzle ; it was often blended with a reddish sensation if it lay from the centre about 3° or 4°. " The recognition of small areas of red or green by the dark-adapted eye is, as regards colour, thus only possible when these are focussed near the fovea, i.e., lie near or at the centre of the field of vision. This is particularly the case with the peripheral vision of green ; presumably these rays, by exciting rods, evoke a sensation of white, which has a dazzling effect and masks the true colour. Such white dazzling sensa- tion is still present, but to a less degree, as the green light approaches the centre of the field of vision, but it ceases when this light is 1° to 2° from the centre. According to v. Kries the explanation is the differential stimulation of both rods and cones, that of the rods giving rise to the dazzling white sensation in addition to the colour sensation produced by the stimulation of the cones." Although the fovea is night-blind compared with the periphery it is capable of some degree of dark adaptation. It is necessary to have very strong preliminary light adaptation, as from looking at the clear sky, in order to demonstrate it (Nagel and Schafer^). Tschermak^ also obtained foveal dark adaptation and thought that it was slower ^ Ztsch. f. Psijchol. u. Physiol, d. SinnKSorg. xxxiv. 271; 1904. - Arch.}, d. ges. Physiol. Lxx. 297, 1898. 6—2 84 COLOUR VISION than peripheral, but he used too large test objects and thus exceeded the foveal limits (Nagel). More recently a slight increase in the foveal sensibility on dark adaptation has been proved by Wolfflin^ and by Dittler and Koike'^. In the experiments of the latter one eye only was dark-adapted and the luminosities of the binocular double images of the illuminated object were compared. The adaptation at the fovea is so slight that one would not expect to obtain Purkinje's phenomenon there under ordinary circumstances. The demonstration of its absence is beset with difficulties — accurate foveal fixation, comparison of bright- ness of very small areas, etc. — and opinions are therefore divided. Tschermak^, Koster* and Sherman^ are in favour of its presence, but their methods were less satisfactory than those of Konig^, Lummer'^ and Nagel and v. Kries**. v. Kries and Nagel have proved fairly con- clusively that Purkinje's phenomenon is absent over a central field not greater than 2°, i.e., not exceeding the rod-free area. The longer the dark adaptation the better was the delimitation of the area. If a red and a blue spot are fixed with this area, the red being brighter than the blue, then on diminishing the intensity of the light the relative luminosity remains the same. On the slightest deviation of the eyes, however, the blue at once appears the brighter and less saturated. Hering^ found that small fields of red and green, equally bright when fixed centrally, become of different brightness when fixed eccentrically (" Purkinje's phenomenon by change of position in the visual field without change of the intensity of the light "), and he also stated^" that the deviations from Newton's law " become less marked the smaller the area of the retina stimulated, and are absent when the field is made sufficiently small." 1 Arch.f. Ophth. Lxxvi. 464, 1910. ~ Ztsch. f. Si7inrsphi/sloL XLVI. 166, 1912. 3 Arch.f. d. ge.s. Physiol, lxx. 297, 1898. * Arch. J. Ophth. xli. 4, 1, 1895. » Wundt's Philo.-i. Stud. xin. 1898. « Konig, p. 338. ^ Verhandl. d. Deutschen. physik. Gcscllschft. vi. 2, 1904. ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxili. 167, 1900; v. Kries, loc cit. ix. 81, 1896 ; Arch.f. Ophth. xlii. 3, 95, 1896 ; CcntralU.f. Physiol. 1896. 9 Arch. f. d. ges. Physiol, lx. 533, 1895. " Ibid. Liv. 277, 1893. 8ectio:n y TEMPORAL EFFECTS CHAPTER I RECURRENT VISION ; THE TALBOT-PLATEAU LAW ; THE FLICKER PHENOMENON We have already had evidence (p. 57) from stimulation of the dark- adapted retina by light of low intensity that two mechanisms are involved. One, which may be called the " scotopic mechanism," in the condition of maximum dark adaptation, responds to light of low intensity by a colourless light sensation, no matter what the nature of the light stimulus so long as it be an " adequate " stimulus {v. p. 19). The other, which may be called the " photopic mechanism," responds to light of higher intensity (which in itself induces a relatively high degree of light adaptation) by a light sensation of greater intensity and greater complexity, the sensation being that of white or colour according to the nature of the stimulus. It may at once be admitted that theoretical conceptions of definite mechanisms subserving these diverse though allied functions are here introduced. They will be allowed to intrude as little as possible in the description of the facts, but the comprehen- sion of the description is facilitated by permitting their use. The sensational response to a single short-lived stimulus is not a single, equally short-lived light sensation. Except in the case of a very feeble stimulus it is "a series of pulses of sensation of diminishing intensity rapidly succeeding one another^." Attention was early called to these recurring responses by C. A. Young in 1872^. He found that when a discharge from a powerful electric machine momentarily illumi- nates a room the objects may be seen, not once only, but two, three or four times in rapid succession, although the spark is single. He called 1 MoDougall, Brit. JI. of Psychology, i. 78, 1904. - Nature, v. 512, 1872 : Philos. Mag. XLin. 343 ; SiU. J I. iii. 262. 86 COLOUR VISION the phenomenon " recurrent vision," and the observation was con- firmed by Shelford Bidwell^. An allied phenomenon was described by Purkinje^, rediscovered by A. S. Davis^, and is commonly called "Bidwell's ghost"; Hamaker^ called it the "satellite", v. Kries "the following image." The ghost is seen in typical form when a light of moderate brightness is moved across a dark field of view at a moderate speed while the eye remains at rest, appearing as a second dimmer image following the primary image after a short interval, usually about -^sec, of darkness. The subject has been specially studied by Bosscha^, Charpentier^, Hess', v. Kries^, Samojloff^ Munki", Exner^i, and McDougalli2. Fig. 34. Fia;. 35. Fig. 34. Appearance of a radial slit 2° in width and 7 cm. in length, with its mid-point 15 cm. from the centre of the disc, rotating at the rate of one revolution per 3" before a glass illuminated by four acetylene-gas jets. (McDougall.) Fig. 35. Appearance with slightly increased speed of revolution — " Charpen tier's bands." (McDougaU.) ' Proc. Roy. Soc. Land. lvi. 1.02, 1894 ; Curioftitict of Light and Sight, London. 1899. 2 Beobachtungen zur Physiol, d. Sinne, ii. 110. 1825. 3 Lond. and Edin. Philosoi^h. Mag. xi.iv. 526, 1872. * Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxi. 1, 1899. 5 Arch.f. Ophth. xl. 2, 22, 1894. « 4^g^_ ./^ physiol. iv. 541, 1892 ; vi. 677, 1896. ' Arch. f. d, ges. Physiol, xlix. 190, 1891 ; Arch. f. Ophth. XL. 2, 259, 1894 ; XLiv. 3, 445, 1897 ; Li. 2, 225, 1900 ; Arch.f. d. ges. Physiol, ci. 226, 1904; ovii. 290, 1905. 8 Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xii. 81, 1896 ; Arch.f. Ophth. XLii. 3, 1896 ; Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxv. 239, 1901 ; xxix. 81, 1902 ; Arch. f. d. ges. Physiol, cm. 167, 1904. * Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xx. 118, 1899. 1" Ibid. xxra. 66, 1900. " Arch.f. d. ges. Physiol, cm. 1905. ^- Loc. cit. RECURRENT VISION 87 The Primary Image. With momentary stimulation by a stationary light McDougall found that a brief stimulus of low intensity provokes a single brief pulse of sensation, and that a stimulus of higher intensity provokes two, three, four or more pulses according to its intensity, and that the extrafoveal region, being more sensitive under dark adapt- ation than the fovea responds to a stimulus of any given low intensity with a rather longer series of pulses than is provoked by the same stimulus when applied to the fovea. If the stimulus is given by a radial slit in a rotating disc the image is spread out into a sector, varying in width according to the rapidity of rotation. This is due to the per- sistency of the sensation beyond the time of stimulation. The sector is not of uniform brightness, but is marked with alternate light and dark bands which vary in intensity and width according to speed of rotation and so on. These fan-like bands are commonly known as " Charpentier's bands\" though they had been previously described by Bidwell (1894). Fig. 36. Diagram illustrating the primary response to a single momentary stimulus. The series of pulses of sensation commences at a and lasts through a period of about f sec, represented by a — i. The height to which each curve rises represents the intensity of the pulse of sensation. The effects of a momentary stimulus of less intensity are represented by a part of the series of curves, e.g., e — i. The overlapping of the curves represents the fact that the intervals between the pulses of sensation are not quite dark. (McDougall.) The Secondary Image. The above is not the only response. In the case of momentary stimulation by a stationary light a secondary image of much less luminosity seems to appear at a great distance at the moment in which the pulsating image comes to an end, and rushing up towards the eye to .stop suddenly, filling the position of the primary I Arch, de Physiol, iv. 541, 1892 ; viii. 677, 1896 88 COLOUR VISION image. The secondary image fades gradually and slowly. In the case of the travelling object the secondary image appears abruptly at a brief interval after the last bands of the primary image, and fills a part or the whole of the track of the image. If the object light is coloured and of high intensity the secondary image is in both cases tinged with the same colour ; but if the object light is coloured and of low intensity the secondary image is grey in the case of blue and green light, but is absent in the case of pure red light. The secondary image is the ordinary positive after-image {vide infra). The interval between the primary and the secondary images is not always completely dark. As the rate of movement of a travelling object light is increased dim grey bands appear in the interval. They are dimmer and broader than the leading bands and differ markedly in quality, the leading bands being yellowish-white, these others a neutral grey. The interval-bands show the following peculiarities : (1) they are brightest when the object light is green, absent when it is red : (2) they are absent where they cross the fovea, even with a green object light; (3) they have the neutral ghostly quality, inclining towards blue, characteristic of the scotopic spectrum ; (4) they are absent when the eye is light-adapted. It is probable therefore that the primary response is a response of the photopic mechanism, any response of the scotopic mechanism being obscured by the preponderant photopic effect. On the cessation of the photopic response the scotopic response becomes manifest under favourable circumstances. There is reason to believe that the latent period of the photopic reaction is shorter than that of the scotopic by about j-g sec. The secondary image is similarly complex. With an object light of low intensity it is a pure scotopic reaction, as shown by its grey quality and its absence with a pure red object light. With a brighter object light it is a combined photopic and scotopic reaction, the photopic preponderating with strong lights, as shown by the colour and saturation of the secondary image. McDougall has further shown that the bright initial reactions which constitute the primary image exert an inhibitory effect upon the im- mediately succeeding reactions. BidweWs gJiost is the last of the series of pulses of sensation, the intermediate members being thus inhibited. In its typical form it is a pure scotopic reaction, but that it is not necessarily so is shown by the fact that it can be obtained by a pure red object light of high intensity and does not then jump the fovea. RECURRENT VISION 89 I have hitherto followed McDougall's excellent paper. It will be readily appreciated that since the response differs in the number and character of the pulses of sensation according to the intensity and nature of the light stimulus, and since the stimulus acts upon two mechanisms of different intensity-response and different latent period, the nature of the phases of the response vary greatly according to the conditions of the experiment. Fig. .37. Appearance of a narrow slit, 2° to 5^ in width, in a large disc, rotating at one revolution per 3", before a milk-glass illuminated with one acetylene-gas burner, as seen by the light-adapted eye which has been kept covered for about three minutes and then fixed upon any point in or near the path of the moving slit — " Bidwell's ghost." (McDougall.) It is not surprising, then, that under appropriate conditions v. Kries found the following series of events : (1) a primary image ; (2) a short dark interval ; (3) a secondary image, complementary in colour to the primary (the satellite or ghost) ; (4) a second dark interval ; (5) a tertiary image (Snellen, Bosscha), faintly tinged with the colour of the primary ; (6) a third dark interval. The secondary image in this series follows the primary by | — | sec. It is absent with foveal fixation. It increases in brightness and extent in the early stages of dark adapta- tion, but is absent after prolonged dark adaptation. It is also absent with red light and its brightness corresponds to the scotopic value of the exciting light (v. Kries). Much discussion has arisen around the colour of the secondary image. According to Charpentier it is violet with low, and colourless with high intensity. Hess describes it as faintly tinged with the same colour as the primary i. With regard to 1 Cf. v. Kries, Ztsch. /. Pmjchol u. Physiol, d. Sinnesorg. xii. 81, ISflO : Hess, Arch f. 90 COLOUR VISION the tertiary image its hue is best appreciated when red is chosen as the stimulus. With increasing dark adaptation the tertiary gains in bright- ness but loses in chromatic value. There is some difference of opinion as to whether the tertiary image can be seen at the fovea. Hamaker described the secondary image as coloured with the complementary colour and lasting | sec, the tertiary as homochromatic and lasting several seconds. Dark adaptation had little influence. The secondary image was the better seen for all colours ; the tertiary better for red and yellow than green and blue. With stationary light he found the following phases : (1) the primary image ; (2) a dark interval, often absent ; (3) the secondary image ; (4) a dark interval (| sec.) ; (5) the tertiary image ; (6) a dark interval, often absent ; (7) with stimulus of 1 to 4 sees, a quaternary image, accompanied by shortening of the tertiary. The quaternary is a true negative after-image, of complementary colour and surrounded by a bright halo. With moving light he found the following phases : (1) the primary image {e.g., 3°), passing into (2) a short homochromatic or white tail (6°) ; (3) a dark interval (10° — 15°) ; (4) the complementary, coloured or colourless secondary image (satellite, 3°, absent with red light) ; (5) a dark interval (50°) ; (6) the tertiary, faintly coloured homochromatic image (about 360°). As regards the region of retina stimulated, with stationary light, red and green gave a very marked complementary secondary foveal image, which was absent for yellow and blue ; the tertiary image was absent with foveal stimulation. With moving light the fovea gave no secondary and apparently also no tertiary image. The " action-time " of a light stimulus (McDougall), i.e., the least time during which a light of given intensity must act upon the retina in order to excite the most intense sensation it is capable of exciting, has been studied by Swan^, Exuer^, KunkeP, Charpentier'*, Martins^, and McDougall^. It is a necessary preliminary to the proper under- standing of flicker phenomena. It may be accepted that the sensation curve rises rapidly to a maximum and falls gradually. The rapidity of the rise varies with Ophth. XLiv. 3, 445, 1897 ; SamojlofE, Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xx. 118, 1899 ; Hamaker, loc. cit. xxi. 1, 1899; P. Miiller, Arch. f. d. yes. Psychol, xiv. 358, 1909; Kinoshita, Ztsch. /. SijDiesphysiol. xixn. 420, 434, 1909; Hering, Arch. f. d. grs. Physiol. cxxvi. 604, 1909; Dittler and Eisenmeier, ;6/(7. cxxvi. 610, 1909. 1 Trmis. R. S. Edin. Ii. 230, 1849 ; xxn. 33, 1861. 2 Sitz. d. Wiener Akad. Lvm. 2, 601, 1868. » Arch.f. d. ges. Physiol, ix. 197, 1874. « C. r. soc. de hiol iv. 1887. ^ Btitr. z. Psychol, u. Philos. i. 2, 1902. « Brit. Jl. of Psychol, i. 151, 1904. RECUKRENT VISION 91 the intensity of the stimulus, and the rate of diminution of the action- time with increase of intensity of the light diminishes as the intensity increases, so that with light of ordinary, fairly high, intensity it is very small. It has been pointed out above that when the duration of stimula- tion is very short the response is a series of pulses of sensation. If the stimulus be other than very brief the sensation shows no such pulses. McDougall has shown that with a light whose action-time is 61'^(yfi^7y sec.) the multiple character of the response persists until the duration of its action exceeds its action-time by about IG*^, and ceases altogether when the duration of the action of the light exceeds its action-time by about 40"^. As regards the variation of the intensity of the sensation with the duration of the stimulus when that duration is less than the action-time McDougall concludes that it follows the "photographic law," i.e., for such duration the intensity of the sensation varies directly with the duration of the action of the light. In mathematical terms the law may be stated thus : Sec I . St, where S is the intensity of the sensation, I the intensity of the stimulus, and 8^ a small increment of time. This law is a corollary of the Talbot-Plateau law {vide infra), and was assumed by Talbot. With very dim lights of scotopic value the action-time is probably slightly greater than 200"^ or i sec. The practical conclusion — that the dimmest light visible to the peripheral retina of the dark-adapted eye, i.e., the dimmest light perceptible under the most favourable conditions, must be allowed to act for a period of not less than 200"^ or i sec. in order to be perceptible — is of great importance in the construction of lighthouse flash-lights and so on. Its practical importance is increased, as well as the theoretical signifi- cance, by McDougall's observations which show that the action-times of red, green and blue lights are the same for lights of equal intensities, i.e., intensities which excite sensations of equal luminosities. Briickner and Kirsch^ have investigated the chromatic action-time or the minimum time during which a coloured light must stimulate the retina in order that the colour may be discriminated {Farbenzeit- schwelle or specifische Zeitschivelle). They found that the time varies with the intensity of the white light acting upon the retina both before and after stimulation with the coloured light and is nearly proportional to the brightness of this " grey " sensation. The chromatic action-time follows Weber's law, at any rate within certain limits. It varies inversely as the width of the pupil and the size of the retinal area stimulated, but not proportionally. The intensity of the following ^ Ztsch.f. SinnespTiysiol. xlvi. 229, 1911. 92 COLOUR VISION white stimulation appears to have a greater effect than that of the preceding white stimulation, though the authors can give no explanation of the fact. If a colour stimulation follows instead of white, the effect on the action-time depends largely upon the similarity or dissimilarity between this light and that which is being tested. THE TALBOT-PLATEAU LAW, When periodic excitations follow each other with sufficient rapidity the resulting sensation is one of continuous light of uniform brightness. Talbot^ and Plateau^ investigated the relationship of the brightness of the individual periodic stimuli and of the resultant sensation. Their conclusions are usually known as the Talbot-Plateau Law, which states that the resultant impression is the mean of the periodic impressions, i.e., the resultant brightness is that which would have arisen if the amount of light intermittently reaching the retina had been uniformly distributed over the whole period of stimulation, v. Helmholtz confirmed the law for ordinary physical intensities of light. Fick^ found some deviations, low intensities giving a continuous sensation brighter than the intermittent. 0. Clrlinbaum^ also found deviations for high in- tensities, the intermittent light being brighter than the continuous. The accuracy or otherwise of the Talbot-Plateau Law is of consider- able importance in the investigation of colour vision, for the principle of the episcotister depends upon its accuracy. Abney finds that the adjustable sectors, i.e., those which can be altered during rotation, are only available for accurate measurement when the angles of aperture lie between 180° and 10°, chiefly owing to the errors in reading being proportional to the angles of aperture. With fixed sectors angles of 2° or even 1° can be used. It is not certain, however, that the method is reliable for very low intensities of light, and for these the annulus {v. p. 5) is to be preferred. The experiments of Abney, Lummer and Brodhun^, Hyde^ and others show that the law applies accurately over a wide range of physical intensities. Hyde's experiments were very carefully conducted, the probable errors of measurement being under 0"1 per cent. The average deviation of the observations for any given angular aperture of the sectors was in no case as large as 0*2 per cent. 1 Load., Edin. and Dublin Phil. Mag. v. 327, 1834. - Ann. d. Phys. u. Chem. xxxv. 457, 1835. * Arch.f. Anat. Phyxiol. u. iviss. Med. 754, 1803. « J. of Physiol, xxii. 433, 1898. ^ Zhch.f. Imtrumentenkunde. xvi. 209 1896. * Bull, of the Bureau of Standards, Wasliington, ii. 1, 1906. RECURRENT VISION 93 It has been seen that McDoiigall arrived at the conclusion that the relationship Sec I . St is probably correct. Viewed from the purely physical standpoint this relationship and hence the Talbot-Plateau law are such as might be anticipated^. From the physiological standpoint one might rather have anticipated that the resultant sensation from intermittent stimuli would follow the analogy of muscle tetanus, i.e., that the application of new stimuli in the course of the curve of sensation produced by the primary stimulus would suffice to prolong the sensation curve at or near the maximum of the primary curve. It appears therefore that the superposition of fresh stimuli during the progress of the sensation curve of the primary stimulus produces positive or inhibitory effects of such a nature that the resultant sensa- tion curve shows rapid oscillations about a mean sensation intensity. If the oscillations are sufficiently rapid a continuous mean sensation results, in accordance with the Talbot-Plateau law. If the oscillations are less rapid the sensation of flicker is felt. The analogy to muscle tetanus is therefore not one to be pressed, but physiologists will have no difficulty in finding analogies in the domain of neurology to the in- hibitory effect of superposed stimuli. One may direct attention particularly to Sherrington's work^. THE FLICKER PHENOMENON. If the oscillations produced by intermittent stimulation are not sufficiently rapid to cause complete fusion, a sensation of flickering is felt. If black and white sectors are rotated with gradually increasing velocity there is first separate vision of the individual sectors. This is followed by a peculiarly unpleasant coarse flickering, which passes into a fine tremulous appearance, after which complete fusion occurs. Care- ful observation reveals further interesting, though complicating details. At a certain stage there is a peculiar glittering, the brightness being greater than that of the continuous sensation after complete fusion. Briicke^ found the maximum with 17 "6 stimuli per second. It may be explained by an absence of those inhibitory effects referred to above, or by reinforcement of each white stimulus by the recurrent image of the preceding sector (Briicke), or by temporal induction {vide infra), or by a combination of such causes, v. Helmholtz noticed that at one 1 Cf. V. Kries, in Nagel's Handb. d. Physiol, d. Mcnschcn, p. 231. - The Infegraiive Action of the Nervous System, London, 1906; ci. also McDougall, Brain, xxvi. 153, 1903. 3 Sitz. d. k. Akad. d. Wissensch., Wien, XLix. 2. 128, 1864. 94 COLOUR VISION stage the forward edge of the white sectors appears reddish, the back- ward edge bluish. Fechner^, by a suitable arrangement of black and white sectors obtained a pattern of yellow and blue. Subsequent observers have succeeded in eliciting all gradations of colour. This phenomenon is most easily seen in Benham's top^. It is probable that successive induction (Section V, Chap, ii) plays a prominent part in these phenomena, and they are clearly of great importance in the theory of colour vision. Exner^, Burch* and others have made important observations on this subject, using coloured lights. More important for our present purpose is the relationship between flicker and the conditions of stimulation — intensity and nature of stimulus, rapidity of stimulation, adaptation of the retina, and region and area of the retina stimulated. Various methods of producing intermittent stimulation other than by the usual rotating pigment- coloured discs, have been devised by Rood^, Whitman^, v. Kries', Simmance and Abady^, Kruss^, Wild^", Ives^^, and Watson^^, and the subject has been investigated not only by them but by others, notably by Ferryi^, Sherrington^^, Haycraft^^, 0. Griinbaum^^, Polimanti^'', T. C. Porter^s, Alleni^, Kennelly and Whiting^o, DowSi, Tufts^a, Millar^^, and Morris Airey^*. As regards the intensity of the stimulus, more rapid stimulation is required for complete fusion with increased intensity. Up to a certain point with alternate darkness and light as the intensity is 1 Ann. d. PhijK u. Chem. xlv. 227, 1838. - Bidwell, Proc. Roy. Soc. lx. 368, 189«; lxi. 268, 18D7 ; Percival, Trans. Ojihth. Soc. XXIX. 119, 1909. 3 Arch.f. d. gc.s. Physiol, i. 375, 1868. ^ J. of Physiol, xxi. 431, 1897. s Atner. J. oj Sc. (3) XLVi. 173, 1893; (4) viii. 194, 1899; Science, vii. 757; vm. 11, 1898. 6 Phys. Rev. m. 241, 1895. ' In Polimanti, Ztsch. f. Psychol, u. Physiol, d. iSinnesorg. xix. 203, 1899. * Proc. Phys. ifioc. xix. 39. 9 Phys. Zeitung, ni. 65 ; Jl. f. Gas u. Wass. XLvn. 129, 1904. 1" London Electrician, 1909 ; The Illuminating Engineer, i. 825, 1908. " Philos. Mag. 1912. i- In Abney, p. 107. " Amer. J. of 8c. (3) xliv. 193, 1892. " J. of Physiol, xxi. 33, 1897. " Ihid. XXI. 126, 1897. " Ibid. xxn. 433, 1898. i' Loc. cit. i« Proc. Roy. Soc. Lond. lxiii. 347, 1898 ; lxx. 313, 1902. " Phys. Rev. XI. 257, 1900 ; xv. 1902 ; xxviii. 45, 1908. 20 The Illuminating Engineer, New York, ti. 347, 1907. 21 Proc. Phys. Soc. xx. 644, 1907; xxn. 58, 1910; Philos. Mag. 120, 1906; 58, 1910; The Electrician, lviii, 609, 1907. 22 P%5.jBei;. XXV. 433, 1907. "^ T he I Hum inatingEngineer,NewYork,i\. 7 Hi), IQOd. 24 J. hist. Eke. Engs. xliv. 177, 1910. RECURRENT VISION 95 increased in geometrical proportion the rapidity of stimulation increases in arithmetical proportion. Griinbaum found that with high intensities the law fails, just as he found the Talbot-Plateau law to fail. If light alternates with less light the point of fusion is lowered by the diminution of the differences of intensity. Filehne^, Schenck^ and others found that with rotating discs the number of sectors influences the result. The more numerous the sectors the greater the number of rotations necessary per second for fusion. The difference is less marked when the disc is viewed through a slit (Schenck)^. Griinbaum worked out the relations and found the explanation in successive contrast (vide infra). Sherrington with coloured discs found the results markedly affected by simultaneous contrast {vide infra). We may therefore conclude that " the point of fusion of intermittent stimuli, so as to produce a continuous sensation, depends, not on the phvsical intensities of the stimuli, but on their physiological intensities, as determined by the condition and nature of the stimulated retina " (Rivers)^. This fact is further borne out by the areal and regional differences. Exner^ found that the duration of the sensory process decreased in arithmetical proportion as the size of the retinal image increased in geometrical proportion ; and Charpentier^ similarly found that increase in the size of the area of retina stimulated raised the point of fusion. Exner found that the fusion frequency was less for the fovea than for a region 1"33 mm. outside it, and all agree that flicker persists longer in the peripheral than in the central areas of the retina. Bellarminoff' found differences between the nasal and temporal areas. T. C. Porter's researches have increased the knowledge of flicker phenomenon materially. Using black and white sectors illuminated by pure spectral lights from the interference spectrum of lime-light, he found that the speed at which flicker was abolished for yellow was nearly double that for violet ; green and the last distinctly visible red occupying the mid-position. Having regard to the increase in speed with increasing intensity of the light, he concluded that as the retinal stimulus increases in intensity the sensation produced retains its maximum for a shorter and shorter time. With constant illumina- tion, altering the relative sizes of the white and black sectors, he found 1 Arch.f. Ophth. xxT. 2, 20, 1885. ^ Arch.f. d. ges. Physiol, cxn. 1, 190G. « Ibid. LXiv. 165, 1896. « In Schiifei's TeM Book of Physiology, ii. 1072, 1900. 5 Silz. d. k. Akad. d. Wisscnsch. Wicn, Lvm. 2, 601, 1868. 6 Arch. d'Opht. x. 340, 1890. ^ Arch. f. Ophth. xxxv. I, 25, 1889. 96 COLOUR VISION that at the point of fusion the effective stimulus at any point of the retina is to the maximum stimulus as the angle of the bright sector to 360°. The fact that the colour stimulus requires a finite time to produce its maximum effect was confirmed by this method. The period during which the sensation remains undiminished appears to decrease as the time of stimulation increases, though within narrow limits of variation one of these quantities is nearly inversely proportional to the other. He found that the relation between the intensity of illumination (/) and the number of revolutions {n) per minute at which a disc half white and half black must be run in order that the flicker may just disappear, the distance of the eye remaining constant, is n = k . log I + k' k and k' being constants. He therefore proved the geometrical- arithmetical relationship between intensity and rapidity of stimulation. The relation was found rigidly true for illuminations from 0"25 metre candle to 12,800 times this value. For intensities below 0'25 metre candle the constant k suddenly changed its value to practically half its former value. We have here further evidence of the duplex mechanism involved, the higher value of k applying to the photopic, the lower to the scotopic mechanism. T. C. Porter also proved that the duration of the undiminished sensation produced by different spectral hues depends solely on the luminosity of the colours and not on their wave frequency. .Ives's researches have elicited several highly important results and are specially valuable owing to the great care taken to secure accuracy in his methods. The flicker method can be applied in two ways. In one, that used by Haycraft, Ferry and Allen, the critical frequency of alternation of the lights is measured, i.e., two lights are regarded as being of the same luminosity when the flicker produced by rapid alterna- tion of each with black disappears at the same speed of alternation. The other method is that employed by T. C. Porter and generally used in flicker photometry. By it two lights are regarded as being of the same brightness when no flicker results on alternating one with the other, the speed of alternation being such that the slightest change of intensity of either light causes flicker. Ives found that the second or ordinary method is more sensitive than the equality of brightness method used by Abney and others {v. p. 44). The results are reproduced with much greater constancy, since in the latter method psychological factors influence the judgment in arriving at the results. On decreasing the illumination the maximum shifts towards the blue (Purkinje's RECURRENT VISION 97 phenomenon) by the equality of brightness method ; by the flicker method it shifts towards the red. On the other hand, decrease of the area stimulated at low intensities shifts the maximum of luminosity towards the red by the equality of brightness method, towards the blue by the flicker method. Ives found the relative positions of the two kinds of spectral luminosity curves generally different. They differ most in position at low illuminations with large areas ; least at high illumina- tions with small areas. The mean curves of several observers show close agreement in the position of the maxima and the shape of the two curves at high intensities, but the areas of the curves are not equal. At low illuminations all observers agree in showing the Purkinje and the reversed Purkinje effects. Haycraft, by the critical frequency method, obtained a pronounced Purkinje shift at low intensities. Ives obtained the reverse effect, except at very low intensities (0"5 metre-candle), when he confirmed Haycraft's result. Ives sought an explanation in Porter's change in the logarithmic rates at which critical frequency varies with the illumination. In Porter's equation n = k . log I + k' k has a different value above and below 0*25 metre-candle. If the critical frequencies are plotted against the logarithms of the illumination for white light two straight lines of different slope, which meet at about 0"25 metre-candle, are obtained. The reversed Purkinje effect occurs above, the true Purkinje effect below this point. When separate colours are investigated and plotted in the same manner, a set of straight lines of differing slope results. The most remarkable curves are those for red (650 ^/x) and blue (480 /xju). The former maintains its direction unchanged ; the latter suddenly changes from a diagonal to a horizontal, i.e., the critical frequency becomes a constant, independent of the (low) illumination. The curves for other colours take an intermediate course. Hence Porter's law for white light holds good for different colours if the values of the constants are changed. i^==/iC,.log^, + /iC/, where F is the critical frequency, S^ is the slit-width, K^ is a constant involving the relationship between critical frequency and intensity of radiation for the individual eye for the colour in question and for the size of the area stimulated, and K^' is a constant involving the quantity of energy emitted by the source, the dispersion, etc. of the instrument, p. c. V. 7 98 COLOUR VISION and the sensibility of the observer's eye to flicker for different colours at a given speed. The Purkinje effect, and its reversal above 0*25 metre-candle, follow at once from these facts. In the flicker phenomenon with colours two causes of flicker are at work, the colour element and the luminosity element^. Of these colour- flicker ceases first. Hence the total flicker effect may be the resultant of the two flicker sensations. Ives has shown that the flicker photometer is largely influenced by the critical frequency phenomenon, but that it does not obey the simple law which would follow were it a mere dove- tailing of two pure flickers. Allen and others^ found that the peripheral retina is more sensitive than the fovea to flicker, as might be expected from its high sensitiveness to movements. Ives, however, found that this result is only true for momentary observation. Adaptation or fatigue sets in very rapidly and then the periphery becomes less sensitive. The fovea is more sensitive to red flicker, the periphery to blue, and this difference is more striking at low intensities (Dow, Ives). If the comparison light is coloured or the stimulating area is sur- rounded by white instead of black the equality of brightness method produces irregular and unsystematic shifts and distortions of the spectral luminosity curves, possibly owing to the increase in distracting psycho- logical factors. Such changes produce no alteration in the luminosity curves by the flicker method. A curve almost identical with the flicker curve can be obtained by the equality of brightness method if it is built up of small steps of slight hue- difference and with small areas. By this so-called " cascade " method the differences in hue are made so small that they do not disturb the judgment of brightness. When the areas of the luminosity curves by different methods are compared it is found that the visual acuity method {v. p. 44) gives a curve many (about five) times as great as the equality of brightness, the flicker and the critical frequency curves, which agree much more nearly. The enormous area of the visual acuity curve is due to the chromatic aberration of the eye. It has been shown by BelP and Luckiesh^ that the resolving power of the eye is much greater for monochromatic than for complex light of the same hue. For a method of measurement to be accurate it should conform to two axioms : things equal to the same thing are equal to one another ; and the whole is ^ Cf. Liebermann, Ztsch.f. Sinnesphysiol. xlv. 117, 1911. 2 Cf. Lolimann, Arch. f. Ophfh. Lxvm. 395, 1908. 3 Eke. World, LVii. 11G3, 1911. « Ibid. Lvm. 450, 1911. RECURRENT VISION 99 equal to the sum of its parts. Most physical measurements conform to these axioms, but it cannot be assumed that the same agreement holds good for measurements involving qualitative differences, such as those of colour. The reliability of a method of heterochromatic photometry must be judged according to the following criteria. First, the shape of the luminosity curve must not be altered by change in the reference standard. Second, the sum of the measurements of the brightness of the parts of the spectrum must be equal to the brightness of the recombined spectrum. 10 8 €^ "^=^~Ci / *N 1 d \ \ / 1 // // t 9V \ \ \ \ o K FULL LINE- EQUALITY OF BRIGHTNESS BY SMALL STEPS DASHED LINE - " , " " "wHITE COMPARISON EIELD CIRCLES - FLICKER METHOD 1 1 1 1 1 1 1 52 54 56 58 60 62 64 Fig. 38. Photopic luminosity curves taken by the flicker and equality of brightness methods. Abscissae, wave-lengths from 520 ^/a to 640 a'M ; ordinates, arbitrary scale. (Ives.) So far as the position of maxima and the general form of the curves are concerned all the methods agree fairly well. Comparison of the different methods shows that the areas of their curves differ most for the visual acuity method as compared with the other three. These latter do not agree well. Of them, the equality of brightness method gives greatest variations, the critical frequency method smaller, and the flicker method least. The crucial test has been applied to the flicker method by Whitman^, Tufts^ and Ives. Of these the last named is by far the most accurate. The method consists essentially in measuring the luminosities of the parts of the spectrum against a standard which ^ Phys. Rev. m. 241, 1896. 2 Ibid. XXV. 433, 1907. 7—2 100 COLOUR VISION is identical with the source of light of the spectrum. Ives obtained a remarkable confirmation of the accuracy of the flicker method, and arrived at the conclusion that it surpasses all other photometric methods in sensitiveness and accuracy. Finally, Ives measured the luminosity curves of eighteen observers with normal colour vision at 25 metre-candles illumination of a mag- nesium oxide surface. The average curve deduced from these experi- ments is shown in Fig. 39 and is compared with Konig's curve as reduced by Nutting^ and with Thiirmel's curve^. ■40 -42 Fig. 39. •44 46 id -50 -52 34 -56 •58 -60 -« •64 66 •68 •70 -ri/i Average photopic luminosity curve of 18 observers. Konig's photopic luminosity curve (equality of brightness method). Thiirmel's photopic luminosity curve (flicker method). (Ives.) In conclusion it may be stated that in the case of colours it is generally agreed that the fusion frequency depends solely on the luminosities of the alternating fields, and the coincidence of the equality of brightness curve with the flicker luminosity curve (Figs. 38, 39) bears out the assumption. We may at any rate assume with confidence that both the methods measure the " something " (p. 44) which we call brightness or luminosity. ^ Bull, of the Bureau of Standards, vii. 2, 235. 2 Ann. d. Physik. xxxm. 1154, 11)10. SUCCESSIVE INDUCTION OR AFTER-IMAGES 101 CHAPTER II SUCCESSIVE INDUCTION OR AFTER-IMAGES After the stimulation of an area of the retina with light, the withdrawal of the stimulus does not result in the obliteration of all visual sensation. The succeeding sensations vary according to the nature and intensity of the primary stimulus, the condition of adapta- tion of the retina, the region of the retina stimulated, etc. If a second stimulus is applied to the same area of the retina during the course of the after-effects of the primary stimulus the resulting sensation is modified by these after-effects. The effect produced by the secondary stimulus can be measured by comparing it with that obtained from a retinal area not previously exposed to the primary stimulus. The stimulus applied to such a resting area is called the comparison lighfi. The sensations obtained from the persistence of the physiological processes set up by a primary stimulus, whether altered by the effects of succeeding stimuli, or unadulterated, i.e., observed under the condition of complete exclusion of fresh stimuli, are commonly called " after- images." After looking for from 20 to 40 seconds at a white spot of light and then directing the gaze upon a white surface a black spot surrounded by a bright halo is seen. This is the " negative after-image." If the spot or light is coloured the after-image is seen to be tinged with the complementary colour. This is the " complementary after-image." If care be exercised a different phenomenon will be observed. Direct the eyes towards a bright spot or light, but keep them carefully covered with the hand for a minute or so. Remove the hand rapidly but with- out any sudden or violent action and quickly replace it, so that the object is momentarily seen. If the experiment is properly carried out the spot will be seen as a persistent after-image in its original brightness and detail. This is the " positive after-image." If the object is coloured it is seen tinged with the same colour ; this is the " homo- chromatic after-image." Although in ordinary circumstances negative after-images are more easily obtained than positive McDougall^ has shown that under suitable ^ In Germany the effect of the pi'imary stimulus is called "tuning" the retina (Hering). The jirimary stimulus is called the "retuning light" (das umsl immende Licht); the secondary stimulus the '-'reacting light" (das reagirende Licht). 2 Mind, X. N.S. 74 sqq., 1901. 102 COLOUR VISION conditions the reverse is the case. If the primary stimulus is an area of white light which is sharply defined and surrounded by a dark back- ground the negative after-image, bordered by a halo as described, is observed. If however the transition from the illuminated area to the background is made gradual a positive after-image is almost invariably obtained. McDougall used a ground glass disc, 12 cm. in diameter, illuminated from behind. On the far side about a dozen sheets of white paper were pasted, each with a circular hole in it con- centric with the edge of the disc. Of these holes the smallest was 2 cm. in diameter, and each of the others was about 1 cm. larger in diameter than the preceding one. Such a " shaded disc," when illuminated, showed a central evenly lighted circle, 2 cm. in diameter, surrounded by a zone, 5 cm. in breadth, in which the brightness diminished regularly to the periphery, where it became negligible. The shaded disc gives with white light a positive after-image without any halo. On the other hand with colours, though the after-image is often homochromatic in the first one or two seconds, with most intensities of light it is approximately complementary through almost the whole of its course. In the case of red and green especially it is very constantly the rule that red predominates in the after-image of green, and vice versa. Sufficient attention does not appear to have been paid hitherto to this fundamental difference between the relation of black to white and that of green to red and blue to yellow in such after- images from " shaded " lights. Though positive and negative after-images are opposed there is no discontinuity between them. If the positive after-image is developed and the eyes are then uncovered and directed towards a uniform moderately illuminated field the negative after-image at once appears. If the illumination of the field is suitably chosen no after- image is seen. The nature of the after-image depends therefore upon the nature and intensity of the primary stimulus and upon the nature and intensity of the secondary stimulus. Great diversity of after-images may therefore be obtained. The after-image of a white spot may be coloured^ ; that of a coloured spot may not be accurately the comple- mentary of the primary stimulus. Anomalous colouration may be due to light passing through the sclerotic and the iris, or to abnormal conditions of the retina (Hilbert^). Moreover, Burch^ has brought ^ Aristotle; Goethe, Farbenhhre (1819), Eastlake's trans, p. IG, 1840. 2 Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. iv. 74, 1893. 3 Proc. Roy. Soc. Land. B. lxxvi. 212, 1905. SUCCESSIVE INDUCTION OR AFTER-IMAGES 103 forward evidence to show that the after-effects of a primary stimulus are much more prolonged than has been generally recognised, and these later after-effects may materially alter the character of the after-image. So complex indeed are the effects that the greatest care should be exercised in making theoretical deductions from after-images. With primary stimuli of considerable intensity a very short exposure is necessary to induce the after-image, which may be so strongly developed as to nullify successive stimuli to the same part of the retina for a considerable time ("positive scotoma"). W. Tschermak stated that primary stimuli with coloured lights of such low intensity as to be below the chromatic threshold were followed by coloured after-images, but it was not found to be so by Titchener and Pyle^. Analogy with other physiological processes would lead us to the conclusion that the positive after-image results from a persistence of those processes which have been set in action by the primary stimulus, and that the negative after-image is the expression of a diminished excitability of the stimulated area to fresh stimuli (Fechner, v. Helmholtz). The fact that a stimulus usually gives rise to a rhythmical response, as has already been shown (p. 87), makes the periodical variations in the after-image, which were first described by Plateau, intelligible. Under suitable conditions, with the eyes motionless, the negative image disappears and reappears at intervals of three or four seconds, sometimes alternating with positive images. Sudden movements of the eyes, distraction of the attention, and other like influences abolish the after- image ; hence they are seldom noticed in every-day life. Moreover with prolonged or strong stimuli the condition of altered sensibility of the affected part of the retina may persist materially longer than the apparent after-effect as shown by the negative after- image. When the latter seems to have quite disappeared if the eye is thoroughly darkened the affected portion of retina gives rise to a sensa- tion of blackness which is more intense than that derived from the surrounding areas. It might seem that the effect of a single isolated stimulus, without any secondary stimulus, was a simple matter. This is, however, by no means the case. The negative after-image of a white object appears as a black spot in the midst of a less black field. There is there- fore a sensation of blackness which is blacker than that obtained from the eye when all external light is excluded. The eutopic sensation when the eyes have long been completely excluded from light is variously 1 Proc. Amer. Philo.-i. Soc. XL. 300. 1908. 104 COLOUR VISION described by different people, but all agree that it is not an impenetrable darkness. There are waves or points of light, and the general sensation is one of very dark greyness. So far as negative after-images are concerned this " light chaos " or " light dust " acts as a grey surface on which the black image is projected. Many explanations of this " intrinsic light " have been suggested. That endogenous stimuli are the cause can scarcely be doubted, but whether these are primarily retinal or of central nervous origin remains Uncertain^, though it is now generally agreed that they are central. The question is of theoretical importance and will receive further treatment in Part III. McDougalP found certain features which, he claims, are quite constant in the after-images of coloured lights. They are as follow : (1) The after-images show in nearly every case a play or succession of colours, in which each of the three simple colours, red, green, and blue, makes itself felt in some phase. (2) The brighter the coloured light fixated, the brighter are the colours of the after-image, and the more keen is the antagonism between the three simple colours, so that these colours fuse less than in the after-images of duller colours. When the primary stimuli are very bright the three simple colours tend to appear pure and saturated in turn in a recurring scale, the unchanging phases of pure colour being separated by periods of struggle between the fading and the succeeding colour, just as is the case in the after-image of bright white light. (3) In the case of fixation of one of the three simple colours, red, green, and blue, the lower intensities are followed by after- images in which the other two colours predominate, more or less fused, i.e., the after-image is predominantly complementary ; while when the colour fixated is very bright the first phase of the after-image is usually homochromatic and of considerable duration. The like holds good for the compound colours, but in their case the phases of the after-image tend to be rather more varied. (4) The order of occurrence and the duration of the different phases vary readily with slight variations in the conditions. (5) With any given light, the vividness and duration of the after-image, observed in the dark, increase with increase of duration of fixation of the light from about 10 sees, to 90 — 100 sees., but with further increase of the period of fixation, the duration and vividness of the after-image diminish, so that after very prolonged fixation the after-image is either dull or of short duration or is not seen at all. As with white light, a sharply bounded patch of coloured light, 1 See V. Helmholtz, 3rd ed. p. 12 sqq. « Mind, x, N. S. 1901. SUCCESSIVE INDUCTION OR AFTER-IMAGES 105 fixated for more than a few seconds, gives rise to an after-image sur- rounded by a halo. If the coloured light is not very bright, and the fixation not very prolonged, the halo is of the same colour. The longer fixation is continued, the brighter and the less saturated is the colour of the halo, until after prolonged fixation it becomes white or even tinged with the complementary colour, and is so much brighter than the after- image itself as to inhibit it partially or wholly. The colours of the after-image and its halo tend to be complementary to one another, for during the observation of an after-image in the dark the conditions are very favourable to contrast effects (See Section VI). The effects of a secondary stimulus on a previously stimulated area are of great importance in considering the facts of colour vision. We have seen that the primary stimulus alters the excitability of the retina so that it is lowered for a succeeding stimulus of the same nature. So far at any rate as colours are concerned we may carry the matter further and say that it is raised for a secondary stimulus of the com- plementary colour. If a complementary after-image of any spectral colour be obtained and the secondary stimulus be this complementary itself, a sensation of this complementary colour is produced which far exceeds in purity and saturation any such colour which is found in nature or in the spectrum. If for instance a complementary after- image of green obtained from a suitable purple field be compared with a patch of the corresponding spectral green, it will appear of extra- ordinary brilliancy and saturation. The question therefore arises whether these after-effects seriously complicate the equations of colour matches. It may be said at once that so far as the peripheral retina is concerned they do, because we have already seen that the periphery values are not the same as the foveal values and they are far more susceptible to variations in adapta- tion. With regard to the fovea, however, it is found that all colour matches still remain valid, no matter what kind of light may have previously stimulated the retina. Thus a match of monochromatic yellow with a mixture of red and green remains a match after previous illumination with yellow and blue. If yellow has been used, both become paler ; if blue, both become more saturated yellow; but they still match. This law has been systematically in- vestigated by Biihler^ under v. Kries' direction^. It may be stated 1 Diss. Freiburg, 1903. ^ See also v. Kries, Arch. f. Anat. 503, 1878 ; and Dittler and Orbeli, Arch. f. d. ges. Physiol, cxxxii. 338. 1910. 106 COLOUR VISION at once that it also holds good for the ordinary types of colour blind (dichromats)^. The law has, however, been denied by Hering^ and Tschermak^, and Watson* has shown that it is at best an incomplete statement of the facts. He confirmed the fact that colour matches remain valid after previous stimulation with another light. He obtained the fol- lowing results : Width of " green " slit Resting Previously stimu- Character of match retina lated retina Correct 22-7 23-5 Toolittle green 20-5 110 Too much green 25-0 27'2 In each case white was matched by a mixture of spectral red, green and violet lights by altering the width of the slits through which the coloured light proceeded. When the match was correct for the resting eye it was found to be also correct for the eye which had been previously stimulated with a colour, e.g., red. If now the slits through which the red and violet lights passed were kept constant, while the green slit was altered, being reduced until the deficiency of green was just observ- able and opened until the excess of green was just perceptible, it was found that the range through which the green slit could be altered while still preserving a correct match with the white light was much greater for the eye previously exposed to red light than for the eye not thus previously stimulated. As shown above, the range between a perceptible excess and a perceptible deficiency of green was 4*5 for the resting eye and 16'2 for the previously stimulated eye. These experiments therefore show that the statement that " all colour matches still remain valid," though true, is not the whole truth, and is indeed misleading. After previous stimulation the range of intensities which give a valid match is much wider than for the resting eye. If we accept the law of the validity of optical equations irrespective of previous stimulation, we can obtain a relation between the changes in appearance of various lights caused by previous stimulation. If R is the measure of the stimulus, the sensation will be aR, where a. is the measure of the retinal excitability for the particular stimulus R. ^ V. Kries and Nagel. Zf.^ch. f. Psychol, u. Physiol, d. Sinncsorg. xii. 1, 1896 ; xxni. 161, 1900. 2 Arch.f. d. ges. Physiol i.iv. 309, 1893. » Ibid. lxx. 297, 1898. * Abney and Watson, Proc. Roy. Sac. Land. A. lxxxix. 1913. SUCCESSIVE INDUCTION OR AFTER-IMAGES 107 Further, if Li at one retinal point gives rise to the same sensation as Lo at another, previously stimulated by a different light, and M\ at the first point gives rise to the same sensation as Mi at the second, then L] + M\ will have the same action at the first as Li + Mo at the second. As a special example, 'projjortiofial alteration of intensity of two stimuli acting on retinal areas which have been previously exposed to dift'erent excitations will produce the same alteration in sensation. This is the " Law of Coefficients " (v. Kries^), adumbrated by Wirth^ as the Fechner-Helmholtz law. From the law it follows that if the change from normal produced by stimulation with three different lights or light mixtures is known the change produced in any other light mixture can be deduced. The law applies only to the photopic condition. It is therefore least subject to deviation for foveal values. For peripheral values, which are so subject to scotopic variations, and for stimuli of low in- tensity, which so readily induce the scotopic condition, it ceases to be valid. Moreover, with feeble stimuli the endogenous stimulus of the intrinsic light becomes measurable relatively to the exogenous stimuli. When the primary and secondary lights are identical the result is a gradual diminution of luminosity (local adaptation, Hering). Such a diminution in differences of brightness occurs when two similar coloured fields of unequal luminosity are fixed for a considerable time. Besides the alteration in brightness there is also a diminution in saturation with coloured lights, and further, though it is scarcely noticeable with- out a comparison light, there is also a change in hue, Voeste^ found that a yellow of wave-length 560 //./i, a green of 500 /x/x, and a blue of 460yLiyLt,show no appreciable change of hue. Hues between 500 and 560yu./Lt change towards 560 yti/x, those between 500 /a^u, and 460 ^jx towards 460 /x/i. on prolonged fixation. These facts are of importance in the equality of brightness method of estimating luminosity matches. In Abney's* method " the angles of the sectors are rapidly altered from " too light " to " too dark " and back again, and the range of angle is gradually diminished until the observer sees both to be equally bright." The change in brightness on prolonged fixation constitutes one of the difficulties of this method with unintelligent examinees, and is an argument in favour of the flicker method. If the primary stimulvs is white and the secondary stimulus coloured 1 Nagel, p. 211. 2 Wundt's PMlos. Stud. xvi. 4 ; xvn. 3 ; xvni. 4. ^ Ztsch.f. Psychol, u. Physiol, d. Sinne/forg. xviii. 257, 1898. * Abney, p. 88. 108 COLOUR VISION the exposure to white changes the chromatic stimulus values. Thus, using a comparison light, and experimenting with rotating discs v. Kries^ found that if the coloured sectors were made the same both for the stimulated and the comparison area there was never even an approximate match ; and if, keeping the same coloured sectors, the white sectors were chosen such that the luminosities were equal, the colours seen by the stimulated area were of far too low saturation. In order to produce a match it was necessary for the amounts of coloured light in the secondary and comparison to be about 3 : 1 (blue sectors 270° : 97° ; red, 270° : 84° ; yellow, 270° : 97°). These experiments have been adversely criticised by Hering^ and diverse results have been obtained by Dittler and Richter^. Dittler and Richter arranged that two contiguous areas of the retina were stimulated with homogeneous blue lights so that an exact match was obtained. The stimulus was then cut off from one area and the other was stimulated with a binary white light obtained by mixing the blue with its complementary colour. The previous blue matches were then replaced and it w^as now found that the blue of the stimu- lated area was much more saturated than that of the resting area. Increase of intensity of the blue of the stimulated area failed to re- establish a match, but addition of white light effected a much better match. // the 'primary stimulus is coloiired and the secondary stimulus white the chromatic excitation causes the white to be coloured with the complementary colour. This is the ordinary complementary after-image, a red, yellow, green, or blue object giving a blue-green, blue, purple, or yellow after-image respectively when a white or grey field is fixed. If the sensation is matched with a comparison light on a neighbouring retinal area it can be shown that it is nearly if not quite as saturated as the spectral colour, even if the exciting light is of moderate intensity and has acted for only 30 to 40 seconds. Exact coincidence with the complementary colour of the exciting light is not generally found**. The colour of the resultant sensation is not a mixture of the colour which would be observed with the darkened eye without any secondary ^ Berichle d. Freiburger Naturf. Gesellschaft, 1894. - Arch. f. (1. ges. Physiol, xciv. 533, 1903. ^ Ztsch. f. Sinnesphysiol. xlv. 1, 1910; Dittler and Orbeli, Arch. f. d. ges. Physiol. cxxxn. 338, 1910. « Cf. Tschermak, Ergeb. d. Physiol, ii. 2, 7G3, 1903. SUCCESSIVE INDUCTION OK AFTER-IMAGES 109 stimulus plus the white light of the white or grey surface. If it were, after stimulation with blue, for example, the yellow of the after-image would always be compensated by mixing the same amount of blue with the secondary light, so that a white after-image would result. By the coefficient law, with increasing intensity of the white light the amount of blue added should increase proportionately, and this is found to be the case (v. Kries). The complementary after-image is therefore not the result of a mere additive process, but is a genuine qualitative change in the white light, though it may also involve a change of intensity. If the primary stimulus is coloured and the secondary stimulus also coloured various coloured after-images are obtained. It has already been mentioned that if the secondary stimulus is the complementary colour of the primary the resulting sensation is that of the extremely saturated complementary colour. This is specially true of the red- yellow end of the spectrum and for secondary lights of high intensity. It may indeed be far more saturated than any spectral colour, so that it is impossible to obtain a comparison light which will match it. This is a fact of profound theoretical significance, since it shows that we are capable of experiencing visual sensations which cannot be elicited by the application of any known physical stimulus to the resting eye. If the secondary stimulus differs from both the primary and its complementary the most diverse results may be obtained, but as a rule the complementary colour of the primary stimulus predominates. After green yellow appears reddish orange, blue purple, and so on. After red, yellow (589 /xyu) matched greenish-yellow (556 fi^) ; after green, 517 fi/u, matched 565 /^/z (Hess^). Attempts have been made by v. Kries^ for white, and by Schon^ for coloured lights, to obtain the time relations of after-images. The sensation derived from the secondary light is matched at rapid intervals with a comparison light. The results are of little value owing to rapidly changing adaptation and the difficulty of avoiding parafoveal stimula- tion. V. Kries* also studied the time relations of recovery from the primary stimulus. It is unknown whether the effect passes off as a smooth curve (Fechner, v. Helmholtz, v. Kries) or in rhythmical waves 1 Arch. f. OpMh. xxxix. 2, 45, 1893. Cf. A. W. Porter and Edridge-Green, Proc. Roy. Soc. Lond. B. lxxxv. 434, 1912; Edridge-Green, loc. cii. lxxxvi. 110, 1913. 3 Arch.f. Ophth. xxni. 2, 1, 1877. 3 Ibid. xx. 2, 273, 1874. * loc. cit. 110 COLOUR VISION (Plateau, Hess^). From analogy the latter is the more probable, but the question requires further elucidation^. The fading of after-images. The observations recorded above deal with successive induction resulting from short exposures. There are a number of careful observations by some of the older authors, notably Plateau^, Fechner*, Seguin^ and v. Helmholtz^, which are of importance in the theory of colour vision. They deal with the subjective impressions which occur during the fading of the after-images of a light stimulus. If the primary stimulus be of sufficient intensity the passage of the positive into the negative or complementary after-image is accompanied by a series of colour sensations. Of particular interest for our present purpose is the subjective development of colour sensations from stimulation with white light. The series varies according to the intensity and duration of the primary stimulus. Momentary excitation of white light produces an after-image which " passes rapidly through greenish blue (green, Seguin) into beautiful indigo blue, later into violet or rose-red. These colours are bright and clear. Then follows a dirty or grey orange, during which the positive after-image generally changes to a negative, and in the negative image this orange often becomes a dirty yellow-green'^." When the negative after-image has developed it appears dark against the background of the " intrinsic light " [v. p. 104). If white light of low intensity is now admitted the after- image passes on into the later coloured phases, and retreats again if the white light is again diminished. Thus, if the image when the eye is covered is blue the introduction of dim light causes it to pass through rose-red to a negative yellow. If the eye is again quickly covered the blue reappears. If the image in darkness is rose-red, a weak white light turns it yellowish-red, and so on. If the positive image has com- pletely disappeared from the darkened eye a dim white light develops a grey or green-grey negative after-image surrounded by a rose-red halo. If the primary stimulus is more intense or acts for a longer time colouration may commence before obscuring the eye (Fechner). It is first yellow, then blue-grey or blue, without passing through green, 1 Arch. /. (I. (jes. Physiol, ci. 226, 1904. ' Cf. Fick and Gurber, Arch. f. Ophth. xxxvi. 2, 245, 1890; Fick, loc. cit. xxxvra. L 118, 1892 ; Hering, loc. cit. xxxvra. 2, 252, 1892; Fick, loc. cit. xxxviii. 3, 300, 1892. 3 Essai dfune Theorie gen. etc. Bruxelles, ] 834. * Ann. d. Physik, l. 220, 1840. * Ann. de Chimie, 3rd series, xli. 415, 1850. • 3rd ed. n. p. 208. '^ v. Helmlioltz, loc. cit. ii. p. 208. SUCCESSIVE INDUCTION OR AFTER-IMAGES 111 then red- violet or red. The yellow phase is shortest. After longer and stronger stimulation by white light the after-image in the darkened eye passes through white, blue, green, red, blue, and on exposure to dim white light blue-green and yellow. Different authors give different phases with stronger excitation, but blue always starts the series. The after-images from direct exposure to sunlight are still more complicated^. The fading phases of after-images have been more recently in- vestigated by Miss Washburn^ and McDougalF. The latter has shown that it is unnecessary to use stimuli of excessive strength. He found that after excitation with white light a very constant feature was the tendency of the three simple colours to follow one another in a recurring cycle of the order green, red, blue, green, red, blue. If the white light be only moderately bright, red is usually the first colour to appear and is then succeeded by green, which persists to the end. After brighter light green comes first, and is succeeded by red, then blue, and then green again. If the light is still a little brighter, McDougall describes the green as being " usually mixed with red from the first, i.e., it appears yellow with a red edge." If the light be very bright then the first phase of the after-image is white or bright blue of very low saturation ; the blue soon passes through blue-green to green, which is then followed by the cycle of pure colours. According to McDougall " an important feature of the after-images of bright white light is that, after a first short period in which two colours fuse to give yellow, or, as is the case after the brightest lights, all three fuse to give white, the colours that in turn occupy the area of the after- image, alone and unchanging for considerable periods, are red, green, and blue only, and these are in every case of exactly the same colour- tone although varying in brightness in different cases and in different stages of one after-image. The red is a rich crimson red, decidedly less orange than the red of the solar spectrum, the blue is a rich ultra- marine, and the green a pure green having no inclination towards blue or yellow^." McDougall states that "they are the purest, richest, most saturated colours that I have ever experienced, and I believe that in this way, and this way only, one may experience absolutely simple, i.e., unmixed and fully saturated colour-sensations." The fading phases of the after-images froyii coloured lights are also not limited to the homochromatic positive and the complementary 1 V. Helmholtz, loc. cit. n. 211. 2 Pgychol. Rev. vii. 1900. 3 31{nd. X. N. S. 235, 1901. * Cf. Part III. 112 COLOUR VISION after-images. As the former passes into the latter in the case of a strong primary stimulus the image does not simply become paler, but un- saturated hues appear. The homochromatic hue of the positive image first disappears and the image resembles that of the after-image of a white object in which the rose-red phase predominates. " Then the complementary colour of the negative after-image gradually developes, but it may be already visible before the positive image has become negative ; it may therefore appear brighter than the dark background. I think that the appearance of the complementary colour may be attributed to the fact that the positive image, which at this phase has become feeble and white, is superposed upon the negative and comple- mentary images induced in the intrinsic light chaos by the fatigue of the eye. It is clear that by such a superposition after exposure to red the positive white and the negative blue-green can together give a greenish white positive image. These positive complementary images are mentioned by several observers (Purkinje^, Fechner-, Briicke^)*." It is an instructive experiment to view the spectrum momentarily and then to observe the phases of the after-images. Burch^ has adduced strong evidence to show that the duration of after-image effects is much longer than has been commonly supposed. He has observed changes occurring in the subjective sensations for as long as two hours. As with all subjective observations of this kind generalisations from them are to be accepted with caution. It is unnecessary to pursue this complicated problem further. The older experiments, especially Fechner's with coloured stimuli, were mostly made with impure lights. The matter requires further investiga- tion with homogeneous lights and due regard for adaptation, etc. CHAPTER III THE EFFECTS OF " FATIGUE The term " fatigue " has been avoided in the previous consideration of the effects of retinal stimulation. With our increased knowledge of adaptation it has become difficult to estimate the importance of fatigue and further discussion must be postponed until we deal with the theory 1 Zur Physiol, d. Sinne, n. 110. 1819. « Ann. d. Physilc, l. 21.3, 1840. 3 Wiener Denkschr. m. 12, 1850. * v. Helmholtz, loc. cit. ii. 213. 5 Proc. Roy. Soc. Lond. B, lxxvi. 212, 1905. THE EFFECTS OF "FATIGUE" 113 of colour vision. Earlier authors, including v. Helmholtz, and most physicists make free use of the conception of fatigue to account for successive induction, but the facts, and especially those of simultaneous contrast {vide infra), negative so simple an explanation. There is, however, a group of phenomena associated with prolonged or intense stimulation to which the term may be fittingly applied, though even in this case it should be done " without prejudice." These phenomena have been studied particularly by Burch^. By flooding the eye with bright sunlight which had passed through a lens of 2 inch focus and a suitable colour screen or by a similar process with a spectrum of wide dispersion Burch was able to study the after-effects of fatigue for the various colours. After fatigue for red, scarlet geraniums appear black, calceolarias and sunflowers various shades of green, and marigolds green shaded with black in the parts that are orange to the normal eye. Purple flowers, such as candytuft and clematis look violet, and pink roses bright sky-blue. Short exposures — a few seconds to two or three minutes^suffice, and the effect is transient, passing off in about ten minutes. After fatigue for violet, violet wools look black, purple flowers crimson, some blues greenish, green a richer hue. A noticeable effect is the tinging of all objects which do not reflect violet with that colour, and the same applies to " dazzling " with other colours. Fatiguing with green makes the landscape look like a picture painted with vermilion, ultramarine and flake-white, variously blended. The foliage is reddish-grey or bluish-grey, blue flowers are dirty blue, red flowers are impure red, and every colour but green is tinged with green. Fatiguing with jmrple {i.e. red and blue) makes everything look monochromatic green. All red, purple or blue flowers look black, and green looks a quite unnaturally brilliant green. The red fatigue passes off first, the observer then being in a condition of violet fatigue, which passes off in 15 — 60 minutes. Fatigue of one eye with purple and of the other with green produced a very weird and exaggerated stereoscopic effect. In fatiguing the eye with the colours of the spectrum comparatively simple changes, differing in degree but not in kind, occur with four regions, viz. the regions which give the sensations of pure red, ^ Phil. Trans. Roy. Soc. Land. B, cxci. 1, 1898; Proc. Roy. Soc. Lond. lxvi. 216, 1900. P. c. V. 8 114 COLOUR VISION unmixed green, unmixed blue, and pure violet. The changes are as follow : (1) All direct sensation of the fatiguing colour is abolished, not merely from the corresponding part of the spectrum, but also from those regions in which it overlaps other colours. It is rather the sensation of the particular colour that is weakened than the sensitiveness of the eye to a particular part of the spectrum. Thus, in red fatigue, the red end is shortened, but the green appears with full intensity at the D line and reaches almost as far as the C line. In green fatigue the green is replaced by red and blue, which not only meet but overlap. (2) The fatiguing colour produces a homochromatic after-effect like a luminous fog, by which the hue of all the other colours is modified if they are relatively weak, but which is unnoticed if they are bright. " A strong light not only fatigues the eye, but dazzles it ; that is to say, the sensation of light persists after the source of light has been withdrawn " (Burch). Each " dazzle-tint " has its own rate of develop- ment and of subsidence, and each is independent of the rest. Red passes oft" first, then green, then blue, and last violet. The positive after-effect of light between the C and D lines is at first red, but it comprises two dazzle-tints — red and green, of which red soon vanishes, leaving green. The dazzle-tint becomes unnoticeable long before the colour sensation is fully restored. (3) Fatigue with one colour has no effect upon the intensity of the remaining colour sensations. Fatigue with intermediate portions of the spectrum produces inter- mediate effects. Thus fatigue with orange or yellow affects not only the appearance of the orange or yellow, but that of the red and green also ; similarly with the blue-greens and the indigo. By fatiguing with red for 30 seconds and then finding pure yellow, fatiguing with E for 30 seconds and then finding the beginning of blue, fatiguing with F for 30 seconds, and then finding the beginning of violet, and again working backwards on the same plan, it has been possible to map out the range of the colour sensations. The averages of 70 people with normal colour vision give the extent of red = 760-6 to 551 /x/x. green = 593 to 484 „ blue = 517 to 443 „ violet = 454 to 397-5 „ THE EFFECTS OF "FATIGUE" 115 " In reality each colour sensation extends so much further that red and blue overlap, and so do green and violet, the only colours which I have not proved to do so being red and violet " (Burch). From the data in Clerk-Maxwell's papers and from his own observa- tions Burch estimates the transition point between Maxwell, red and green, 583"7 ; Burch (average of 70 persons) 573 ; ,, green and blue, 500*3 ; Burch (average of 70 persons) 500*3 ; „ blue-indigo = 449"8 ; Burch, blue and violet = 448. Pure colours — - Maxwell— red = 630-9 ; green = 529-7 ; blue = 457-4. Burch (average of 70 persons) — red = 625 ; green = 525 ; blue = 471 ; violet = 415. Burch therefore thinks there are usually four colour sensations — red, green, blue, and violet. He is unable to detect the existence of a separate yellow by this method. It should be pointed out that these are facts independent of any theory of colour vision, but it must be remembered that the fatiguing lights were of great intensity and might well produce changes which are quasi-pathological. Edridge-Green and Marshall ^ controvert some of Burch's statements, especially those relating to fatigue for the D line. Their experiments differ from his in that the intensity of the fatiguing light was much less. They find that after exposure of the eye to the sodium flame for from three to fifteen minutes and then looking at the spectrum the yellow is entirely obliterated and only a faint band of orange separates the red and green. If the eye is still further fatigued this also is obliterated, the red and green meeting. The red looked rather more purple, the green bluish. The blue and violet appeared diminished in intensity. There was no shortening of the red end of the spectrum. It should be noted in this connection that Burch does not deny the existence of a yellow sensation in some people, but states that he finds no evidence of it himself. On the other hand Edridge-Green and Marshall's experiments do not produce nearly so complete a fatigue of the sensations — whether true yellow (according to them) or red and green (according to Burch) — as in Burch's experiments. A. W. Porter and Edridge-Green^ have also investigated the effects of fatigue with spectral lights. The eyes were fatigued for about 20 sees. with monochromatic light and then fixed a spectrum of weaker intensity 1 Tram. Ophlh. Soc. xxix. 211, 1909. 2 Proc. Roy. Soc. Lond. B. lxxxv. 434, 1912. 8—2 116 COLOUR VISION so that the after-image formed a band across the middle of the spectrum. After fatigue with 654 — 675 /x/x the extreme red was slightly diminished, there was no change in the orange, yellow, or green, and the blue and violet became darker and bluer. If only red and orange were viewed the red disappeared and the orange remained. The sodium flame appeared unchanged (Porter), slightly greener (Edridge-Green). After fatigue with orange, 619 — 631 /x/u,, the dark blue after-image was seen right across the spectrum except in the region of the orange, which appeared unaffected. If only red and orange were viewed the red was replaced by the green-blue after-image, whilst the orange appeared unaffected. After fatigue with orange-yellow, 585 — 595 /x/x, the purple after-image appeared as if painted over the spectrum, the red being affected most. Fatigue with yellow-green, 545 — 550 /x/x, gave a similar result, with least effect in the orange. Fatigue with blue-green, 496 — 500 /x/A, gave the same effect,a purple (not red) after-image, with orange least affected : the sodium flame appeared unchanged (Porter), slightly redder (Edridge-Green). After fatigue with blue, 478 — 480 /x/x, there was no change in the red and orange, and the reddish purple after-image was seen over the rest of the spectrum. With blue, 445 — 455 /x/i, the after-image was yellow-green (Porter), orange (Edridge-Green) ; the violet and blue were cut off, red and green became yellower ; if only green and red were viewed the former was obliterated, the latter un- affected. With violet, 425 — 436 /x/x, the after-image was green ; the red and orange were unchanged, the green became yellow-green, and the violet and blue appeared green. The authors found that ' ' the stability of the after-image was remark- able ; it did not change colour, and was not influenced by subsequent light falling on the retina when this was not of too great intensity." Thus, if after viewing the reacting spectrum the eyes were turned to the dark screen the after-image appeared as a uniform dark band, darker than the screen. In contradiction to previous observers the authors found that " the complementary to the exciting light is never strengthened in the spectrum on the screen by the after-image." They also found no change in colour of the after-image as it fades. Wanach^ found that the eye fatigues more quickly for some spectral colours than for others, and least for the more refrangible. Macular fatigue lasts longer than peripheral. The consideration of Abney's researches on colour fatigue will be postponed (Part III). ^ Ztsch. /. Sinnesphysiol. XLni. 443, 1908. SECTION VI AREAL EFFECTS. CHAPTER I THE LOCAL QUANTITATIVE EFFECT Some effects of the size of the area stimulated have already been touched upon {v. pp. 51, 79, 95). They are of considerable importance and demand more detailed consideration. With regard to foveal vision it may be said that given a sufficiently intense illumination for a given condition of adaptation a mathematical point of light will be visible. Its image on the retina is always a diffusion area, the central parts of which afford at least a minimal effective stimulus if they anywhere impinge upon a retinal cone. If the effective area is so small as to occupy an interconal space a slight movement of the eye must be predicated in order that the point may be visible. It is probable that for a given condition of retinal adaptation a subminimal stimulus for a single cone may become a minimal or eft'ective stimulus if spread over several cones. Ricco^ conducted a very careful series of experiments bearing upon this point. The experiments were carried out in six different ways^. He found that at the threshold of sensibility the quantity of light entering the eye is constant, or, in other words, the light intensity and the area of the retinal image are reciprocal functions, or the product of the area into the light intensity is constant. In terms of the visual angle, the law is that the minimum visual angle varies inversely as the square root of the light intensity, or the product of the minimum visual angle and the square root of the light intensity is constant. The limit of the law is determined by the size of the foveal region ; it ceases to be accurate for visual angles above 40' to 50'. 1 Ann. di Ottal. vi. 1877. 2 Parsons, Roy. Lond. Ophth. Hasp. Rep. xix. 1, 114, 1913. 118 COLOUR VISION Charpentier^ published a series of experiments which confirm and extend those of Ricco. He used small bright squares up to 12 mm., viewed at 20 cm. distance. Below 2 mm. the smaller the surface the greater the minimum illumination necessary for perception. Two millimetres at this distance correspond to about 0"17 mm. on the retina, i.e., about the size of the fovea. For larger areas the area has no effect. It follows, therefore, that in order to produce a luminous sensation at the fovea the total quantity of light, i.e., the product of the area and the illumination, must attain a certain value, and that that value is constant for a given condition of adaptation. " The fovea centralis forms a sort of autonomous territory, in which the luminous excitation diffuses itself, and which always requires a certain quantity Of light to be set in activity." Charpentier showed, in answer to criticisms by Leroy, that the diffusion could not be accounted for by irradiation due to dioptric aberrations. Asher^ found that for the range of light intensities used by him up to a visual angle of 2' to 3' the apparent size depends entirely upon the quantity of the light. According to him, therefore, vision of objects subtending angles up to this size is a function purely of the light sense and not of the form sense. The earliest observations bearing upon this aspect of the subject were by Volkmann (1863) and Aubert (1865). Aubert used lines 2 mm. wide and 50 mm. long, and determined the distance the lines had to be apart in order that the interspace might look the same as the breadth of the lines. He found that when the breadth of the lines was varied by Volkmann's macroscope so that they subtended visual angles of from 10" to 45" the angular distance apart of the lines varied from 104" to 112" for black lines on a white background, and from 140" to 153" for white lines on a black background. Asher used small black, white, and grey squares and rectangles of paper and determined the distance at which a difference of size could be detected, the light intensities of the papers being calculated by the colour top and by Hering's polarisation photometer. He found that visual angles between 23" and 78" might be increased by 25" to 100" (average 58") under the given differences of light intensity (from 360 to 6) before a difference in size could be detected. At a great distance from the different sized objects they appear either of equal size and brightness, or, if the objective brightness of the smaller is much greater than that of the larger, of equal size but unequal brightness, the smaller being the 1 Compt. rend. xci. 1880 ; Arch. d'Ophf. n. 234, 487, 1882. 2 Ztsch.f. Biol. XXXV. 394, 1897. THE LOCAL QUANTITATIVE EFFECT 119 brighter. >^metimes the smaller appeared the larger. On approaching the objects the difference in apparent brightness increased whilst the sizes remained equal. Then followed a stage in which the difference in apparent brightness diminished, the sizes remaining equal ; both might even appear of the same brightness. On still further approximation the difference in brightness again became manifest, and simultaneously or shortly afterwards the larger object showed indefiniteness of the edges and greater apparent size. In many cases the size and brightness remained the same until the difference in size became distinguishable. Asher's explanation is as follows : So long as the objects subtend so small a visual angle that they cover a single sensibility area^ they appear equal, since the same quantity of light acts upon the sensibility area. With increase of the visual angle the influence of aberration becomes manifest. The larger object has a larger light area which is larger than a sensibility area, but the periphery of the light area has so low an intensity that the effective light area is not larger than a sensi- bility area. The smaller object, on the other hand, has a smaller light area, the effective part of which, however, is as large as a sensibility area. The ordinates corresponding to an effective brightness may extend farther from the centre for the smaller than for the larger object, so that the smaller may appear the brighter. So long as the relationship of the objects is such that the sensibility areas of both cover one or an equal number of sensibility areas they must appear of equal size, though the aberration areas may be very different. The conditions of light and contrast may easily be such that the smaller sensibility area may belong to the larger aberration area. Asher denies that any proof has yet been given that it is possible to produce a retinal image so small as to stimulate only one cone. Schoute^ does not agree with Asher that it is impossible to stimulate a single cone, and he holds that for single cone images the impression of size is dependent solely upon the product of the area into the light intensity. If only one cone is stimulated the object always appears 1 Mach (1866) first distinguished between the physical distribution of light over the area of a retinal image and the physiological distribution of brightness over the same area. If the retina be imagined flattened out and ordinates erected upon it, the lengths of which correspond with the intensities of light at the given spots, the area obtained by joining the summits of the ordinates will give an area representing the light intensity, or more briefly a " light area." If the ordinates represent the apparent brightness of the light at the spots as seen by the observer, the area will represent the sensibility and is briefly termed a " sensibility area." 2 Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. xix. 252, 1899; Ztsch. f. Augenhlk. vm. 419, 1902. 120 COLOUR VISION round, but differences in size are still appreciable. If the objects are of equal size the brighter appears the larger, i.e., the apparent size varies with the light intensity. Schoute, like v. Helmholtz, attributes this to psychological causes, i.e., to an error of judgment. Loeser^ has entirely confirmed Ricco's law for foveal vision. The law is stated in one form thus : The product of the minimum visual angle and the square root of the light intensity is constant. Loeser's results are shown in the following table : )istance of Diameter of Visual angle. Sq. rt. of light Product object, E object, D D/E intensity, J DJ/E m. mm. 8 200 2-5 0-87 2-18 >» 14-0 1-75 1-27 2-22 9? 8-5 1-06 ^•4 2-5 9« 50 0-63 3-45 2-26 The photochromatic interval {v. p. 60) can be demonstrated by altering the area of the retina stimulated. Thus Bonders^ found that in full daylight the hue of intensely coloured papers on a dark back- ground could be distinguished when they subtended a visual angle of 0"7 minute (1 sq. mm. at 5 metres distance). The dependence of the discrimination of hue on visual angle is readily demonstrated with relatively unsaturated colours. Coloured objects on a white background appear dark or grey under the smallest angles at which they are visible. It is possible to select a grey background of such a luminosity that the colourless interval is abolished, i.e., as the visual angle is increased the colour of the object is recognised as soon as the object becomes visible. Charpentier^ determined the absolute and the chromatic thresholds from the area stimulated for the fovea of the dark-adapted eye. Since there is little adaptation at the fovea, this factor is of relatively slight importance. He obtained the following results, the measurements being the diameters of the diaphragm of the photometer. Colour (sunlight spectrum) Extreme red Orange Yellow Green Blue . . 1 Hirschberg's Festschrift, 1905 2 Ann. d'ocul. LaXIX. 1878. Absolute Chromatic Ratio Threshold Threshold 05 mm. 1 mm. 4. 0-9 „ 21 =, 5-5 1 3-1 „ 9-6 0-3 „ 4-2 „ 196 0-3 „ 7-5 „ 625 Feilchenfeld and Loeser, Arch. f. Ophth. lx. 97, 1905. 3 La Lumiere ct les Couhurs, pp. 213, 238. THE LOCAL QUANTITATIVE EFFECT 121 Conversely, diminution of the area stimulated causes the colour to lose in apparent intensity — red less than green, green less than blue. The law associating area stimulated and intensity is therefore not so simple as that for white light. Abney^ in his experiments on the extinction of colour and light, i.e., on the point in the diminution of the intensity of light which just causes, first the colour, and then the light to become invisible, made a series of investigations on the influence of the area stimulated. He found, as was to be expected from the results of previous observers, that the smaller the disc the less reduction in intensity of the ray was required to extinguish it and the same ratio existed between the extinction of the different colours. Plotting curves with aperture diameters in powers of 2 as abscissae and logarithms of light intensities as ordinates, with apertures less than 1| inches diameter the curves become straight lines, all of which are parallel. Hence " from that point the intensity of a light which will be just extinguished with a certain diameter of aperture may be increased 10 times and yet be invisible when an aperture with one quarter of that diameter is employed ; if the intensity of the light be increased 100 times, we have only to diminish the diameter of the aperture to y'^ and it will again disappear, or if to -^}-^, the light may be increased 1000 times." When the angidar aperture exceeds 4° apparently the upper limit is reached, all extinctions being the same beyond it. With regard to the point of extinction, Abney^ says : " The light from a square, or a disc, or an oblong, just before extinction, is a fuzzy patch of grey, and appears finally to depart almost as a point. This can scarcelv account for the smallest width of an illuminated surface determining the intensity of the light just not visible ; but it tells us that the light is still exercising some kind of stimulus on the visual apparatus, even when all sensation of light is gone from the outer portions. The fact that the disappearance of the image takes place in the same manner whether viewed centrally or excentrically tells us that this has nothing to do with the yellow spot, or fovea, but is probably due to a radiation of sensation (if it may be so called) in every direction on the retinal surface. Supposing some part of the stimulus impressed on one retinal element did radiate in all directions over the surface of the retina, the effect would be greatest in the immediate neighbourhood, and would be inappreciable at a small distance, but the influence exerted upon an adjacent element might depend not only on its distance, but also upon whether it was or was not itself excited independently. 1 Abney, pp. 169, 174. 2 Ibid. p. 177. 122 COLOUR VISION Following the matter out further we should eventually arrive at the centre of an area as the part which was the recipient of the greatest amount of the radiated stimuli, and consequently that would be the last to disappear. With a slit aperture the slit is visible till extinction is very nearly executed, but it finally merges into a fuzzy spot at the moment before it finally fails to make any impression of light." Fig. 40 shows that there is an angular aperture or size of retinal area stimulated at which any ray will be extinguished both for colour and light at the same time. The aperture is largest for red. -a -J -4 -5 -6 Diameter of apertures xrv powers of z. -7 Fig. 40. Extinction of colour and light curves with different areas of stimulation. SSN 44 = 548-l/iM; SSN 50-0 =i) line: I and II, different observers and different in- tensities. (Abney.) Abney has also compared the luminosity of the spectral light from apertures of different size. Fig. 41 shows the curves for scale numbers THE LOCAL QUANTITATIVE EFFECT 123 27"3 and 50'6 (yellow D line, 589 ju/x), the logarithms of the annulus values of the two apertures being plotted as abscissae and ordinates. For points on the curve the luminosity of the large aperture is equal to that of the small, equality being obtained by means of the annulus. If the luminosity of the spots of light were always equal, irrespective of size, the inclination of the curves would be 45°. I <3 I '•J I _J ^ •J ^ / ,^ ^ y^ jy ^ ^* ^ ^ • ^ k ^ + / 0 I a Log readings with large aperture 3 Fig. 41. Equality of brightness curves for different sized apertures. SS:S 27-3 =575 M/" ; SS'S 50-6 =Z> line. (Abney.) The relation of area stimulated to intensity for the jperi'pliery of the retina has been studied by Piper^, Loeser^, Henius^, and Fujita*. Piper gives the following table for the dark-adapted periphery : Area 1 10 25 100 x^Area or Angular size 1 315 5 10 Relative Product of angular Threshold stimulation size and threshold value value value 10-0 1 100 2-94 3-4 9-3 1-96 5-1 9-8 102 9-8 10-2 ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxxn. 98, 1904. 2 Hirschberg's Festschrift, 1905 ; Feilchenfeld and Loeser, Arch.f. Ophth. hx. 97, 1905. Ztsch. f. Sinnesphysinl. xLiii. 99. 1008. * Ibid. XLHi. 243. 1908. 124 COLOUR VISION There is therefore obviously not the same relationship between area and threshold intensity that appertains to the fovea. There is, how- ever, a relationship between the square root of the area (which is equivalent to the visual angle subtended by circular objects) and the threshold intensity. Piper states the law thus : For the dark-adapted periphery the stimulation value of a luminous surface is proportional to the square root of the area of the retinal image ; or in other words, the product of the threshold value and the square root of the area of the retinal image is constant. Henius and Fujita showed that the law is valid only for mixed white light and objects subtending a visual angle up to 10°. Above 10° the sensitiveness of the retina increases much more slowly than the visual angle increases, and still more slowly for red light. Fujita showed too that the law is not valid below 1°. For the light-adapted periphery Fujita showed that there was no simple relationship between area and threshold intensity. For small objects, however, subtending less than 2° the threshold value diminishes with the visual angle, but not so quickly as for the fovea. Owing therefore to the great difference between the photopic and scotopic periphery threshold values, especially for relatively large luminous areas, the condition of adaptation is of great importance, especially as it itself changes with the actual stimulation. The relationship of the region of the retina and the size of the area stimulated respectively to the sensations derived from colourless mixtures and equations have been studied by v. Helmholtz^, v. Frey and V. Kries^, and Hering^. Hering observed colourless mixtures of complementary monochromatic lights of a given area and studied the effects of altering the size or shifting the fixation point. A mixture of red and blue-green became redder and darker on diminution, greener and brighter on augmentation of the area. The foveal colourless mixture became greener and brighter on direct fixation ; the peripheral colourless mixture became redder and darker on central fixation. A colourless mixture of yellow-green and violet became yellow-green on diminishing the area, rose-red on increasing the area or observation by indirect fixation, but the effects were much less than in the first case. Hering found that colourless equations with spectral lights become invalid, both with light and dark adaptation, if the area is increased or fixation changed from direct to indirect (contrast p. 69). So too in a match between a mixture of spectral red and blue-green (A) and yellow 1 1st ed. p. 301. 2 jirch. f. Physiol. 330, 1881. 3 Arch. f. d. ges. Physiol. Liv. 277, 1893. THE LOCAL QUANTITATIVE EFFECT 125 and blue (B), A becomes green and brighter, B lilac and darker. These changes are probably due to macular absorption. Tschermak^, for example, was unable to distinguish definite variations in parafoveal matches on increasing the eccentricity of fixation, thus confirming earlier results by Hess and v. Kries. Under such conditions, however, he found slight changes on increasing the size of the area stimulated, and these were of the same kind as those found by Hering. The effect of the size of the area stimulated on the photochromatic interval at the periphery and on the size of the field of vision has already been discussed (p. 79). CHAPTER II SIMULTANEOUS CONTRAST OR SPATIAL INDUCTION We have seen (Section V, Chap, ii) that stimulation of a retinal area is followed at a certain interval by sensations, potential or kinetic, of an opposite nature, a phenomenon known as " succedaneous " or " succes- sive contrast." There are numerous facts which tend to show that the sensation resulting from stimulation of a retinal area is modified by the condition of the surrounding areas, and per contra that stimulation of a retinal area itself causes a change in the condition of excitability of the surrounding area. This reciprocal interaction of retinal areas, or, in terms less open to criticism, of contiguous or nearly contiguous areas in the field of visual sensations is called " simultaneous contrast." The term contrast (Chevreul) in each case indicates opponent activities. Less open to objection is the term " induction " (Briicke), though this term has unfortunately been used in a more restricted sense by Hering. He uses " simultaneous induction " to connote change during fixation, " successive induction," change occurring after removal of stimulation. We shall use " induction " in the broadest sense, temporal induction being any change in a visual area due directly to stimulation (and there- fore including adaptation), and spatial induction being any change in other areas resulting from the primary stimulation. To avoid circum- locution we shall speak of " retinal " areas, without prejudice as to the exact site at which the change manifests itself. Spatial induction manifests itself by changes both in luminosity (or brightness) and hue. A patch of grey paper on a white background 1 Arch. f. d. ges. Physiol. Lxxxn. 559, 1900. 126 COLOUR VISION looks darker than a patch of the same paper on a black background (Goethe). Innumerable experiments have been devised to demonstrate luminosity (or brightness) contrast. Especially instructive are those with Hering's double-room or double-screen method, in which an aperture in a screen is variously illuminated from behind, whilst the illumination on the front of the screen can also be varied independently at will. It has been recognised since the time of Leonardo da Vinci that coloured surfaces are altered in hue by their surroundings, and vice versa. Briicke called a coloured surface surrounded by white the inducing colour. It must, however, be well borne in mind that the phenomenon is reciprocal, no matter what the luminosity or hue of the two fields may be. Tschermak^, who has written an excellent resume of our knowledge of contrast, enumerates five methods of demonstrating colour contrast. (1) The background method was early studied by Chevreul^. In its simplest form white, grey, black, or coloured patches or strips are laid on coloured backgrounds. The contrast is accentuated by placing tissue-paper over both^. Maxwell's top can also be used (Dove, V. Helmholtz, H. Meyer, Hering^, and others), or a projection apparatus (Rollet). (2) The mirror method, used by Goethe, is familiar in Ragona Scina's well-known experiment, which has been modified and improved by Hering''^, (3) The method of coloured shadows dates from Goethe, and has been much used, notably in recent years by Hering*" and Abney for the detection of colour-blindness. (4) The double-image method was introduced by v. Helmholtz'^ and modified by Hering^. (5) Negative after-images have also been used by Hering^. Chevreul, Fechner, and Wundt^° showed that actual contiguity of the reacting surfaces was not essential. Aubert^^ showed that weak and localised stimulation altered the excitability of the whole retina. The 1 Ergebnisse d. Physiol, ii. 2, 726-798, 1903. ^ De la Loi du Contraste simultane des Couleurs, Strassburg, 1839 ; new ed. Pa.ris, 1890. 3 Joh. Miiller, Handb. d. Physiol, n. 1837 ; H. Meyer, Ann. d. Physik. xcv. 170, 1855. « Arch. J. d. rjes. Phy.^iol. xli. 1, 1887. * Ibid. xli. 358, 1887. « Ibid. XLii. 119, 1888. ' 1st ed. p. 406; 2nd ed. p. 559. * Arch.f. d. ges. Physiol, xlvii. 236, 1890 ^ Sitz. d. Wiener Akad. 1872-4. " Philos. Stud. IV. 112, 1887. " Physiol, d. Netzhaut, 1865. SIMULTANEOUS CONTRAST OR SPATIAL INDUCTION 127 extent of the background, as well as its luminosity and colour, modifies the luminosity and saturation of the contrast colour. The contrast effect is not uniform over the whole surface, but is most marked at the edges. A distinction is therefore sometimes made between " surface-contrast " {Fldchenkontrast) and "border-contrast" {Randkontrast). The disparity diminishes on prolonged fixation (" simul- taneous induction," Hering). Border contrast is easily demonstrated with black and white, or coloured sectors on a rotating disc (Mach^, Aubert, v. Helmholtz), or by the shadow method (Goethe, A. Fick^). It is visible by momentary stimulation and is reversed in the negative after-image (Mach) : in the latter case it may be more obvious than in the primary image. With colour-contrast the saturation is greater at the borders than in the middle. Contrast effects are diminished by separation of the contrasting fields, and are therefore seen better with small than large fields. A narrow black line between the fields diminishes contrast markedly (v. Helmholtz), and the effects are greatest at contiguous contours (Mach). Jurin (1783) and Brandes (1827) attempted to explain simultaneous contrast by successive contrast, due to slight movements of the eyes. Though this factor is a frequent complication and requires special heed to its elimination it is not the cause (Fechner^, Hering). With the greatest care in controlling the movements of the eyes and keeping accurate fixation, simultaneous contrast manifests itself immediately. Moreover it occurs with colours of such weak saturation that no coloured after-image is induced. Aubert, Mach, Meyer'* and Rollet^ have shown that it occurs on momentary stimulation. Attempts have been made to obtain quantitative measurements of simultaneous contrast by means of a comparison field by Lehmann^, Ebbinghaus'^, Hess and Pretoria and others. For black and white or pure luminosity (brightness) contrast Ebbinghaus found that the increase in brightness of a patch on a darker background is proportional to the difference of the two light intensities, irrespective of their absolute values. Hess and Pretori found that the apparent luminosity of a small bright surface on a dark background ^ Sitz. d. Wiener Aknd. Lii. 2, 303, 1865. 2 Hermann's Handh. d. Physiol, iii. 1, 3, 1879. 3 Ann. d. Physik xliv. 513, 1838 ; L. 193, 427, 1840. ^ Amer. J. of Sc. XLVi. 1, 1893. ^ Sitz. d. Wiener Akad. lv. 424, 1867. « Wundt's Philos. Stud. iii. 497, 1886. ' Sitz. d. Aknd. d. Wiss. Berlin, 994, 1887. » Arch. f. Ophth. xl. 4, 1, 1894. 128 COLOUR VISION remains constant when the increase of illumination of the surface bears a constant ratio to the increase of illumination of the background. The contrast effect therefore varies directly with the inducing stimulus and is independent of its absolute value. Ebbinghaus and Hess and Pretori agree therefore as to increase of brightness by contrast or contrast-brightening. Hess and Pretori find that the law also applies to contrast-darkening, therein differing from Ebbinghaus, but the results of the latter are probably due to error in technique. Kirschmann^ found that the amount of contrast varies as the square root of the area of the inducing field, i.e., the relationship between the area and the intensity of the inducing light is reciprocal. For colour-contrast Fechner showed that the brightness of a grey background must be appropriately chosen. Under optimum conditions an inducing colour of very slight saturation suffices to induce marked contrast. The statement, however, that simultaneous contrast is most evident with slight differences of colour between the reacting fields (v. Helmholtz) is not true (Fechner, Rollet). Hering^ showed that with optimum black-white background conditions contrast increases with the saturation of the coloured inducing field, an observation confirmed by Pretori and Sachs^. Kirschmann (1892) had previously arrived at the same result, but found that the increase was not proportional, but probably a logarithmic function. Colour-contrast, as might be expected, is accompanied by change in luminosity, and this subject has been exhaustively studied by Pretori and Sachs under , Bering's guidance. Their results will be considered in Part III. The most important element in colour-contrast from our present point of view is the opponent or complementary effect, early noticed by Briicke*. Under ordinary circumstances the induced colour is not the precise complementary of the inducing, as was known to Goethe, and has been fully investigated by Hering^. Various physical factors conduce to the effect, such as the variable nature of "white" light and " grey " surfaces, macular pigmentation (Hering, Sachs), the pig- ment of the lens (Hering), the brown pigment of the retinal epithelium (Tschermak), and the reddening by the blood of the light which passes through the sclerotic (Hering, Hess). These do not suffice to explain all the facts. There are other, physiological, factors, such as the 1 Wundt's Philos. Stud. vi. 417, 1890 ; vn. 362, 1891 ; Amer. J. of Psychol, iv. 4, 74, 542, 1892. 2 Arch.f. d. ges. Physiol. XLi. 1, 1887 ; XLii. 117, 1888. 3 Ibid. LX. 71, 1895. * Ann. d. Physik, lxxxiv. 418, 1851, * Zur Lehre vom Lichtsinne, 1876. SIMULTANEOUS CONTRAST OR SPATIAL INDUCTION 129 previous stimulation of the eye (Hering, Hess^, Tscherniak) and fatigue (Tschermak and Krause). The discrepancy between the induced and the complementary colour is said to be absent when the eye is dark- adapted (Mayer, Kuhnt^). There are also psychological factors, which form at any rate the most probable explanation of simultaneous con- trast over colour-scotomatous areas (Tschermak^) and over the blind spot. The entoptic visibility of the blind spot* is itself evidence of contrast, as are also the effects of stimulation of the optic nerve by the constant current (G. E. Miiller^). Simultaneous contrast is under ordinary conditions limited to the eye stimulated (Hering^), but binocular contrast can also be proved to occur'^. 1 Arch.f. Opkth. xxxv. 4, 1, 1889 ; xxxvi. 1, 1, 1890. 2 Loc. cit. xxvn. 3, 1, 1881. 3 Arch.f. d. gca. Physiol, i.xxxii. 559, 1900. * Brewster; Puikinje; Aubert ; v. Helmholtz; Cliarjjentier, Compt. rend, cxxvi. 1()34, 1898. — ^ Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. xiv. 329, 1897. « Loc. cit. I. 18, 1890. ' Fechiier ; H. Meyer ; v. Helmholtz ; Hering, in Hermann's Hundb. d. Physiol, in. i, (JOO, 1879 ; Ebbinghaus, Arch.f. d. ges. Physiol. XLVi. 498, 1894 ; Chauveau, Compt. rend. cxiii. 1891 ; I). Axcnfeld, Arch. ital. de Biol. xii. 28, 1889 ; xxvii. 103, 1897 ; Burch, Jl. of Physiol, xxv. 1900. P. C. V. SECTION VII THE EVOLUTION OF COLOUR VISION CHAPTER I INTRODUCTION A priori we should expect some light to be thrown upon the fully developed colour sense of man by a knowledge of the stages through which that colour sense has evolved. The sources of our information on the evolution of colour vision are few and the methods of in- vestigation difficult and arduous. Only recently have the researches been carried out in a scientific manner and yielded valuable results. Positive evidence is derived from three chief sources. In the first place we naturally appeal to the visual sensations of lower animals. These are extremely difficult to investigate since we are almost wholly dependent upon observation of motor responses which the animals make to various light-stimuli, though some deductions can be made from the structure of the visual organs. In the invertebrata little can be done beyond recording the phototropism of the animal, i.e., its attraction or repulsion by lights of different wave-length and intensity, as exhibited by its movements towards (positive phototropism) or away from (negative phototropism) the light^. As we ascend the animal scale the increase in complexity of the nervous system and of the visual organs is accompanied by a corresponding increase in complexity of the motor responses, associated with a greater difficulty in their inter- pretation. On the other hand, as we descend the animal scale from man there is an unwarranted tendency to interpret the apparently purposeful responses of the animal in an anthropomorphic manner which is not necessarily justified on neurological and psychological grounds. For example, we have little knowledge of the psychology ^ Mast, Light and the Behavior of Organisms, New York, 1!)11. INTRODUCTION 131 of the lower mammal, with its less highly developed nervous system. The temptation to interpret such an animal's actions in terms appropriate only to the human mind has proved very great and has undoubtedly given rise to error in the past. Our deductions must of necessity be anthropomorphic, since such terms as visual sensation, attraction, repul- sion, pleasure, pain, and so on, have no meaning for us except in so far as these processes form a part of the contents of our own minds^. Yet it should be a guarded anthropomorphism, neither exaggerating the psychological elements nor flying to the impossible antithesis of imagining that the anthropomorphism can be eliminated by a new terminology. In the second place a study of the colour vision of primitive races may throw some light on the evolution of visual sensations. It may be that some primitive races are in a condition of arrested development — of vision, as of other faculties. We have only just crossed the threshold of this part of the investigation and it is to be hoped that no time will be lost in carrying it forward, lest the material for the research be obliterated by the march of civilisation. The third source of information is the development of visual sensa- tions in the infant. It is generally admitted that " ontogeny is a compressed phylogeny," — that each individual passes rapidly through the same stages of development which have marked the upgrowth of the race. Here again, little progress has been made, and the investiga- tion is arduous and full of pitfalls. Besides these main sources there are others of less security. We are familiar with congenital defects of vision, and it may be that some of them are atavistic, that development has become arrested at a stage which corresponds with an earlier stage in the development of the race. Some arguments too may be derived from the careful study of normal colour vision, but the evidence derived from both these sources is too uncertain to be of much value. CHAPTER II THE COMPARATIVE PSYCHOLOGY OF COLOUR VISION I shall briefly review the comparative psychology of vision in verte- brates only, laying particular stress upon points of theoretical interest. I Cf. Washburn, The Animal Mind, New York, li)OS. 9—2 132 COLOUR VISION The most recent and most exhaustive experiments have been made by Hess^. Mammals. Graber^ experimented on nine dogs by his " preference method." A j^oodle and a fox terrier showed definite preferences. When black and white were presented white was chosen 56 times, black 4 : when light red and blue, blue 53 times, red 7 ; when the same and a much darker blue, blue 32 times, red 28. The results may have been due to preference for the brighter colour. Lubbock^ experimented with a poodle. Three pairs of cards coloured blue, yellow, and orange, were used. One card from each pair was placed on the floor and the remaining card of one of the pairs shown to the dog in the hope that she would learn to select from the three cards before her the one that was of the same colour as that held up. Training for ten weeks proved entirely without success. Lubbock himself remarks that negative results prove nothing as to the colour vision of the animal. Himstedt and Nagel* taught a poodle to fetch coloured balls. On being ordered to fetch " red " it brought first scarlet or bright red, then orange, but only the latter when there were no more conspicuous red balls left. Balls coloured with Bismarck brown were a cause of difficulty to it. " Blue and grey were obviously quite different to it from red, even in all degrees of brightness of each." Elmer Gates^ experimented with dogs and food receptacles of various colours, including grey ; only receptacles of one particular colour actually contained food. In some experiments coloured metal plates were laid on a passage, certain colours being connected with a battery, so that if the dog stepped upon them it received a shock. Gates concluded from his researches that dogs could distinguish not only colours but relatively fine difl^erences of hue. The description is so meagre that it is impossible to determine whether due precautions were observed. Samojloff and Pheophilaktowa*" tested whether a dog could be taught by much training to distinguish between different colours as well as between different brightnesses. Three boxes were placed in a row ; 1 Vcrglcichende Physiolofjie des Gcsichisinnes, Jena, 1912 (Bibliography) : " Birds " — Arch. /. Augenhlk. lvii. 4, 298, .317, 1907; Lix. 2, 142, 1908 : " Fishes "—Zoc. cit. lxiv. Erganzungsheft, 1, 1909; Arch. f. d. ges. Physiol, cxxxiv. 1910; cxlii. 1911; Zool. Jahrb. 1912; " Eeptiles and Amphibia" — Arch. f. d. ges. Physiol, cxxxii. 255, 1910. - Grv.ndlinien zur Erforschung des Helligkeits- u. Farhciisiiines der Thiers, Prag, 1884. ' On the Senses, Instincts, and Intelligence of A)nmah, p. 280, Londoji, 1888. * Festschrift d. Univ. Freiburg, 1902. = The Moni.sf, p. 574, 1895. 8 Centralhl. f. Physiol, xxi. 133, 1907. THE COMPARATIVE PSYCHOLOGY OF COLOUR VISION 133 each had a disc in front, 25 cm. in diameter, of which two lield arev papers and one a saturated green. The " green " box contained a biscuit. After training it was found that out of fifty papers, ranging from black to white there was no grey which, when presented to the subject for discrimination from the standard green, yielded a percentage of errors greater than 3L Later tests showed that such discrimination only occurred under certain conditions. If the grey was presented as a square and the green as a circle, green was chosen : but if the green was a triangle or square, the grey square was invariably chosen. Hence the discrimination of a colour from a series of greys is only possible when the conditions have been thoroughly learnt. Nicolai^, did not succeed in teaching his two dogs to distinguish between red and green bowls. Colvin and Burford^ experimented in the same manner with three dogs, a cat, and a squirrel, food being placed in red receptacles amongst similar empty receptacles of other colours. The colour was distinguished in 87 '3 per cent, of trials, the squirrel responding best. Kinnaman^ attempted in a series of similar feeding experiments to eliminate the ambiguity of discrimination of brightness and colour-differences. Monkeys were tested with glass tumblers covered wdth papers of different colours, and when it had been shown that they were able to identify a vessel of a particular colour as associated with food, the power of distinguishing the colour from greys of the same brightness was tested. Kinnaman came to the conclusion that the capacity to distinguish colours as such in monkeys was undoubted. J. B. Watson^ criticising Kinnaman's results, says that the use of coloured papers can never give a satisfactory test of colour vision in animals. He himself used a spectrometer apparatus which illuminated a screen w^ith two monochromatic red and green patches. Food boxes w^ere placed beneath the patches, one being empty, the other containing a grape. If the red box were opened the grape was obtained ; if the green, the monkey was pulled back. In early tests with red and green the animals failed to react to red. Blue-vellow discrimination arose more rapidly than red-green in all cases (three monkeys). The experi- ments were early vitiated by the onset of position habits. Yerkes^ made very exhaustive experiments on the Japanese dancing 1 .//./. Psychol, u. Neurol, x. 1907. - Psychol. Rev., Psychol. Monographs, xi. 1. 1909. ■■' Amer. J. of Psychol, xiii. 98, 173, 1902. * ./. ofComp. Xeur. and Psychol, xix. 1. 1909. 5 y/,p Dancing Mouse, New York. 1902. 134 COLOUR VISION mouse. The method consisted in teaching the animals to associate one of two differently illuminated compartments with a disagreeable electric shock. Light blue and orange, green and red, violet and red, were distinguished even when their luminosities were considerably varied. The possibility that these discriminations were made from brightness rather than colour was not wholly eliminated. Yerkes concluded, how- ever, that the mice have a certain degree of ability to distinguish red, green, and violet as colours. No ability to discriminate green and blue was shown unless there was a great difference in brightness. There is evidence that the red end of the spectrum is much darker to the mouse's than to the human eye. Waugh^ on the other hand found that his mice could only distinguish red from white and light greys with great difficulty. Davis and Cole^ experimented on racoons by the feeding method and found that while discrimination of black from white, yellow, red, blue or green appeared easy, that of blue from yellow and red from green was difficult. Davis thought it probable that the animals were colour- blind, but Hess does not agree with this conclusion. Washburn and Abbott^ made attempts to discover the brightness value of red for the light-adapted eye of the rabbit. They arrived at the conclusion that this animal can discriminate between a saturated red and a grey paper, and that the discrimination is based upon luminosity rather than colour. Red has a low stimulus value for the rabbit. The experiments furnish no evidence that it sees red as a colour, but do not prove that it does not do so. An elaborate series of researches has been carried out in Pawlow's laboratory^ with a view to establishing " conditional reflexes." The salivary secretion was used as the indicator. An indifferent stimulus, e.g., a strong electrical stimulus to the skin, is at first applied simul- taneously with the placing of food or acid in the mouth. After a short time the salivary secretion is called forth by the application of the indifferent stimulus alone. Orbeli^ studied the effects of coloured optical stimuli. After exposure of a red square on a white screen reflex salivary secretion was obtained. The same recurred when other colours were exhibited. If red and grey squares were exhibited suc- cessively and the dog fed only after exhibition of the red square the ^ J. of Comp. Neur. and PsycJtoI. xv. 549, 1P05. " Cole, he. cit. xvii. 211, 1907. ^ J. of Animal Behavior, ii. 145. « Pawlow, Brit. Med. Jl. ii. 473, 1913. ^ Comptes rend dc la Soc. med. russe, St Petersburg, 1907 ; Arch, des Sc. hiol. xiv. 1908. THE COMPARATIVE PSYCHOLOGY OF COLOUR VLSION 135 salivary reflex eventually followed the red stimulus only. It was found, however, that the reflex in dogs was conditioned by the change in intensity of the light, not by variations in the wave-length. The experi- ments are not of a nature to demonstrate colour-blindness in the dogs. Kalischer^ has published interesting experiments on dogs. His animals were placed in a dark chamber which could be illuminated by means of light from a lamp, transmitted through coloured glass. On the appearance of red light the dog was given food, which was, however, withheld during the period when light was excluded. When a perfect reaction habit had been formed, the experiment was varied sometimes by transmitting the light through blue, instead of through red glass, though the subject was still only fed when a red light was exposed. In order to render the comparison of brightness between the different lights more difficult, their successive exposure was always followed by a dark interval. At first the subjects reacted in a similar manner to either coloured light, but after some further training the difference was duly learnt, and eventually discrimination became so perfect that no matter how the brightness of the red or the blue light might be changed the reaction was almost invariably correct. Observations were also made with green, yellow and red-violet lights, but discrimination was found to be the most definite when red and blue were used ; red and yellow proving very difficult to learn, and red with red-violet almost impossible. Colvin and Burford likewise found that when their subjects were required to discriminate between red and violet, only a relatively low percentage of right choices was obtained. In later experiments Kalischer left his subjects untested for two or three days, and then presented a blue or a green light previous to the exposure of the food-colour, red. Though these experiments were repeated many times and under varied conditions, the animals never failed to react correctly to the food-colour when it appeared. It is of importance to note that not only did the animals learn to discriminate, during the course of the experiments, between different colours irrespective of the degree of their luminosity, but that in addition the behaviour shown on the appearance of each colour was characteristic, e.g., on the exposure of a blue light the subject would appear frightened and quickly withdraw his head from the food ; whereas when a green or a yellow light was exposed great hesitation was shown, the head was allowed to fall gradually nearer and nearer towards the meat as though to seize it, and was then suddenly jerked away. 1 Arch.f. Anat. p. 316, 1909. 136 COLOUR VISION The conclusions which Kalischer draws from his experiments are as follow : (a) That there is no doubt of the ability of dogs to perceive differences of hue as well as differences of brightness. {b) That whereas such colour discrimination is cortical, brightness discrimination is to some extent sub-cortical, reactions to changes of illumination being found to occur even after extirpation of the visual cortex. (c) That there exist considerable individual differences in the sensi- bility of dogs to colours. By far the most exhaustive experiments on dogs have been carried out bv Miss E. M. Smith^. The apparatus used is shown in Figs. 42 and o o r-t _J L home-box h k\ 7E F\ u trial-box o o ^ ■window-door ^ Fig. 42. Ground plan of experimental dog box. (E. M. Smith.) 43. It will be noticed that the " home box " is separated from the " trial box " by a partition in which there are five glass screens Sj.^. Cards of various colours could be introduced into these screens. The dog was placed in the home box and let into the trial box through Sj, which was then closed. The preliminary " general training " con- sisted first in teaching the dog to pass through the screens. These 1 Brit. Jl. of Psychol, v. 119, 1912. THE COMPARATIVE PSYCHOLOGY OF COLOUR VISION 137 were left transparent and a biscuit was placed within view in the home box. Next, cards of medium grey were placed in S,., S:i, S4, and S5. Finally, only one shutter contained a grey card. The dogs were thus taught always to go to the opaque screen in order to get through into the home box and receive the reward. The various test series were then commenced. (1) "Colour prefer- ence series." Maximum blue, red, yellow and green cards were put in four screens and the frequency with which each was selected was noted in 70 tests. For one dog the numbers were red 32, blue 26, yellow 8, green 4. The red and blue were much darker to the human eye than the yellow and green. (2) " Approximate brightness value series." f- ^■^....■v -J^.'. ^ 'V — ■■■■"■■;^-^' fl t f I // ?■* ^ i I // i s L I I * / ii . . . ■.-^. $t . . . .1. .....t^ > .. 4,/ ... Fig. 43. Elevation of experimental dog box. (E. M. Smith.) Red, blue, yellow, green, and 14 shades of grey were used. First, the particular colour was presented simultaneously with the grey which was most remote from it in brightness. When the animal had given proof of its ability to discriminate between these two the next grey, differing only slightly less in brightness, was substituted. Second, all the greys, in the same order as before, were presented on the same day for dis- crimination from the standard colour, each grey being exposed 5 times in succession. Next, the colour and the greys were presented in irregular order. It was found that in the case of the most intelligent dog the average percentage of right reactions was much higher for green and yellow than for blue and red : further the first colour tested, blue, 138 COLOUR VISION had the smallest, and the last, green, the largest percentage of right cases. To test if this were due to general improvement from increased practice rather than specific qualitative differences the " repeat bright- ness value series " was given. In this series all four colours were re- tested with additional precautions. (3) " Hue discrimination series." (a) Coloured cards and equivalent greys were exhibited in pairs : and (6) two coloured cards were shown simultaneously, such that (o) the pairs were equal in brightness and saturation ; {/3) the standard colour was slightly or obviously darker than the other ; (7) the standard colour was slightly or obviously lighter than the other. (4) " Grey difference threshold series." (5) " Colour threshold series." Enough has been described to show that the tests were exhaustive, and elaborate efforts were made to determine the nature of the dis- crimination evinced. Perhaps the most serious objection which can be brought against the experiments is their extreme elaboration. Anyone who has had experience of testing the colour vision of untrained human observers knows the frequency with which luminosity and colour matches are confused. How much more this source of error is likely to be present in dogs needs only to be mentioned. It was allowed for to a certain extent by a graduated method of marking. As Miss Smith says, " In view of their relative difficulty, sensory discrimination tests on the higher vertebrates might aptly be compared to complicated intellectual tests on man, in which the subject, unable to grasp completely the point at issue, has to rely on a series of sporadic gleams." Only a summary of the chief results can be given here, the reader being referred for details to the admirably thought out original paper. In favour of the view that dogs are only able to perceive differences of brightness are (1) the result of the " colour preference series " ; (2) the greater the difference in brightness between a standard colour and grey, the more definite is the discrimination ; (3) the length of time required to establish a successful discrimination habit between blue and red ; (4) the fact that whereas change of colour frequently appeared to pass unnoticed, a marked change of brightness often caused confusion ; (5) the fact that green and yellow were not at all, or only with the greatest difficulty, discriminated from white ; (6) the fact that during a prolonged series increased confusion and uncertainty in discrimination were shown from day to day. In favour of the view that dogs possess rudimentary colour dis- crimination are (1) the decreased time for training required by each new colour in the " approximate brightness value series " ; (2) the unique THE COMPARATIVE PSYCHOLOGY OF COLOUR VISION 139 physiological effects produced by the difEerent colours even when of approximately the same brightness (greater affective value of red compared with blue, etc.) ; (3) the evidence afforded by the " approxi- mate brightness value series " that on the whole no grey was found that was indistinguishable from the standard colours employed, training being still imperfect ; (4) the difficulty and confusion of the " grey difference threshold series " as compared with the parallel " approximate brightness value series " ; (5) the results obtained from the " hue discrimination series " ; (6) the difference in the approximate thresholds of the four colours and grey as determined by the " colour threshold series." Whereas the facts recorded in the first group of items of evidence are not incompatible with the assumption that dogs possess some sort of colour vision, those in the second group are quite irreconcilable with the view that the subjects are limited to brightness vision. Miss Smith draws the following conclusions : (1) That while evidence has been obtained to show that some dogs possess a rudimentary power of colour discrimination, such discrimination is highly unstable and cannot be supposed to play any part in the animal's normal existence. (2) That the colour sense is, as shown by the high colour threshold, very weak. (3) That colour discrimination, even where clearly established, may be readily inhibited by differences of luminosity or position. Piperi investigated the electrical reactions of dogs, cats and rabbits. In dogs and cats he found the reactions in both light- and dark-adapted eyes to correspond with the achromatic scotopic values for man. In rabbits the strongest photopic value was 570 /x/z, scotopic 540 ixfi. Hess tested a monkey in the following mannner. Grain was scattered over a black surface, upon part of which a spectrum was thrown. The monkey gathered all the grains from the extreme red to the extreme violet, leaving those unilluminated. When the animal was dark-adapted and the intensity of the light diminished until the grains were visible to the dark-adapted experimenter only in the yellow and green, only these grains were gathered. Hess concluded that the extent of the visible spectrum is the same for the monkey as for man, and that it is brightest for the dark-adapted monkey in the region where it is also brightest for the dark-adapted man. Hess also studied the effects of different spectral lights and the lights transmitted by coloured glasses on the pupil reactions of cats and rabbits. 1 Arch. f. A)iat., Suppl. ]fl0.5. 140 COLOUR VISION It was shown by Sachs^ that in man the extent of the pupil reaction varies with the subjective luminosity of the light. Similarly AbelsdorfE^ showed that the maximal pupillomotor effect for the light-adapted eye was at 600 /x/x, for the dark-adapted eye at 540 [xjjl. Hess found the photopic and scotopic values for cats and rabbits about the same as for man. It appears therefore that mammals have the same spectral limits as man ; that they have a similar capacitv for dark adaptation ; and that their luminosity curves are similar. Birds. Lloyd Morgan^ experimented with chickens. A chicken which picked out bits of yolk of hard-boiled egg from the white was given pieces of orange peel, which it found distasteful. Afterwards for some time it was suspicious of yolk. On the other hand after having learned to avoid bad-tasting black and yellow caterpillars, it did not object to wasps, but probably points of difference other than colour were apparent to it. J, P. Porter* investigated the colour vision of the English sparrow and other birds by feeding experiments. A number of glasses of like size and shape were covered inside and out with coloured papers, including dark and light grey. The glasses were placed in a row on a board, and food was always put in the same coloured glass, the position of which was changed. The sparrow and cowbird learned to pick out the right vessel. Rouse^ investigated pigeons by Graber's method of allowing a choice between compartments illuminated by different coloured glasses. The pigeons showed a preference for green and blue. A greater rapidity of breathing was recorded under green and blue lights than under others. Hess experimented with birds of diurnal habits (chickens, pigeons) and nocturnal habits (owls) in the same manner as for mammals. For diurnal birds corn or rice grains were scattered over a matt black surface, part of which was illuminated by a spectrum. They usually pecked at the grains in the orange and red first, then those in the green and yellow, and some in the blue-green. Those in the green- blue, blue and violet were left untouched. Dark-adapted chickens 1 Arch.f. d. ges. Physiol, lii. 79, 1892 ; Arch.f. Ophth. xxxix. 3, 1893 ; Ztsch.f. Psychol u. Physiol, d. Sinnesorg. xxn. 388, 1900. 2 Zisrh. f. Psychol, u. Physiol, d. Sinnesorg. xxi. 451. 1900; Arch. f. Aiigenhlk. XLi. 155, 1900; Arch.f. Anal. u. Phy.<^iol. .561, 1900. 3 Habit and Instinct, Tjondon, 1896. ■* Amer. J. of P.^ychol. xv. 313, 1904; xvii. 248. 190G. * Harvard Psych. Studies., ii. 580, 1906 THE COMPARATIVE PSYCHOLOGY OF COLOUR VISION lil generally pecked somewhat farther towards the blue, but seldom in the blue and violet. When the spectrum was reduced until it was almost colourless for the dark-adapted human eye they pecked chiefly or solely in a region corresponding to the yellow and orange-yellow, i.e., rather red-wards of the human scotopic maximum luminosity. Chickens 48 hours after hatching behaved in the same way as older birds. Hawks {Falco tinnunculus), fed with pieces of meat, showed the same shortening of the violet end of the spectrum. All the birds first pecked in the region which is brightest for man, and then sought regions of less luminosity. If half the field were illuminated with light transmitted through a red glass and the other half with blue light, light- or partially dark-adapted birds pecked first in the red half, turning to the blue half only when the red was cleared of grains. By suitable alterations of the strengths of the lights, the blue half was left untouched, even though the red grains appeared to the human eye very dark red and scarcely visible whilst the blue were clearly seen. Blue grains, however, were as eagerly pecked as red if the light were made sufficiently strong. Dark adaptation increases the sensitiveness of birds' vision very markedly. The " peck- ing-limit " is for them only slightly, if at all, higher than the extrafoveal threshold visibility of the grains for a 1 — 2 hours' dark-adapted human eye. Adaptation appears to be slower in hawks and buzzards than in the human eye. Nocturnal birds were investigated in the same manner. If an owl, which has perched for several hours in sunlight, is examined with a moderately bright spectrum illuminating white pieces of meat, these are snapped up first in the red, yellow, green, or blue-green. The spectrum extends further violet-wards for nocturnal than for diurnal birds, but is still much shortened at this end. When thoroughly dark- adapted owls snap at meat in the blue and violet; they snap first in that part of the achromatic spectrum which is brightest for man, or slightly farther towards the violet. Dark adaptation causes a rapid rise of sensibility during the first half-hour and the maximum sensibility appears to be much greater than for man. The pupil reactions of birds are somewhat complicated. The maximum constriction for diurnal birds (fowl and pigeon) is in the region of the orange and yellow, diminishing gradually towards both ends of the spectrum, more rapidly towards the violet the less the dark adapta- tion. For the dark-adapted eyes of nocturnal birds the maximum is in the yellow-green and green, falling slowly towards the violet end. It appears therefore that diurnal birds see the spectrum shortened 142 COLOUR VISION towards the violet end, much as it would be seen by the human eye through a reddish yellow glass. The shortening is similar but rather less for nocturnal birds. Both types possess a power of dark adaptation the same as or very similar to that of man. Reptiles. Hess investigated several varieties of tortoise, and found that their vision nearly resembles that of birds. They have the same shortening of the spectrum towards the violet end. More remarkable is the fact that they possess a high degree of dark adaptation, the increase in sensibility being seemingly the same as that of the human eye looking through a suitable orange-tinged glass. It may be mentioned here that these animals are generally admitted to possess no rods or visual purple in their retinae. Hess noticed that the only animals which he examined and found to possess a shortened spectrum towards the violet end were such as possess coloured oil globules in their cones. At the same time he points out the difficulties in accepting this fact as a complete and satisfactory explanation. Amphibia. Yerkes^ caused a green frog to go through a simple labyrinth leading to a tank of water. At the point where the first choice between two paths occurred a red card was placed on one side and a white on the other. When the frog had learned to take the correct path the cards were exchanged and the confusion of the animals showed that they discriminated between the cards, though it may have been a luminosity- rather than a colour-discrimination. Loeb^ found that frogs showed negative phototropism, and preferred red to blue light. Torelle^ found that two species of frog showed positive phototropism, associated with a tendency to prefer blue to red light. The phototrop- ism persists when the animal is blinded, although in the normal animal the eyes are involved in the reaction since it occurs when the skin is covered and the eyes left intact*. Cole^ found that the phototropism of Rana clamata differs according to the surrounding temperature. Pearse^ found some amphibia positively, and others negatively helio- tropic. 1 Harvard I'.sijch. Studies, i. 570, 11)03. - Der HcUotropismuH der Tie.re, Wiirzburg, 1890. 3 Amcr. J. oj Phymol. ix. 466, 1903. * Koranyi, Centralbl.f. Physiol, vi. 6, 1892 ; G. H. Parker, Amcr. J. of Physiol, x. 28, 1903. 5 J. ofComp. Neur. and Psychol, xvii. 193, 1907. " Proc. of the Amcr. Acad, of Arts and Sc. xlv. 16 J, 1910. THE COMPARATIVE PSYCHOLOGY OF COLOUR VISION 143 Hess examined an American salamander {Diemictylus viridescens), the ordinary toad {Bufo vulgaris), and the African spurred frog {Xenopus Mulleri). He found that the spectrum does not appear shortened to them. He also found that their range of adaptation appeared to be the same as for the human eye. Fish. Graberi tested two species of fish, but no convincing proof of their powers of colour discrimination were obtained. Bateson^ placed food on coloured tiles and found that fish picked it ofE most readily from white and pale blue, and least readily from dark red and dark blue, a fact which can be explained by luminosity differences. Zolotnitzki^ fixed dark red Cheironomus larvae and white, green, yellow, and red pieces of wool of the same size and shape as the larvae to the wall of the fish tank. The fish showed decided preference for the red, swam past the white and green, while some of them paused at the yellow. Washburn and Bentley^ investigated the chub {Semotilus atromaculatus). Two dissecting forceps were used, alike except that to the legs of one were fastened, with rubber bands, small sticks painted red, while to those of the other similar green sticks were attached. The forceps were fastened to a wooden bar projecting from a wooden screen, which divided the circular tank into two compartments, and hung down into the water. Food was always placed in the red pair of forceps, which were made frequently to change places with the green ; and the fish was caused to enter the compartment half of the time on one side, and half of the time on the other. This was to prevent identification of the food fork by its position or the direction in which the fish had to turn. The animal quickly learned to single out the red fork as the one important to its welfare, and in forty experiments, mingled with others so that the association might not be weakened, where there was no food in either fork, and where the forceps and rubber bands were changed so that no odour of food could linger, it never failed to bite first at the red. More- over, the probability that its discrimination was based upon brightness was greatly lessened by using, when experimenting without food, a different red much lighter than that in the food tests. The fish success- fully discriminated red from blue paints in the same way, and it was ^ Grundlinien znr Eijurschuufj des HelUgkeUs- u. Farbcn/iiimes dcr Thierc, Prag and Leipzig, 1884. - Jl. of the Marine Bwl. Afi.soc. U. K. i. 225, 1887. 3 Arch, de Zool. exper. ix. 1901. * J. of Comp. Near, and Psychol, xvr. 11.3, I'JOG ; Waslihuin, The Animal Mind, New York, 1908. 144 COLOUR VISION afterwards trained by putting food in the green fork, to break the earlier association and bite first at the green. Reighard^ fed grey snappers {Lutianus griseus) with small Atherinae which were artificially coloured with seven hues. On simultaneous presentation of white and blue the white was generally seized first, with blue and light or dark red or yellow the blue first ; no difference was apparent between blue and green. If Cassiopea tentacles were fastened to the Atherinae the snappers soon learnt to avoid them. After they had learnt to avoid red Atherinae with tentacles it was found that they also rejected red ones without tentacles, but that they still took those stained any other colour. The researches of Zolotnitzki, Washburn and Bentley, and Reighard are consistent with the view that the fish are totally colour-blind. Hess examined a number of species of fish and found that they appeared to behave exactly as if they had the visual perceptions of a totally colour-blind man (see Part II). In the light of the spectrum young Atherinae rapidly congregated in the yellow-green and green, i.e., the brightest part of the human achromatic scotopic spectrum. Far fewer were found in the yellow and scarcely any in the red ; there were rather more in the blue. If the spectrum were displaced the fish rapidly rearranged themselves in the same fashion. By moving a card along the spectrum so as to intercept certain rays the fish could be driven up to the blue or violet ends. If the red-yellow and green-blue regions were intercepted the fish spread in both directions towards the ends. If only the red and yellow-red and the blue and violet were exposed many more fish collected at the violet than at the red end though to the human photopic eye the red is much brighter. Just as to the human achromatic scotopic eye, the red end was shortened so that the extreme red produced no more effect on the fish than complete dark- ness. By lighting one half of the basin with homogeneous light and the other with equivalent mixed light which could be varied in intensity it was possible to get the fish to arrange themselves evenly in the two halves. In this manner the luminosity values of different parts of the spectrum were worked out and it was found that the luminosity curve agreed very well with that of the achromatic scotopic or totally colour- blind human eye. Hess was also able to measure the effects of dark adaptation. He found that after 15 — 20 minutes in the dark the sensi- bility of the fish for light was many hundred times as great as im- mediately after transference from light to dark. By using red and blue ^ Canicgio Inst., Washington, ii. 257, l'JU8. THE COMPARATIVE PSYCHOLOGY OF COLOUR VISION 145 lights of suitable intensity he was able to prove that the Purkinje phenomenon is not shown by fishes. Owing to the forward movement of the pigment in the pigment epithelium of fishes when exposed to light they possess a certain amount of physical adaptation. Owing to the absorption by the pigment they are in the photopic condition less sensitive to the violet end of the spectrum. Hess used various coloured imitation baits against different back- grounds, and showed that their visibility to the fish depended upon their luminosity and not upon their colour. If the brightness of the grey background corresponded to the achromatic scotopic brightness of the bait it was not seen. Hess's results have been criticised by Bauer^, v. Frisch^ and Franz^, without, however, seriously shaking his position. Amphioxus. The response of amphioxus by movements to light show that the curve of stimulus values of different homogeneous lights agrees nearly, if not completely, with the luminosity curve for fishes and for the totally colour-blind man, and that the adaptative changes of sensibility to light in amphioxus with its primitive visual organs, nearly resemble those of higher vertebrates. CHAPTER III THE COLOUR VISION OF PRIMITIVE RACES Attention was first drawn to the colour vision of primitive people by Gladstone, in 1858. In his Studies on Homer and the Homeric Age he drew attention to the vagueness of the colour terminology of Homer, and concluded that the ideas of colour at that time must have been different from our own. Ten years later Geiger* came to the conclusion that the colour sense of the ancients must have been defective. Not only in Greek literature, but also in the Indian Vedic hymns, in the Zendavesta, in the Norse Eddas, and in ancient Chinese and Semitic writings there is evidence of the paucity of colour terms, especially in 1 Ce7itralbl. f. Physiol, xxm. 1909 ; Arch.f. d. ges. Physiol, cxxxni. 7, 1910 ; oxxxvii. 1911. 2 Verhandl. d. Deutsch. Zool. GeselUch. 220, 1911. ^ Inter nat. Rev. d. ges. Hydrobiol. u. Hydrogr. 1910. * Contributions to the History of the Development of the Human Race, p. 48, 1880. P. c. V. 10 U6 COLOUR VISION the case of names for green and blue. Geiger advanced the view that the sensation of red had been evolved first, then yellow and green, and finally blue. Magnus^ came to the same conclusions from still more extensive researches, and Gladstone^ returned to the subject in the same year. These views were strongly criticised by Grant Allen^. It was shown that modern poems showed similar peculiarities, red occurring much more often than blue in Swinburne's Poems and Ballads and in Tennyson's Princess. La Fontaine used an epithet for blue once only in all his poems (Javal*). Philological evidence on matters of this kind is notoriously open to doubt, but it is probable that the opposition to the views of Gladstone, Geiger, and Magnus was carried too far. The examination of two parties of Nubians, who were travelling in Germany in 1877, by Virchow and others did much to bring about this result. It was found that they used the same word for blue as for black and other dark colours, yet they sorted coloured papers and wools correctly. The use of pigments by the Egyptians, Assyrians, Greeks and others has also been brought forward in opposition to Geiger's views. Green and blue pigments were used by the Egyptians long before the time of Homer, and green and blue beads are found in the prehistoric Egyptian tombs. Benzaky^ has collected the evidence from ancient monuments and considers that it conclusively disproves any colour defect in the Egyptians and Greeks. As Rivers points out, however, the colour sense of the Egyptians — and they appear to have had two names for green and one for blue — has no direct bearing on that of the Greeks, who may have remained in a state of arrested development. Moreover, even in Egypt human statues with blue hair have been found, and in the Acropolis there may be seen a blue bull, a blue horse, and a man with blue hair and beard, all dating from later than the time of Homer^. Magnus' pointed out that the same defect of terminology for green and blue which characterises ancient writings still exists among many primitive races. ^ Die geschichtlich? E nhuickelung des Farbensinves, Leipzig, 1877. ' Nineteenth Century, ir. 366, 1877. ^ The Colour Sense, London, 1879. * Bull, de la Soc. d' Anthropologie de Paris, xii. 480. 1877. * Du Sens chromatique dans Vantiquite, Paris, 1897. ^ Cf. E. A. Gardner, Handbook of Greek Sculpture, p. 28, 1902. ^ Untersuchungen u. d. Farhensinn d. Naturvolker, Jena, 1880 ; Ueber ethnol. Unter- such. d. Farbensinnes, Breslau, 1883. THE COLOUR VISION OF PRIMITIVE RACES 147 By far the most important contributions on the subject are those of Rivers^. In the Torres Straits expedition Rivers examined the colour vision of two tribes of Papuans, of natives of the island of Kiwai, and of members of several Australian tribes. The languages of these people showed different stages in the evolution of colour terminology, which correspond in a striking manner with the course of evolution derived from ancient writings by Geiger. Some Australian natives from the district of Seven Rivers used only three colour epithets ; one for red, purple, and orange ; another for white, yellow and green ; and a third for black, blue, indigo and violet. In Kiwai there was a very definite name for red, and a less definite for yellow ; greens were called by the same names as white and black or light and dark ; blue, violet, and black had the same name ; other Kiwaians had a name for green which was also applied to blue and violet. The natives of Murray Island and of the Island of Mabuiag showed two further stages in the evolution of colour language. In these four stages the lowest possesses only a definite term for red apart from white and black. In the next stage there are definite terms for red and yellow, and an indefinite term for green. In the third stage there are definite terms for red, yellow, and green, and a term for blue has been borrowed from another language ; while in the highest stage there are terms for both green and blue, but these tend to be confused with one another. The series corresponds with the order of the natives in general intellectual and cultural development. The absence of a definite term for blue is very common. In some the word for blue and black is the same, e.g., Kiwaians, Hovas, Bushmen, and many Australian and Melanesian tribes. In many others the word for blue and green is the same, e.g., in African and South American races and in Jamaica. The absence of a term for brown is also common in Australian and Melanesian languages, as well as in Tamil, Singhalese, Eskimo, Welsh, etc., a brown object being called red, yellow, or dark. Rivers tested 150 natives of Murray Island with Holmgren's wools and failed to find any who confused red and green, but blue and green and blue and violet were constantly confused. Examination with Lovibond's tintometer showed a relatively greater sensibility to red and less to blue than in the European. 1 Po}). Sci. Mo. Lix. 44, 1901 ; Jl. of the. Anthropological Ins. xxxi. 220, I'JOl ; Bcp. of the Cnmhridge Expedition to Torres Straits, n. 1, 1901 ; "Observations on the Senses of the Todas," Brit. Jl. of Psychol, i. 321, 1905; The Todas of the Nilgiri Hills, Cambridge, 190(3. 10—2 148 COLOUR VISION Rivers suggests that the insensitiveness to blue may be due to greater macular pigmentation. The natives were able to recognise blue readily on the peripheral retina. Rivers tested 50 natives of Upper Egypt with Holmgren's wools, Nagel's cards, and Lovibond's tintometer. He found much confusion in their colour names, but general agreement with the characteristics of primitive colour nomenclature in other parts of the world. There was a very definite word, ahmar, in use for red, but it was also applied to colours containing a red element, such as orange, purple, violet and brown. A somewhat less definite term, asfar, was employed for yellow, also for orange and brown and occasionally for green and faint red. The word for green, akhdar, was still less definite, being very often applied to blue, violet, grey and brown. There was no definite word for blue. The word azraq, usually regarded as the Arabic equivalent of blue, was never used by these people for light blue, and was applied by them more frequently to black than to an indigo blue. The nearest approach to a word for blue was Idhdni, milk colour, which was, however, often used also for green, grey, and brown. Azraq and iswid (black) were used indiscriminately for black, blue, and violet, and also for dark brown. The Arabic term for brown, asmar, was never once used for this colour though occasionally applied to blue and grey. Over twenty different names were used for brown papers and wools, but generally ahmar or asfar. Of 80 natives of Upper Egypt tested by Randall-Maclver and Rivers four were certainly colour-blind (" red-green blind "), making the characteristic mistakes with Holmgren's wools (see Part II). Others made the same kind of matches as the Torres Straits' natives, i.e., they behaved normally with the red, pink, and yellow test wools, but compared green with blue, and blue with violet. Two distinct tendencies were also noted, viz., to match according to saturation rather than according to colour tone, and to put together wools which would be called by them the same name. The colour thresholds for red and yellow were low, that for blue much above the European average. Amongst the Todas there was a definite word for red which also meant blood. Orange was often called " blood " or " earth." Yellow was called drsena. probably a borrowed word. Green had many names — " leaf," " moss," including the Tamil name and nil, blue, and Mg, dark or black. Blue and indigo were generally called ml, a word used by all the Dravidian races of Southern India, but sometimes THE COLOUR VISION OF PRIMITIVE RACES 149 hag or hdthiti, black. The nomenclature for violet and brown was very indefinite. There was a definite name for white, and greys were called white or black, or occasionally " ashes," a common term for this colour in Southern India. Rivers tested 503 Todas with Holmgren's wools : 43 were found to be definitely colour-blind. The normal individuals matched the wools well, but most confused red and pink, green and blue, violet and blue, and there was a general tendency to confuse faintly coloured wools. One confused blue with grey and brown, and two blue with brown. The defects of this kind were less marked than among the Uralis and Sholagas examined in India^ and the Egyptians. All showed a tendency to discriminate greens, blues, and violets less definitely than reds and yellows, indicating that the deficiencies in nomenclature for the former group of colours were accompanied by a certain degree of deficiency in their discrimination. Quantitative observations on the colour thresholds were made with Lovibond's tintometer. The table gives the results obtained by Rivers in five different communities. No. ex- amined. Red M.V o- C. v. Yellow M.V. f J \ \ \ / \ ^ / 1 r r \ \ \^ ' \ / / / / \ \^ / \ \ / / / / / i k\ \ \ , / ^ / — 1 / 1 1 ^ L J L Fig .720 700 680 660 640 620 600 580 560 540 520 500 480 460 440 420 400 380 a B C D E b F G H 44. Equal-area gauging curves for dichromats and nionochromats {H). Ifj , deuteranopic " warm " curve ; W2, protanopic " warm " curve ; K, " cold ' curve common to both types of dichromats. Abscissae, wave-lengths of the interference spectrum of sunlight ; ordinates, arbitrary scale. (Konig.) It is to be noted that the curves are so drawn that the areas enclosed between each curve and the axis of abscissae are all equal. They therefore represent the relative magnitudes of the stimulus values of the given elements — red or blue — in the unanalysed light from which the spectrum is derived {v. Part III, Section III, Chap. 11). They do not establish any quantitative relationship between the stimulus values for different individuals, since these are qualitatively of a different order. V. Kries^ used as his gauging lights a red (589-2 /x^u, for protanopes, 645 /x/x. for deuteranopes) and a. blue (460-8 ^a^u.), and determined the relative amount of each which must be mixed in order to match exactly the spectral colours. 1 V. Kries and Nagel, Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. xii. 1, 1896 ; v. Kries op. cit. xm. 241, 1897. DICHROMATIC VISION 165 The method of procedure was as follows, v. Helmholtz' spectro- photometer was used, one half of the field being illuminated with the homogeneous light, the other half with the mixture. In this apparatus the lights are derived from two sources. The mixture is obtained from one source by forming two spectra with the aid of a doubly refracting prism, so that these spectra are polarised at right angles to each other. The intensities of the lights in both halves of the field are altered by varying the widths of the slits transmitting the lights and by means of Nicol prisms. Suppose that a deuteranope is being examined and that the mixing lights are 645 yn/x and 460-8 /x/x. One half of the field is illuminated with a homogeneous red, say 670-8 /x/x. With this light a perfect match, both in colour and brightness, can be made with 645 ju/x merely by altering the intensity and without adding any of the 460-8 /x/x light. The red or warm {W-) value of 670-8 fxfM is expressed in arbitrary units, which depend upon the width of the slit, which transmits this light from the second source of light, when the two exactly match. Let us suppose that it equals 33 under the given conditions. A second wave-length is then chosen, e.g. 656 /XjU, the width of the slit which transmits this light from the first source of light being kept unchanged. We now find that the 645 /x/x slit must be opened more in order to establish a perfect match, say to 48. For each succeeding wave-length this slit must be opened wider until about 600 /x/x is reached. Beyond this point the slit-widths must be reduced. At about 540 /x/x it is found that no perfect match can be made with the 645 /x/x light alone. For example, 536 /x/x appears to the deuteranope less saturated than 645 /t/x. It is necessary to add a small quantity of light of shorter wave-length. On adding the appro- priate amount of 460-8 /x/x light a perfect match, both in colour and brightness, is obtained. This light, 536 /x/x, therefore possesses both a red or IF-value and a blue or cold {K-) value, and these are recorded, viz.. If -value 41, Z-value 6-3. The readings are not quite so simple as would appear from this description. For example, it is found necessary to alter the width of the slit which transmits the homogeneous light, otherwise the second slit has to be opened so wide that the two spectra derived from it are impure. Moreover the relative amounts of 645 /x/x and 460-8 /x/x lights mixed, as well as the brightness of the resultant mixture, are altered by rotation of the Nicol prism, and corrections have to be applied according to the angle of rotation. 166 COLOUR VISION If however the apparatus is kept as constant as possible and the same procedure is adopted in each series of observations concordant results are obtained froro the same individual and comparable results from different individuals, v. Kries and Nagel compare directly the results obtained from deuteranopes, using the mixtures of 645 /x/a and 460-8 jjLjx lights, with those obtained from protanopes, using mixtures 200 190 180 170 160 150 140 130 120 1 10 100 90 80 70 60 50 40 30 20 10 f \ // ' \ 1 ^ / 1 <^ Xy\ •i \ \ I ■. '\ / ■ : \ \ \ '■. 1 1 / / : 1 '\-. \\ / / \ ,7 s \ \ 1 / • • > ( ( 1 / / 1 ■ / » / \ / / / / f / ^ V- \ St., \ > !/ •^ ! i ■\ \ \ '/ 1 i VV \ \ .V V, \ \_ \' .V S ^ '••.. "v. ..^ '^ ■^=: ^iii:^ .0 o 1 3 10 11 2 13-5 15 16-5 18 19-; Ol Ol Ol en en A x» ro K> -< O O ID CD 0) Ol en Ol Ol a> t 21 Fig. 45. Gauging curves for dicliroinats : " warm " curves. O 23 i>- Ol CO 24-7 o Fig. 4G. Gauging curves for dichroraats: "cold" curves, as in Fig. 45. (v. Kries.) Same observers by Konig and his subjects. The latter are indeed chromatic scotopic curves, in which there is some dark adaptation combined with a large residual photopic condition. Another factor, however, is the much greater luminosity value of red compared with that of blue {vide infra). Hence the arbitrary units adopted by v. Kries do not signify equal stimulus values of red and blue for the same unit-number. In their gaslight dispersion spectrum one part of lithium red has about the same luminosity as 8 — 12 parts of blue (480 ^/x) for the deuteranope. In V. Kries' units one part of red is equivalent in luminosity to about 168 COLOUR VISION 20 parts of blue. Hence 6 or 10 parts of blue will have relatively- little stimulus value when mixed with 30 to 35 parts of red, whilst a trace of red will manifest itself by definitely increasing the brightness when mixed with 60 parts of blue. The red curves show distinctly the difference in the two groups of dichromats. No blue is necessary for the match until beyond 530 ju/x. The red protanopic maximum is at 571 jUju, ; the sharp fall towards the red end shows the relatively low stimulus value for the long-wave light. The deuteranopic maximum is at 603 jm/x, and the curve does not fall to so low a value at the red end. Rivers^ has therefore intro- duced the terms scoterythrous and photerythrous for protanopes and deuteranopes respectively. The advantage of these terms is that they are purely descriptive of observed facts and are quite independent of any theory. Unfortunately they have not been generally adopted, and it has been thought best in this book to use v, Kries' terminology. The protanopes are relatively more sensitive to the short-wave light, the deuteranopes to the long-wave. In order that a yellow and a red may look equally bright a protanope will require a much more intense red than a deuteranope. This fact cannot be explained on physical grounds, such as macular pigmentation, since these rays are not appreciably absorbed by the yellow pigment. Further to prove that the fact is true of the two groups v. Kries made thirteen equality of brightness observations on 9 deuteranopes and eleven on 11 protanopes, using the red lithium line (670-5 /z/x) with the yellow sodium (D) line (589-2 /x/x). The protanopes required the following quantities of red : 214, 213, 211, 205, 196, 198, 210, 200, 210, 203, 235 : the deuteranopes required : 36-5, 36-3, 36-3, 36-5, 38-4, 37-3, 37-0, 37-0, 37-8, 37-0, 36-9, 38, 40. The protanopes therefore require on an average five times as much red as the deuteranopes in order that the red may appear as bright as the standard yellow. These observations confirm earlier results obtained by Bonders^. The low stimulus value of red for the protanopes accounts for the so-called " shortening of the red end of the spectrum " in this class of dichromats. The limit of visibility of the red end, however, varies with normal people, and depends upon the intensity of the light, whether the whole spectrum or only the isolated colour is observed, and so on. Too much stress has often been put upon this point. The blue curves show a general agreement, but more individual 1 Schafer's Text Book of Physiologij, 1091, 1900. 2 Arch.f. A licit, u. Physiol. 528, 1884. DICHROMATIC VISION 169 variation, which is to be explained on physical grounds. This is the part of the spectrum where macular pigmentation makes itself most felt. Moreover, here too adaptation is most likely to introduce complexity. It is probable that there is no crucial difference between the blue stimulus values of the two groups. Macular pigmentation also probably accounts for the difference between the two red curves (Fig. 45) beyond 552 /x^u,. The ratios of the red of one observer to that of the other throughout the spectrum agree well with the determination of the absorption of the yellow pig- ment by Sachs. Similarly the ratio of blue to red in two persons gave the same value (0'3) as for the normal (v. Kries). The estimation of macular absorption is much easier in dichromats than in normal trichromats. Macular pigmentation is a serious complication in the accurate determination of the neutral point. This point, where the mono- chromatic colour appears colourless or white to the dichromat, should be situated w^here the W- and /i-curves intersect, and hence it should be situated rather nearer the violet end in protanopes than in deuter- anopes. If, however, there is much macular absorption the light will look yellower and the neutral point will be displaced slightly towards the red end. Indeed, the neutral point of a protanope with densely pigmented macula may be nearer the red end than that of a deuteranope with slightly pigmented macula, v. Kries has made the following estimates, specially directed to elucidate this matter. Wave length of homogeneous match Nature of light Daylight — reflected from magnesium oxide coated sur- face Light reflected from clouds — weakened by ground glass :, „ „ „ weakened by smoked glass These estimates were made with a comparison white light. The existence of a neutral point shows that dichromats receive the same sensory impression from what normal colour-sighted indi- viduals call white, and also from a certain homogeneous spectral light and from the mixture of red and blue which to them matches that homogeneous light. The homogeneous light appears to the normal a highly saturated bluish-green ; the red-blue match appears a red- purple. If we suppress nuances it may therefore be said that one of the chief characteristics of both groups of dichromats is that they confuse red and green. But this is a very imperfect expression of Deuteranope Two protanopes 499 mm 498 fill 490 n/i 499 ,, 497 „ 489 „ 495 „ 494 „ 486 „ 170 COLOUR VISION the facts, for owing to the very low stimulus value of red for the protanope he requires much more red in his red-blue match of a blue- green homogeneous hue than does the deuteranope for the same amount of blue. Hence a protanope confuses a slightly bluish red (in the physical sense) with a dark green, e.g., scarlet with olive green, whereas a deuteranope matches a much bluer red with a green which is of about equal brightness to the normal. " The red which appears to match a given green differs markedly both in colour-tone and intensity in protanopes and deuteranopes " (v. Kries). We have considered the conspicuous difference in dichromatic vision from trichromatic vision. Its conspicuous similarity in an important respect was pointed out by Seebeck in 1837, viz., that all colour mixtures which appear equal to the normal eye, also appear equal to the colour-blind. This statement is too general, and the following is more accurate, all colour equations valid for normal vision are also valid for dichromatic vision, or colour matches which are valid for the trichromatic are also valid for the dichromatic eye. If this statement can be fully substantiated it follow^s that the dichromat possesses no variable which the trichromat lacks, but lacks a variable which the trichromat possesses. In other words, dichromatic vision IS, Q> reduction form (v. Kries) of normal vision, and not a fundamentally different kind of vision. V. Kries^ has made a very extensive series of observations of the equivalence of trichromatic and dichromatic matches, and they strongly support the validity of the law. Especially is this the case for the red end of the spectrum, where macular pigmentation is least disturbing and where the typical differences in the two groups of dichromats are most obvious. Monochromatic yellow is matched by a red (670-8 yLfx) and yellow-green (550 /z/x) mixture. As the ratio of red to green is altered a suitable intensity of yellow can always be found which will match the mixture. The match for the protanope, however, is not valid usually for the deuteranope. For the former a very red mixture matches a dark yellow and a very green mixture a bright yellow. The deuteranope finds the mixture too bright in the first case and too dark in the second. When the mixture is such that it exactly matches the yellow for both dichromats it also matches for trichromats, and that with extraordinary exactitude. In fact when we succeed in finding matches with which both protanopes and 1 Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. xiii. 274, 1897. DICHROMATIC VISION 171 deuteranopes agree trichromats will also agree with them. The follow- ing table (v. Kries) bears out this statement. Ratios of red (670 8 mm) to yellow-green (550 ntj.) in mixtures wliich match the homogeneous light are AVave-length homogeneous of liglit for deuteranopes too dark for protanopea too light for trichro valid 639 ^M 0012 0026 0016 625 „ 0038 0-062 0-044 613 „ 007 0-12 0-09 589 „ 0-22 0-49 0-33 569 „ 1-00 300 1-34 The stimulus values of the red and green can indeed be calculated from the matches for the normal. The latter are given on p. 34 and are repeated in the next table. We see that the normal requires 88-5 units of standard red to match, both in colour and brightness, 670-8 /xjLt, and 71 units of standard green to match 550 /j-fx. A certain deuteranope required 33 and 64 units respectively for the same matches. Now for the normal trichromat, Qk = Qr + Qg where Qk is the quantity of homogeneous light, Q,. the quantity of red (670'8/i/x), and Qg the quantity of green (552 /iyu) to make a perfect match. But in the method used, i.e., with the spectrophotometer, Q^ belongs to one spectrum, and Qr and Qg belong to two spectra derived from another source of light. By making r=\, we obtain the value of Q,- in terms of the Qa spectrum, and similarly for g = \. Now, if observations under identical conditions by different observers are directly comparable, Q,. = 88"5 ; Qg=l\ for the normal trichromat. Q;. = 33 ; Qg=Qi for the deuteranope. Therefore, in the intensity scale of the deuteranope Qa = Q'r + Q'g Where ^'' = 8^5' ^^^^ Q'^-^^- For example, the normal trichromat requires 202 units of red and 67 units of green to match 591 /iyu., or Q,,, = 202Q, + 67Qg. We shall expect the intensity value of 591 [xn for the deuteranope to be 202x33 67x64 ,^, -^^^+-Ti"=i^^- 172 COLOUR VISION V. Kries found the observed value to be 137, and so on, as in the follow- ing table. Wave-length of homo- geneous light Quantities of 670 8 HM and 552 HM in normal match 670-8 fi/j. 552 MM Stimulus value for deuteranope Calculated Observed Stimulus value for protanope Calculated Observed 670-8 88-5 — 33 33 4-9 4-9 628 251 10 106 107 28-8 38-5 615 276 27 126 147 54-2 63 603 270 49 145 151 86 84 591 202 67 135 137 108 105 581 123 76 114 124 117 113 571 73 91 110 103 137 126 561 21 80 76 82 111 106 552 — 71 64 64 101 101 From this table we also see that the stimulus value of the yellow- green (552 fMfju) is to that of the red (670-8 fiiA.) as 101 : 4-9 or about 5 : 1 for the protanope, whereas for the deuteranope it is only as 64 : 33 or about 2:1. The luminosity of the spectrum. The luminosity curves of dichromats were investigated by Hillebrand (1889)^, who found the protanopic maximum at 560 — 658 /x/x, and by Konig and Ritter (1891)^. The following table gives the values for two trichromats and three dichromats at the intensity value H. Trichromats Dichromats Deuteranope Protanopes Wave-length Konig Kottgen Brodhun Ritter X 670 0-855 1-120 0-540 0-0518 0-071 650 2-381 2-137 1-368 0-155 0-183 625 3-460 3-413 2-630 0-493 0-517 605 3-650 3-247 3003 0-996 0-976 690 3-030 2-645 2-539 1-389 1-370 575 2-358 1-923 2-183 1-615 1-477 555 1-695 1-389 1-661 1-412 1-339 635 1- 1- 1- 1- 1- 520 0-554 0-553 0-576 0-606 0-700 505 0-224 0-250 0-225 0-314 0-492 490 0-0994 0-092 0-0846 0-152 0-250 Figs. 47 and 48 show the curves. Ritter's protanopic curve for the lowest intensity A is almost identical with Konig's. Konig thus early demonstrated the identity of the protanopic achromatic scotopic and the normal achromatic 1 Sitz. d. Wiener Akad. xcviii. 3, 70, 1889. - Konig, p. 144. DICHROMATIC VISION 173 scotopic luminosity curves. As the intensity increases the divergence increases. At the highest intensity the protanopic maximum is at about 560 nix as compared with the normal 615 fxfji. Brodhun's deuteranopic maximum is almost identical with the normal. The deuteranopic achromatic scotopic curve is also nearly identical with both the normal and the protanopic, having a maximum at 535 /x/z. Konig emphasises the fact that nothing as to heterochromatic luminosity values can be deduced from the equivalence of colour matches of two individuals, as of course follows directly from the validity of such matches for both normal and dichromatic vision {v. p. 170). Light intensities FG H E D C B A 670 650 625 605 590 575 555 535 520 505 490 B C D B F 470 4 50 4 30 Fisr. 47. Protanopic luminosity curves for different intensities of light, A being the lowest, H the highest. Abscissae, wave-lengths of the prismatic spectrum of gas- light ; ordinates, arbitrary scale. (Konig.) Konig's curves agree well w^ith Abney's^ (Fig. 49). His normal maximum is 585 /xjli, protanopic 559*6 /x/x, and deuteranopic 599-6 juju. Polimanti^ obtained similar results with the flicker method (Figs. 50 and 51). The curves show the general resemblance of the flicker periphery luminosity values. The normal maximum is 589 jtx/x, the protanopic 565 /x/x, and the deuteranopic 606 /x/x. It is noteworthy, that though Polimanti found his flicker and periphery values appreciably different from Nagel's (deuteranope), v. Kries found no appreciable difference between his own and Nagel's periphery values. However this may be explained it is certain that the divergence of the deuteranopic luminosity curve from the normal is less than that of the protanopic. Watson^ has also published a series of observations by the flicker ^ Abney, p. 279; Proc. Roy. Soc. Lowl. A, Lxxxni. 1910. ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xix. 263, 1899. * Proc. Roy. Soc. Land. A, Lxxxvm. 404, 1913. 174 COLOUR VISION method, which agree with those already mentioned. We may more conveniently discuss his results at a later stage (Part III). The spectrum as seen by the dark-adapted eye. Tschermak^ has stated that the process of adaptation varies according to the type of /-~>, _ > 3-8- 3-6- ( \ " Trichromats 3-4- ji f \ \ \ \ ~*'*'^^"""" 3-2- j 1 1 i A Deuteranope\ ") f Dichromats \ Protanopes ) 1 i \ \ 3-- 11 / / \ \ , . . . . Rfinhfnoco t'rflllP9 nffhf* rnrrntt 2-8- / / / / / ■•. \ of the Spectrum 2-6- / / / / 2-4- ;' 1 1 \ ;\A i ' I \\\ V 2-2- /.' /" 1 V •' \ 1 i ^■^■ •. \ f \ 1 (\ ^ ^\ 2- If 1 1 1 / \ A V A 1-8- 1/ i i / v>\ 1-6 / i; / / / ,. \\ 1-4- 1-2- •'if •' 1 / 1 j 1 f •V. ^4 1 A* ■■.'A f /.' "'^-'--'•^ **\ V-- [ ' i i 1 ^ '/^ 0-8- / 1 f / 0-6- ■; j 1 0-4- J X. N. N. 0-2 ^ ..•/ "• 1 f'l 1 1 • 670 650 625 605 590 575 555 B C D 535 520 505 E 490 Fig. 48. Photopic luminosity curves (intensity H) of trichromats and dichromats. Abscissae, wave-lengths of the prismatic spectrum of gaslight; ordinates, arbitrary scale. (Konig.) colour blindness, being normal in deuteranopes and much below normal in protanopes. Nagel and v. Kries, however, are fully convinced that ^ Arch. f. d. ges. Physiol, lxx. 297, 1898 ; Ueber ■physiol. u. j)ath. Anpassung des Auges, Leipzig, 1900 ; Ergeb. d. Physiol, i. 2, 700, 1902. DICHROMATIC VISION 175 adaptation occurs exactly as in the normal in all types of colour blindness. It has already been mentioned that the achromatic scotopic luminosity curve is identical with the normal in all cases of colour blindness. With regard to Purkinje's phenomenon, owing to the relatively far greater ratio of the achromatic scotopic values of the yellow-green region of the spectrum to the red as compared with normal vision all C7 f ^ ^7 1 ■?'/ V 1 . if \ 7 4^ in.(t Ji 7 1 \ / / \ G _Li^ ^ hi 1 1 D. 1 ^- 1 ■i^^ 100 90 eo 70 60 S 50 a 40 30 20 10 10 15 20 25 30 35 40 45 50 Scale of Priam^atir Spectrum. 55 60 65 Fig. 49. Photopic luminosity curves of trichromats and dichromats. (Abney.) the effects are accentuated. This fact was early brought out by Preyer^ and van der Weyde^ and more conclusively by Konig^. Thus K5nig found a change of maximum brightness from 605 /xjn to 535 /x/x for the deuteranope Brodhun and from 575 fx/x to 535 [xfi for the protanope Ritter on diminution of the intensity of the light, associated of course with increasing dark adaptation. Konig and Tonn made a 1 Arch, f.d.ges. Physiol. XXV. 31, IS^^- " Arch. f.Ophth.xxvni. 2, 1,1882. 3 Konig and Brodhun, Sih. d. Akad. d. Wiss. Berlin, 311, 1887 ; Brodhun, Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. v. 323, 1893 ; Tonn, op. cit. vn, 279, 1894. 176 COLOUR VISION number of matches between a mixture of two spectral lights and mono- chromatic light of an intermediate wave-length, and observed the efiect of diminishing the intensity of the light, v. Kries and NageP took greater precautions as to the condition of adaptation and found that a scotopic colourless homogeneous blue-green possesses a six or seven times higher scotopic value than the photopic equivalent red-blue mixture. The greatest difference is shown between homogeneous reds and yellow-greens. The latter, from 544 /x/x violetwards, give with red .^. 100 ^'■^ ^ f^ \ s \ 60 / f/i / \ \ / / / / / '; 1 \ ^ / / -I f / j 1 1 V \ \ \ s 50 AD 1 1 1 i V =^ \ > V 1 1 1 1 / i 1 ; 1 VS. N > ■v V ■■-'\ 30 on ■I / f /' 7 >>. i / / / '-^ s 10 0 y ' ■r 664 1 1 — 642 624 606 ^ 1 — 589 565 1 553 — 1 — 543 526 509 Na-Q. Na-^-bN&-y /Va-0-5 Na. /V«+4 /Va+i A'a+1'5 Ha-^1 Na*3 Fig. 50 Photopic luminosity curve of a protanope (flicker method). Periphery himinosity curve of the same protanope. Photopic luminosity curve of a trichromat (flicker method). Periphery luminosity curve of the same trichromat. (Polimanti.) exactly equivalent photopic values, whereas the achromatic scotopic value of the yellow-green is more than 100 times that of the photopic equivalent red (642 /x/x). These remarks apply to the deuteranope and are illustrated in Fig. 52. Similar though less marked differences occur in protanopes. The differences to which reference has already been made (Part I, Section IV, Chap, i) between the photopic and scotopic periphery values of the trichromat are accentuated in the dichromat. Colours ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xii. 1, 1896. DICHROMATIC VISION 177 which, when viewed by peripheral parts of the retina, accurately match under light adaptation and with high intensities of light, may possess very unequal achromatic scotopic values, and the inequality is much greater with dichromatic than with trichromatic vision. Regional effects. As might be expected the dichromat, like the trichromat, sees colours best with foveal stimulation. For certain conditions, indeed, his powers of discrimination with central vision 120 t1 0 100 90 80 70 60 50 40 30 20 10 A u ■7 1 .-^ H / / / It 'tS; // V ^ ■ I' : T' / > Vy \ 1 // \ ^!>. ^^ 7 1 '^> \ y ill \ S "-«^ /i 'I / ^•N. >>s ""■■■■•. ^ 1 1 *'s. J 687 I— 1 — 664 1 1 1 G42 1 624 606 — ' 589 1 1 1 565 L— + — 563 543 1 • 1 520 50 9. t^a-lb Na.-'2N^.-■\■^Na-^ Na.-0-b Na. ;7a+1 //SH ; 5 A;2+2 A£+a ;Va+4 Fig. 51. Photopic luminosity eurve of a deuteranope (flicker method). Periphery luminosity curve of the same deuteranope. Photopic luminosity curve of the same trichromat as in Fig. 50 (flicker method). Peripliery luminosity curve of the same trichromat. (Polimanti.) are markedly superior to those of the normal. Thus, in the protanope a yellow-green light (545 jit/x) possesses at least 100 times the achromatic scotopic value of its photopic red (670 ju/x or higher) match, whereas in the normal the maximum ratio is 6:1. He is, therefore, in a peculiarly advantageous condition for determining whether the fovea is sensitive to changes of adaptation. Nagel^, a deuteranope, was ^ V. Kries and Nagel, op. cit. xxin. 161, 1900. P. c. V. 12 178 COLOUR VISION unable to discover any difference produced by dark adaptation in colour matches with small central fields. A greenish yellow spot on a red background appeared to match the background both for the light- adapted and dark-adapted eye so long as it was fixated exactly. On — " ,' — 300 / V \ / 1 / \ / \ f \ 1 / 1 / / \ \ \ \ \ ' \ 1 i \ \ \ \ / \ / \ 80 / \ / \ 60 \ 50 / \ V 40 / \ , 30 / \ 20 / / V 10 V 1 1 ^ — __ __ — ' 74.7 0 D 2 C ~ i 5 ( 5 7 8 f ) 1 0 1 1 19 1 q.«= 1 ^ 1 fi.e; 1 R 1 q.p o 1 o n o q ' 05 O 0> CTJ O CT> Cl ocntocDCico-. — _._. oi oi CI o en 05 Ol -t* CO ■fw O) CO CD 4^ (S o < 05 ■ft ■ft O) o CO Fig. 52. Scotopic values of the homogeneous lights. Scotopic values of the photopic equivalent red (642 ^/u) — blue (460-8 mm) mixture. Abscissae, wave-lengths of the prismatic spectrum of gaslight ; ordinates, arbitrary scale, (v. Kries and Nagel.) slight deviation (1° to 2°) of the dark-adapted eye the spot appeared much brighter than the background. The scotanopic foveal area subtends a visual angle of about 1-5° (Nagel), but its exact dimensions are difficult to measure. DICHROMATIC VISION 17& With regard to the periphery of the retina v. Kries^ compared a long series of normal colour matches with those of dichromats. His results have been confirmed by van der Weyde^. He found that whilst the deuteranopic matches agreed well with the normal, the protanopic matches were very different. The difference can be shown by the different brightness of various coloured lights without any comparison light (Levy^, Schenck*). The divergence is most marked for colourless red-blue mixtures. Such a mixture which matches a grey background for the normal also matches for the deuteranope within the somewhat wide range of experimental error for such observations. For the protanope such a mixture looks distinctly blue and much too dark ; his match contains less blue and much more red. 100 90 80 70 60 50 40 30 20 10 0 --''"' ^^ 7 -' ^ Z ^' ^ ^v_ V ^ ^-. - t / -4- -^^it X / -' H^ / /'' ^r^ .680 651 629 608 589 573 55£ 530 513 Fig. 53. Photopic perij)heral luminosity curves for the normal trichromat ( ) and for the protanope ( ). Abscissae, wave-lengths of the prismatic spectrum of gaslight; ordiaates, arbitrary scale, (v. Kries.) Dichromats have a monochromatic peripheral zone as have trichro- mats. The normal and protanopic periphery values for this zone are shown in the curves in Fig. 53. The diminished sensitiveness for light of long wave-length and relatively increased sensitiveness for light of short wave-length in protanopes is clearly shown. The deuteranopic periphery values indicate that the summit of the curve is rather red-wards of the normal. Angier^ claims to have proved more marked differences from the normal in this group of dichromats, but as Frl. v. Maltzew^ has shown considerable individual variations in normal trichromats the question must be left open. ^ Zisch. /. Psychol, u. Physiol, d. Sinnesorg. xv. 266, 1897 2 Onderzoekingen, Utrecht, iv. 3, 2, 1882. 3 Ztsch. f. Psychol, w. Physiol, d. Sinnesorg. xxx^^. 74, 1904. * Arch. f. d. ges. Physiol. cx\Tn. 174, 1907. ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxxvn. 401, 1905. * Ztsch. f. Sinnesphysiol. xun, 76, 1908. 12—2 180 COLOUR VISION Areal Effects. NageP, who was a deuteranope, found that he was only completely dichromatic for small fields. With larger fields, involving the extra-foveal regions, he became partially trichromatic. By fatiguing his eyes with red and orange lights he became completely dichromatic. His perception of red was much better than that of green, possibly due to red not causing strong contrast effect. Good dark adaptation had the same effect as fatiguing with red. He found a similar condition in 30 other dichromats. Stilling^ regards Nagel's observations as erroneous. The Colour Sensations of Proteinoses and Deuteranopes. Abney has given illustrations of the names applied by dichromats to the various parts of the spectrum. Since they have only two true colour sensations the distinctions are based chiefly upon variations in luminosity, and are therefore often contradictory. Most observers think that the two sensations experienced correspond most closely to normal yellow and blue^. Uncomplicated cases of unilateral congenital colour blindness would afford valuable evidence, but most such cases recorded are of doubtful value. One case for example was regarded by Holmgren* as a typical protanope with shortening of the red end of the spectrum, but V. HippeP found that he could see the rubidium line y. Both agree that his colour sensations were yellow {D line) and blue (indium or caesium line). Holmgren recorded a second case and Becker^ a third. Hayes'' has recorded a case of a woman with partial protanopia in one eye and normal vision in the other. So far as I am aware no other congenital cases have been recorded, and pathological cases are value- less for the purpose under consideration. Tritanopia. Cases of tritanopia or so-called blue-blindness are rare and mostly due to disease. Since the defect causes little risk of con- fusion of red and green lights it is not of much practical importance. It is, however, of considerable theoretical importance, and a bibliography of the literature on the subject is appended. Goethe^ referred to the condition as "akyanoblepsia^." Konig^" investigated five pathological 1 Arch.f. Anat. u. Physiol. 543, 1907; Zhch.f. Sinnesphy-siol. xliv. 5, 1909. ^ Ztsch.f. Sinnesphysiol. xliv. 371, 1909. ' Pole, Phil. Trans. Boy. Soc. Land, cxlix. 322, 1859. * Centralbl.f. d. med. Wiss. 898, 913, 1880. 5 Arch.f. Ojihfh. xxv. 2, 205, 1879 ; xxvi. 2, 176, 1880 ; xxvn. 3, 47, 1881. « CentralU. f. Augenhlk. 353, 1888. ' Amer. J. of Psychol, xxn. 369, 1912. « Zur Farbenlehre, 1798. » Konig, p. 4. " p. 396 ; Sitz. d. Akad. d. Wiss. Berlin, 718, 1897. DICHROMATIC VISION 181 cases. The subjects match yellow (566 — 570 jxfx) with unsaturated complementary blue. They confuse blue-green with blue, greenish yellow with grey and rose-purple, yellowish green with bluish violet, orange with reddish purple. They appear to be a reduction form of trichromatic vision. Tritanopes call the red end of the spectrum red, the violet end green or blue. Vintschgau and Hering's case was congenital, but showed changes in the red and green, thus approximating to total colour blindness. There was a neutral band in the yellow (598- -568 /x/x), but the violet end also appeared colourless, though of good luminosity and without marked shortening. Yellow and blue could be matched with grey, but a match could not be made between red and green. The most saturated green was at 532 /xfx, and the site of maximum brightness was between 558 fifj, and 565 /x/x. In Bonders' case the spectrum was shortened at each end and the neutral band in the yellow occupied one-third of the spectrum. Tritanopia appears to be commonest in cases of detachment of the retina. It is simulated in cases of jaundice and sclerosis of the crystal- line lens, these being due to absorption by yellow pigment. The same ambiguity may arise in people with unusually dense macular pig- mentation^. Bibliography of Cases of Tritanopia Stilling. Klin. Monatshl.f. Augenhlk. xm. Supij]. 2, 41, 1875 ; xiv. Suppl. 3, 1, 1875. Cohn. Studien uber angeb. Farhenhlindheit, 139, Breslau, 1879. Donders. Ann. d'ocul. xxxiv. 212, 1880. Holmgren. Med. Zentralbl. xym. ; Centralbl. f. Augenhlk. v. 476, 1881. Hermann. Inaug. diss. Dorpat, 1882. Wundt. Phil. Studien, vm. 173, 1892. Vintschgau. Arch. f. d. ges. Physiol. XL\Tn. 431 ; Lvn. 191, 1894. Hering. Arch. f. d. ges. Physiol. Lvn. 308, 1894, Konig. Sitz. d. Akad. d. Wiss. xxxiv. 718, 1897 (Pathological). Piper. Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxxvin. 155, 1905 ^ Pathological). Levy. Arch. f. Ophth. Lxn. 464, 1906. CoUin and Nagel. Ztsch. f. Sinnesphysiol. xli. 2, 74, 1906 (Pathological). Schenck. Arch. f. d. ges. Physiol, cxviu. 161, 1907. Krienes. Ztschr. f. Augenhlk. xx. 392, 1908. Kollner. Die Stdrungeii des Farbensinnes, Berlin, 1912 (Bibliography), ^ Cf. Abney, p. 343 ; Proc. Roy. Soc. Loud. xlix. 1891 ; Hess, Arch. f. Anat. lxi. 29, 1908. 182 COLOUR VISION CHAPTER III ANOMALOUS TRICHROMATIC VISION Slight variations in the colour matches made by people with apparently normal colour vision are commonly met with, and varia- tions occur in the same individual under as nearly as possible identical conditions at different times of examination. Setting aside physical causes, such as differences in macular pigmentation, pigmentation of the lens, etc., ample reasons for slight variations will be found in successive induction, fatigue, the psychological state, and so on. There yet remain, however, cases which cannot be explained on any such grounds. The majority on exhaustive examination show inter- mediate grades linking them with the normal on the one hand and the dichromats (protanopes and deuteranopes) on the other. Konig called this form of colour vision anomalous trichromatic vision. As Green- wood saysi, " It is not improbable that these abnormal trichromatics are extreme variants oi a frequency system representing the whole range of visual types. The matter can only be settled when the quantitative mixing ratios for a definite match have been determined on a large number of persons taken at random ; we may then find that the results are in accordance with some well-known frequency distribution, the normal equation representing the modal value." We owe the discovery of probably the two largest classes of these anomalous trichromats to Lord Rayleigh^. They approximate to the protanopes and deuteranopes. Lord Rayleigh found that if homo- geneous yellow (Z) line, 589 ju^u,) is matched with a mixture of homo- geneous red {Li line, 670-8 /x/z) and homogeneous green {Th line, 535 /xju.) some persons require much more red, others much more green than the normal. The latter are the more numerous class and may be called partial deuteranopes {deuteranomal, Nagel). Of them, Bonders^ found four cases amongst 60 people examined ; Konig and Dieterici^ 3 amongst 70. Many cases have been thoroughly investigated by Hering^, Lotze^, v. Kries'^, and Abney and Watson. Of the partial ^ Physiology of the Special Senses, p. 137. 2 j^Tature, xxv. 64, 1882 ; Brit. Assoc. Rep. 728, 1890. » Arch f. Anat. u. Physiol. 520, 1884. * Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. iv. 293, 1892. ^ Lotos, VI. 1885. * Dissertation, Freiburg, 1898. ' Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xix. 64, 1899. ANOMALOUS TRICHROMATIC VISION 183 protanopes {profanomal, Nagel) Bonders and Konig do not record any case, V. Kries one only^, whereas NageP, Guttmann^, and Abney and Watson record a considerable number. The two last have exceptional opportunities as examiners for the Board of Trade. Their results, of which only a very small proportion have been published, will be discussed later (Part III). v. Kries and his pupils made a series of investigations on the ratio of green to red in the matches with intermediate homogeneous lights for normal and anomalous trichromats. The quotients of long series of average ratios is given in the following table, I for v. Kries (normal) and Lotze (abnormal), II, for Halben (normal) and Lotze (abnormal), III, for Polimanti (normal) and Zehender (abnormal). jgeneou 3 light Quotients I II Ill Homogeneous light Quotients (Levy) 628 mm 4-51 4-35 4-82 ■ 625 Mu 0019 615 „ 3-74 — — 613 „ 0123 603 „ 3-15 3-7 417 601 „ 0-230 591 „ 314 — • — 589 „ 0-278 581 „ 2-68 30 3-92 579 „ 0-262 571 „ 2-48 — — 569 „ 0-249 561 „ 2-15 2-6 2-98 559 „ 0-176 552 „ 2-12 — 2-09 550 „ 0-080 The table shows conclusively that the anomaly is the result of a physiological condition, not of physical conditions, such as macular pigmentation as suggested by Hering*. If it were due to the latter cause all the quotients of any series would be the same. On the contrary the quotients of partial deuteranopes diminish 23ari passu with the diminution in wave-length of the comparison homogeneous light. The divergence is even more marked for the partial protanope (Levy) ; here the quotients also vary with the wave-length, increasing up to 589 [xfji and then diminishing. If the results are compared with Breuer's^ for macular pigmentation we find that his ratio of green to red for central fixation (1° field) was I'l times that for paracentral fixation (3° field, 3° — 6° paracentral). The difference therefore between normal and anomalous trichromats ^ Levy, Dissertation, Freiburg, 1903. 2 Klin. MonatsU. f. Augenhlk. xlii. 356, 1904. 3 Ztsch.f. Sinnesphysiol. XLn. 24, 250, 1907 ; XLni. 146, 199, 255, 411, 1909 « Lotos, \^. 142, 1885. ^ Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. xm. 464, 1897. 184 COLOUR VISION equally manifests itself with paracentral fixation. There might, however, be more widespread dense macular pigmentation or pigmen- tation of the refractive media. This possible factor can be tested by comparison of the achromatic scotopic values of spectral lights for a normal and an anomalous trichromat with a fixed blue (460 (jlix) comparison light. The ratios of the quantities of blue to match the homogeneous lights for Halben (normal) and Lotze (partial deuteranope) were Wave-lengths . . 591 581 571 561 552 544 536 529 Ratios . . 103 1-49 1-31 110 1-24 (0-99) 1-18 0-90 109 The difference is therefore not due to more extensive pigmentation. As might be expected the luminosity curves of the anomalous trichromats show corresponding differences from the normal (Watson^). It has already been mentioned {v. p. 179) that v. Maltzew^ found individual variations in trichromats, and these formed a complete series between the trichromats and complete deuteranopes by way of the partial deuteranopes. The protanopes and partial protanopes showed a quite different luminosity distribution. In red-green mixtures the red component is much more powerful than the green. She con- cludes that the differences in colour matches between the normal and the anomalous have a different cause from the individual variations of the normal. The latter correspond with the luminosity differences, whereas the former do not. Guttmann^ has made important contributions to the knowledge of anomalous trichromatic vision. He is himself a partial deuteranope. He finds that anomalous trichromats show seven chief differences from normal trichromats. (1) For partial deuteranopes the discernible difference in hue is smaller in the yellow and greater in the green than for the normal. The difference in the green is not marked, but that in the yellow is very appreciable. Guttmann's mean error {v. p. 19) at 589 /xfi was from 12 to 13 /x/x, whereas that for the normal under like conditions was 1 to 2 fx^. (2) The anomalous deuteranopes are more dependent on intensity. 1 Proc. Roy. Soc. Land. A, Lxxxvm. 1913 ; see Part III. 2 Ztsch. f. Sinnesphysiol. xun. 76, 1908. 3 Neurol. Centbl. 491, 1904; Ztsch. f. Sinnesphysiol. xlh. 24, 250, 1907 ; XLm. 146, 199, 255, 1908-9; Munch, med. Wochenschr. 566, 1910 ; Ztsch. f. Sinnesphysiol. l\ti. 271, 1910. ANOMALOUS TRICHROMATIC VISION 185 They can only distinguish colours at the optimum intensity, and their optimum is narrower than the normal and of greater intensity. (3) They are more dependent on luminosity differences, which they are expert in translating into hue, thus showing a specially well developed capacity for distinguishing differences in brightness of coloured objects. (4) They need a considerably larger area of stimulation or visual angle for the accurate perception of hue, a fact confirmed by Nagel^. Diminution in field causes a marked diminution in their capacity for distinguishing differences in hue. (5) They need longer time for the correct perception of colours, a fact of great practical importance in certain occupations. (6) Fatigue for colour, and with it defective discrimination, occurs more rapidly than in the normal, (7) They have a very marked increase in simultaneous contrast effects as compared with the normal. Contrast depends to an excep- tional extent upon the luminosities of the contrasting fields, and the anomalous trichromatic colour thresholds are largely dependent upon it. NageP also drew attention to the acuteness for contrast of the partial deuteranopes. He found that a red and a yellow light, side by side, of equal luminosity to the normal appeared different to them, and they often called the yellow greenish or green, whereas in the absence of the red light the yellow was correctly named. Guttmann found his own periphery values agreed with the normal. Like Donders, he calls anomalous trichromatic vision a " weakness " of colour vision, a term to which Nagel objects. Guttmann found that -the " extreme anomalous dichromats " nearly resemble complete dichromats, and show diminished differential sensibility for hue, rapid fatigue and increased perception of simultaneous contrast effects. He also found the after-images of anomalous dichromats agreed with those of the dichromats, but only with homogeneous lights. With pigments the after-images appeared to resemble those of the normal, and no accurate deductions could be drawn without comparison lights, since the naming of colours is very liable to lead to mistakes. Koffka^, himself a partial protanope, confirmed Guttmann's general conclusions for this group. They show the same greater sensitiveness ^ Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxxix. 1905. 2 Klin. Monatshl. f. Augeiihlk. XLn. 356, 1904. ' Ztsch. f. Sinnesphysiol. XLm. 123, 1908. 186 COLOUR VISION to simultaneous contrast effects, but these are spread over the whole spectrum with the exception of yellow (compare the partial deuter- anopes). The contrast-responding field is equal in luminosity to that of the contrast-exciting field. Contrast extends over a larger area and commences more rapidly than in the normal. It is strongest in the green, and is greatest for small fields except towards the violet end. Colour-fatigue sets in more quickly with them, as with the partial deuteranopes. The two groups are thus easily confused. The normal variations in the Rayleigh equation — matching D light with mixed thallium- and lithium-line lights — are more restricted in one group of partial deuteranopes than in the partial protanopes. It has been found that the partial dichromats vary relatively little in their individual matches for the sodium line. If, however, they match a wave-length materially shorter than 589 /z/x their matches become very indefinite and variable. Considerable quantities of red can be added without upsetting the match. On the other hand, for 536 /x/x for example, the red can be reduced almost to zero without the fields ceasing to be at any rate nearly equal. These are extreme cases of partial deuteranopia, and they therefore show very defective sensibility for colour difl'erences in the neighbourhood of 570 — 535 /x/x. Similarly there are extreme cases of partial protanopia in which pure yellow and especially lights of somewhat greater wave-length appear almost the same as the spectral red. CHAPTER IV MONOCHROMATIC VISION There are certain persons who appear to have no perception of colour ; their vision is monochromatic. The condition is congenital, and is known as total colour blindness or achromatopia. The subjects, apart from the colour defect, usually have very bad central vision (less than -f^), marked photophobia, and nystagmus. Several members of a family (11 groups) may be affected either with the same condition or less commonly with other forms of dyschromatopia. Consan- guinity (five cases) and heredity appear to be factors in the incidence of the disease. Males are afi^ected about twice as often as females. Nettleshipi has published 10 pedigrees containing 34 cases affected : 18 males, 15 females, and 1 of unrecorded sex. 1 Trans. Ophlh. Soc. xxix. cxci, 1900. MONOCHROMATIC VISION 187 Grunert^ has collected all the recorded cases up to the date of his paper, adding four others. Cases have since been recorded by Bj errum^, Ronne^, Hessberg^, Kollner^, Juler^, and Gertz'^. The total of published cases is 84 (Nettleship, 1909). Refractive errors are common, generally moderate myopia ; correction produces little improvement of vision. One case recorded by v. Hippel and Uhthoff^ could see Rontgen rays, but this observation has not been confirmed and was negatived in Ronne's case. The only case with good central vision was Frau Prof. R., examined by Raehlmann^, but this case is peculiar in other respects. Obj ective causes for defective central vision have been found occasionally : — macular changes^", pallor of the discs^^, moderate albinism^^, and corneal nebulae^^. In the majority of cases the ophthalmoscopic appearances are quite normal. In Konig's experiments^* on the influence of intensity of illumination on visual acuity, as the intensity rises the curve shows a well-marked angle where the intensity changes from scotopic to pho- topic. In the wholly colour-blind Konig found that this angle was absent, the scotopic part of the curve being continued in the same direction without interruption until an intensity was reached which caused dazzling and diminished visual acuity. In monochromatic vision moreover the usual diminution in acuity in the parafoveal region is not noticed. Photophobia and nystagmus are almost always present. The former increases with the intensity of the light and is absent with low illuminations. Bright light causes an unpleasant sensation of a luminous cloud before the eyes. Central vision is also slightly improved by 1 Arch.f. Ophth. LVi. 1, 132, 1903. ^ Hospiialstidende, 1145, 1904. 3 Ibid. 1230, 1906 ; Klin. Monatshl. f. Augenhlk. xliv. Beilageheft, 193, 1906. * Klin. Monatshl. f. Augenhlk. xl\ti. 2, 129, 1909. ^ Ztsch. f. Sinnesphysiol. xlhi. 409, 1909. « OpJithal. Rev. 65, 1910. ' Arch.f. Augenhlk. Lxx. 202, 228, 1911. » Bericht. d. Ophth. Gesellschft. Heidelberg, 150, 158, 1898. ^ Wochenschr. f. Therapieu. Hygiene d. Auges, ii. 165, 1899 ; Ztsch. f. Augenhlk. ii. 315, 403, 1900. " Nettleship, St Thomas's Hosp. Rep. x. 37, 1880; Uhthoff, Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xx. 326, 1899; Nagel, Arch. f. Augenhlk. xliv. 153, 1901; Hess, Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxix. 99, 1902 ; and Grunert, loc. cit. " Landolt, Arch. d'Opht. iv. 211, 1893 and Grunert, loc. cit. 12 Uhthoff, op. cit, xxn. 1, 200 and Konig and Dieterici, the same case as Uhthoff's, Sitz. d. Akad. d. Wiss. 805, 1886. " Raehlmann, Arch.f. Ophth. xxn. 29, 1879. 1* Sitz. d. Akad. d. Wiss. Berlin, 559, 1897 ; see also Parsons, Roy. Loud. Ophth. Hosp. Rep. XIX. 2, 283, 1914. 188 COLOUR VISION lowered illumination. Uhthofi found the maximum at 12 metre- candles ; the normal maximum is 30 metre-candles or over. The scotopic visual acuity resembles the normal, and the light sense of the totally colour-blind, whenever it has been tested, has been found good. No objective cause can be found for the photophobia. The unpleasant sensation experienced in a bright light is due rather to the interference with vision than to any painful sensation. Thus Nagel and May's patient^ could look at an Auer burner without discomfort ; she was, however, practically blind for several minutes after doing so. Simi- larly, examination with Thorner's ophthalmoscope and dilated pupil caused no discomfort, but great diminution of vision followed for a quarter of an hour. The nystagmus is not found quite so often as the amblyopia and photophobia. Amongst Grunert's last 23 cases it is not mentioned by Fukala^, nor in two cases recorded by Hess. In Nagel's case it was absent in one eye. In v. Hippel's first case it was present, but had disappeared four years later when examined by Hess and Hering^. Grunert noticed diminution in one case in the course of eleven years. These were all elderly people, whereas most of the recorded cases were children. The nystagmus, sometimes rather slow in the condition of rest, becomes very rapid, with short excursions, on fixation. There is a striking resemblance between the nystagmus and that met with in miners*. Many cases showed divergent strabismus, and when absent it could be elicited by eliminating the tendency to binocular vision, as on shutting the eyes or going to sleep. The defect of foveal vision renders binocular fixation difficult and the eyes readily take up a position of rest, which is usually one of divergence. The field of vision is generally normal, though Grunert found ring scotomata in one of his cases. Of greater theoretical interest is the frequency of an absolute central scotoma. Owing largely to the accompanying nystagmus its presence is very difficult to demonstrate, and certainly undue importance has been attached to it since Mrs Ladd- Franklin (1892) and Konig (1894) hypothecated it on theoretical grounds. Nearly 30 cases have been carefully examined as to this point. Central scotoma was absent in 14 cases, published by V. Hippel (1), Pfluger (1), v. Kries (1), Hess (5), Grunert (2), Juler (3), ^ Ztsch. f. Sinnesphysiol. XLn. 69, 1907. - Klin. Monatsbl. f. Augenlilk. xxxvi. 175, 1898. 2 Arch. /. d. gcs. Physiol. Lxxi. 105, 1898. * Llewellyn, Miners' Nystagmus, London. 1912. MONOCHROMATIC VISION 189 and Gertz (1). Gertz used a special method and was convinced that if there were a central scotoma it was less than 50'. There was a central scotoma in nine cases: Konig (1), Nagel (1), Uhthoff (3), Grunert (3), and Hessberg (1). Bjerrum and Ronne were unable to decide with certainty; in Bjerrum's case the blind spot could not be demonstrated. The positive cases include those with macular changes and albinism, but several had quite normal fundi. Whether all cases have an absolute central scotoma or not, it is certain that foveal vision is very defective even as compared with parafoveal, as is admitted by Hess. 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 y ^7 "^ \ J ' V' *^ \ ^ / t k, ^ V \ ,._ J \i s 1 h \ \ /' K y \ \ y \ ih \ s ■> / 1 S v' , y K "V ^ )- / vj ^ ^ -* — 1 ji— i >- ** — 1 r — 1 L JTrn 0 1 2 3 4 5 6 7 8 9 10111213-5 1516-5 18.19-5 2122 2324-7 262728 29 30 31 CT Cl 01 o> o ^ CJi -^ ^5 — o fn »o o) CJ1 Oi c-i tn O lO CJ CO — - c^ m cji oi m -. _. ro ^ o ■t' *■ ^ ^ 0> o> ^ u o o CO 01 > ~s -fc >^ i Fig. 54. Achromatic scotopic luminosity curve of a deuteranope. Luminosity curve of a monochromat. Abscissae, wave-lengths of the prismatic spectrum of gaslight ; ordinates, arbitrary scale, (v. Kries.) The spectrum appears to the totally colour-blind as a monochromatic strip, which is usually described as grey. No colour distinctions are made, but the luminosity varies in different parts. The part which appears to them brightest is what the normal call green, and when the luminosity curve is worked out it is found to agree in a very remarkable manner with the normal achromatic scotopic luminosity curve. This fact was first discovered by Hering and Hillebrand^ and has received conclusive confirmation, notably in the researches of Konig and ^ Sitz. d. Wiener Akad. xcvin 70, 1889. 190 COLOUR VISION Dieterici^ and Konig and Ritter-, Hering^, v. Kries'*, Nagel and May^, and Abney and Watson. The maximum luminosity with the gaslight spectrum is about 536 /x/x, with sunlight about 527 ju/x. The red end is shortened, but not the violet (Figs. 54 and 55). Fig. 55. — Achromatic scotopic luminosity curve of a deuteranope. Luminosity curve of a monochromat. Abscissae, wave-lengths of the prismatic spectrum of the Nernst light ; ordinates, arbitrary scale. (Nagel and May.) Whereas, however, increasing the intensity of the light shifts the maximum for the normal towards the yellow (see Section III, Chap, ii), it merely increases the luminosity for monochromatic vision. The curve is also independent of adaptation. We have already drawn 1 Sitz. d. Akad. d. Wiss. Berlin, 805, 1886 ; Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. IV. 241, 1892. 2 Konig, p. 144. ^ Arch. f. d. ges. Physiol, xlix. 598, 601, 1891. * Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xiii. 239, 1897. 8 Ztsch. f. Sinntsphysiol. xlii. 69, 1907. MONOCHROMATIC VISION 191 attention to the similarity of the scotopic luminosity curve and the curve of bleaching values for different monochromatic lights for the visual purple (p. 55). T. C. Porter's and Schaternikoff's experiments, especially the latter, (p. 95), have shown that the rate of change of fusion frequency of intermittent stimulation of the retina is lower for achromatic scotopic than for photopic vision. Uhthoff ^ examined one of his cases and found that the totally colour-blind reacts like the normal with achromatic scotopic vision. A rotating disc with black and white sectors ceases to flicker with an intensity of light even above the normal photopic threshold whilst it is still flickering to the normal eye. Nagel has confirmed this observation. The facts of monochromatic vision are inconsistent with the view that it is a simple reduction form of trichromatic vision, though some cases have been regarded as of this nature. The theoretical explanation will be discussed in Part III, Section I. Bibliography of Cases op Monochromatic Vision Huddart. Phil. Trans. Roy. Soc. Lond. Lxvn. 260, 1777. Schopenhauer (1812). Griesebachsche Ausgabe, vi. 81, 1875. Rosier. Obs. sur la 'physique, etc. xm. 87, 1779. d'Hombres-Firmas. C. r. acad. des sci. n. 1849. Rose. Arch. f. Ophth. xn. 2, 98, 1860. Galezowski. Du diagnostic des maladies des yeux par la chromatoscopie retinienne, Paris, 1868. Bonders. Klin. Monatsbl. f. Augenhlk. 470, 1871. Raehlmann. Arch. f. Ophth. xxn. 29, 1879 ; Zisch. f. Augenhlk. n. 315, 403, 1900. Cohn. Studien uber angeb. Farbenblindheit, Breslau, 1879. Becker. Arch. f. Ophth. xxv. 2, 205, 1879. Nettleship. St Thomas's Hasp. Rep. x. 37, 1880 ; Roy. Lond. Hasp. Ophth. Rep. xi. 373, 1887. Magnus. Centralbl. f. Augenhlk: iv. 373, 1881. Landolt. Arch. d'Opht. i. 114, 1881 ; xi. 202, 1891. Scholer and Uhthoff. Bcitrdge z. Path. d. Sehnerven, BerUn, 1884. Dor. Rev. gen. d'Opht. iv. 434, 1885. Konig and Dieterici. Sifz. d. Akad. d. Wiss. Berlin, 805, 1886; Uhthoff, Arch. f. Ophth. xxxn. 1, 200, 1886 : Konig, Arch. f. Psychiat. und Nervenkr. xxi. 284, 1889 ; in Konig, p. 206. Kressig. Mitt. a. d. ophth. Klinik zu Tubingen, n. 332, 1890. Qucrenghi. Ann. di Ottal. 35, 1891 ; Ann. d'ocnl. cvi. 333, 1891. Hering. Arch.f. d. ges. Physiol. XLix. 563, 1891. V. Kries. Zisch. /. Psychol, u. Physiol, d. Sinnesorg. xm. 241, 1897. Colburn. Amer. J. of Ophth. xiv. 237, 1897. 1 V. Kries, Zisch. f. Psychol, u. Physiol, d. Sinnesorg. xxxn. 113, 1903. 192 COLOUR VISION A. V. Hippel. Festschrift, Halle, 1894 ; Bericht. d. Ophth. Gesellschf. Heidelberg, 150, 1898. Hess and Heriiig. Arch. f. d. ges. Physiol, lxxi. 105, 1898. Pfliiger. Bericht. d. Ophth. Gesellschf. Heidelberg, 166, 1898; Internal. Congress, Moscow, 315, 1898. Fukala. Klin. Monaishl. f. Augenhllc. xxxvi. 175, 1898. Uhthoff. Bericht. d. Ophth. Gesellschf. Heidelberg, 158, 1898 ; Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xx. 326, 1899 ; xxvn. 344, 1902. Hess. Arch. f. Ophth. Li. 225, 1900 ; Ztsch, f. Psychol, u. Physiol, d. Sinnesorg. xxix. 99, 1902. Abney. Proc. Roy. Soc. Loud. Lxvi. 179, 1900. Nagel. Arch. f. Ophth. XLiv. 153, 1901. Pergens. Klin. Monatsbl. f. Augenhlk. XL. 2, 46, 1902. Grunert. Arch. f. Ophth. lvi. 1, 132, 1903. Wehrli. In Nagel's Jahresbericht f. Op)hth. xxxiv. 92, 1903. Bjerrum. Hospitalstidende, 1145, 1904. Piper. Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxxvrrr, 155, 1905. Ronne. Kliii. Monatsbl. f. Augenhlk. xliv. Beilageheft, 193, 1906. Nettleship and Holmes Spicer. Trans. Ophth. Soc. xxvrn. 83, 1908. Nettleship. Trans. 02)hth. Soc. xxix. cxci, 1909. Hessberg. Klin. Monatshl. f. Augenhlk. XLvn. 2, 129, 1909. KoUner. Ztsch. f. Sinnesphysiol. xlhi. 409, 1909. Juler. Ophth. Rev. 65, 1910. Gertz. Arch. f. Augenhlk. lxx. 202, 228, 1911. PART III THE CHIEF THEORIES OF COLOUR VISION SECTION I GENERAL REVIEW CHAPTER I INTRODUCTION In the previous parts I have endeavoured to set forth the best- established facts of normal and abnormal colour vision. There are other equally well established facts, but they are so intimately associated with the theories which we have now to consider that it has been thought better to defer consideration of them. We have already, however, accumulated sufficient to warrant an attempt at this stage to classify them. Such as have been mentioned form the basis of the chief theories of colour vision and will enable the reader to obtain a grasp of the significance of these theories. No theory has ever been propounded which suffices to explain all the facts of the department of science to which it refers. If it succeeds in offering a complete solution of all the problems it ceases to be a theory and merits at least as secure a position in the domain of knowledge as the facts themselves. As McDougall^ says, " the most fruitful hypothesis ever developed is perhaps that of an ethereal medium to which an impossible and inconceivable combination of properties is assigned by physicists." Theories must be judged solely according to their efficacy as working hypotheses. In so far as they serve the purpose of sign-posts, pointing out the paths of future research, so far are they of value. " Sterile theories easily relinquish immortality. Fruitful theories hand down their immortal part to their children, while their epheiueral shell falls to pieces " (Hering)^. 1 Brain, xxiv. 611, 1901. 2 L^iog^ n. f. i. 15, I88O. P. c. V. 13 194 COLOUR VISION The task which we have set before us is that of discovering in the realm of vision " what psychical event a and what physical stimulus a are universally connected in the order of nature, and of finding the law by which a undergoes a definite change and becomes /3, when a by a change equally definite (but definable only by a physical standard and not a psychical one) becomes b (Lotze^)." In the accumulation of the facts already set forth we may discern two distinct methods which I will term the synthetic and the analytic. In every case there is a stimulus which excites and a sensatipn of which the individual is conscious. In most cases — one might say in all — both are complex. In the investigation two methods may be adopted. In the first the stimulus is reduced to its simplest form and the resulting sensation observed. The sensation is evoked by the simplest possible physical stimuli : hence the designation synthetic. In the second method the sensation is reduced to its simplest form and the varieties of stimuli which elicit this simple sensation are classified and analysed. This is the analytic method. A priori one is inclined to attach more significance to the results obtained by the synthetic method. It is the method which prevails in the exact sciences, chemistry, physics, and so on. Our sensations are notoriously liable to play us false. They are complex and they are modified by all the other sensations to which the individual is subject at the moment of observation. It is impossible to nullify all the conflicting sensations and it is easy to overlook those potent to mislead. On the other hand the problem itself deals with the sensation and its modifications, and it may therefore be reasonably urged that the analytic method of attack is of the essence of the problem. In the exact sciences the sensations involved are generally simple and well-defined and their modifications either complicate the observations little, or are so disturbing as to indicate their own corrective. They are one cause of " errors of observation," and the manner in which these are minimised in the exact sciences is familiar to all. The two methods are not antagonistic but complementary, and as such should lead to identical results. That they do not is the fault of ignorance and is not due to any innate defects. If we had certain knowledge of all the effective factors in the physical stimuli and of all the efiective factors in the resultant sensations and of all their quantita- tive and qualitative relations, we should arrive at consistent results. But that is a truism. It is the role of physiology to deal with com- plex relations of this nature, and especially the role of physiological ^ Metaphysik; Bosanquet's translation, 295, 1894. INTRODUCTION 195 psychology. A more material analogy may be drawn from metabolism. Here we have more or less accurate information as to the nature and constitution of the food material supplied to an organ or to the body and as to the nature and constitution of the secreted or excreted material. The intermediate steps are full of gaps. In the domain of the special senses physiology is the realm of the intermediate steps, physics and psychology the termini. It is natural that the physicist should adopt the method characteristic of his branch of science, the psychologist that characteristic of his own. But if we survey the advance of psy- chology in recent years we cannot but be struck by the fact that it has been largely due to the adoption of the synthetic method. Experi- mental psychology and the elaboration of " psycho-physical laws " are results of this tendency^. It appears to me, therefore, that more credence is to be attached to the results of the synthetic method. They alone are submissible to approximately accurate quantitative estimates, and it is only by quantitative estimates that the facts can be conclusively correlated. We thus obtain a group of correlations which themselves attain the dignity of precise knowledge. Outside this aristocracy of facts is a vast multitude of undigested facts. Either they have not yet been sub- mitted to accurate quantitative relationship or are as yet incapable of so being. In so far as they fall into line with precise correlations they are of confirmatory value. In so far as they are antagonistic they are impotent to destroy these correlations, but ofier a well-stocked field for further research. Not that they themselves are incapable of correlation, but it is qualitative in nature and therefore less precise. Always remembering their relative smaller value as coins of the realm these qualitative correlations are well worthy of extended study. We have therefore in quantitative relations a criterion which can be applied to the grouping of the facts which we have already brought forward. If we rapidly survey the sections of Part I we shall find that the quantitative relationships are best established for luminosity and colour with the photopic fovea and for luminosity with the achromatic scotopic eye. When we consider peripheral vision, temporal, and areal effects, both for photopic and scotopic vision, the quantitative relations are far less well established. The same grouping applies to Part II. Bearing in mind the relative precision of our knowledge of the various facts we can at once hypothecate certain relationships with ^ Cf. Myers, Experimental Psychology, Chap, i, London, 1909. 13—2 196 COLOUR VISION greater or less accuracy. First, there is the divergence between photopic and achromatic scotopic vision, a divergence so striking as almost necessarily to hypothecate diverse mechanisms. Second, there is the outstanding triplex relationship between colour and colour-less sensa- tions that three colour-components suffice to arouse almost the whole gamut of colour and colour-less sensations. Third, — and here the relative precision of our knowledge is less — there is a certain opposition evinced by certain colour sensations to each other, between black and white, and between the complementary colours. Observations by the analytic method particularly indicate that there are specially differentiated opponent activities between certain particular colour sensations, such as those of red and green, yellow and blue. CHAPTER II HISTORICAL REVIEW OF MODERN THEORIES OF COLOUR VISION In 1866 Max Schultze^, as the result of anatomical researches, human and comparative, came to the conclusion that the rods were the more primitive organ of vision and were concerned with the perception of light, but were incapable of initiating impulses leading to colour percep- tion. The cones he regarded as more highly differentiated and capable of initiating impulses leading to both light and colour perceptions. The only arguments available to him were (1) the parallel diminution of visual acuity and colour sensations in indirect vision and of the cones in the periphery of the retina : (2) the absence of cones in certain animals of nocturnal habits (bat, hedge-hog, mole, eel, etc. ; relative paucity in the owl) and the paucity of rods in certain animals of diurnal habits (lizard, snake) {v. p. 10). The discoveries of the visual purple by H. Mliller (1851) and that it was bleached by the action of light by Boll (1876) were followed by the exhaustive researches of Kiihne (1878). They led Klihne to the con- clusion that we could not only perceive the spectrum by means of the rods and the visual purple, without the intervention of the cones, but that it would under these circumstances appear grey and colourless as it does to the totally colour-blind. He hypothecated the existence of other visual substances to account for the colour phenomena of vision 1 Arch f. milcr. Anat. n. 1866 ; iii. 371, 1867. HISTORICAL REVIEW OF MODERN THEORIES 197 and thought that one such was the so-called visual yellow. Haab^ elaborated similar suggestions on the basis of Schultze's theory and Hering's theory. Charpentier (1877 sqq.) distinguished " perception lumineuse brute " or " vision diffuse " from " vision nette," the former with vague localisa- tion, the latter with sharply defined. The " photaesthesic elements " initiate impulses which mutually modify each other as in the inter- ference phenomena of light and thus the two apparatus, each capable of initiating only colourless perceptions alone, become capable of giving rise to colour-perceptions. He did not attempt to define the exact locality of the apparatus ; it is supposed to be partly peripheral or retinal, partly central or in the nervous system. He conjectured that diffuse vision is initiated by the rods and visual purple, but only in achromatic scotopic vision, since the visual purple can itself be viewed entoptically^. Parinaud^, in ignorance of Schultze's theory, elaborated much the same views with more extensive material. He attributed the disease known as night blindness -(often badly termed hemeralopia) to defects in the rods and visual purple'*, and first demonstrated that central vision is intact in this complaint. He asserted the failure of adapta- tion at the fovea in 1884, having previously in 1881 pointed out that the photochromatic interval is absent here. These facts, confirmed by the night blindness of fowls and pigeons owing to absence of visual purple, led him to the conclusion that the cones, which are alone present in the fovea, are incapable of dark-adaptation, whereas achromatic scotopic vision is a purely extrafoveal phenomenon and is the function of the rods and visual purple. This is the " theory of two retinas." Colour-perception according to Parinaud is a purely cerebral function. Similar views were expressed by Liesegang^ and by Parinaud's pupil Weiss^. Berry'^ suggested that the pigment epithelium is the seat of colourless light sensations. Konig^ also regarded the visual purple as the excitant of the rods (KUhne) and with low intensities of light stimuli the basis of achromatic scotopic vision. He considered the visual yellow, the first product of ^ Correspondenzbl. f. Schweizer Aerzte, ix. 1879. ^ C. r. soc. de biol. 1890 ; C. r. acad. des sci. 1891. 3 C. r. acad. des sci. 1881, 1884, 1885; Ann. d'ocul. lxxxv. 113, 1881; cxii. 228, 1894 ; La Vision, Paris, 1898. * Arch. gin. de med. 1881. s photogr. Arch. 1891. 6 Rev. gin. des sci. 1895. ' Ophth. Rev. 1890. 8 Sitz. d. Akad. d. Wiss. Berlin, 1894. 198 COLOUR VISION bleaching of visual purple with higher intensities, as the basis of the blue sensation, which is therefore also carried out by the rods and is absent from the fovea. V. Kries^, also independently of any knowledge of Schultze's paper and apparently of Parinaud's writings, elaborated a similar theory, the duplicity theory (DupUzitdtstheorie), which has received wide accepta- tion. It will be considered in Section II. The physical basis of colour vision, as exemplified in the wonderful researches of Sir Isaac Newton, opened up the ground for the modern theories. Newton himself compared the simple colour notes of the spectral scale to the tones of the musical scale. This idea was elaborated by Hartley, Young, Drobisch, and others, but was soon found wanting^. Attempts to explain the facts of colour-mixtures directly from the undulatory theory of light by Challis, Grailich, and later Charpentier have also proved inadequate^. More recently Hartridge^ has attempted to give a purely physical explanation of the sensation of yellow. Yellow (650 [xfu. — 560 [jl[jl) is accurately matched by a suitable mixture of red (740 /x/x — 690 /^/a) and green (550 ju./x— 510 yu/x) if a small quantity of white light is added to the yellow. By drawing on mathematical paper two sine curves of equal amphtude, which have periods corresponding approximately to the mean wave-lengths of red and green components of this compound yellow, it will be found that their summation gives a curve which approximates closely to a sine curve of wave-length 620 /x/x. It shows beats at every 2000 ^/x ; at the nodes the amplitude is almost zero whereas at the antinodes it is nearly double that of either of the curves of which it is compoimded. By compounding a red and green wave motion one has obtained a yellow wave motion, the mean wave-length of which approximates closely with that found by experiment. If this hypothesis were accurate the combination of the same red and green, polarised at right angles to each other, should fail to arouse the sensation of yellow ; but such is not the case. Further, although the compound curve obtained by combining the sine curves corresponding to the red and green has between any two consecutive minima a period equal to that of the yellow light, yet at each minimum the curve has ^ Especiallj' in Bericht d. naturf. Ges. zu Freiburg i. B. ix. 2, 1894 ; Arch. f. Ophth. XLn. 3, 95, 1896 ; Ztsch.f. Psychol, u. Physiol, d. Sinnesorg. ix. 81, 1896. 2 V. Helmholtz, 3rd ed. n. 97. ' Ibid. n. 129 ; Charpentier, La Lmniere et les Couleurs, Chap, xii, Paris, 1888. * Proc. of the Physiol. Soc, J. of Physiol. XLV. p. xxix, 1913. HISTORICAL REVIEW OF MODERN THEORIES 199 an abrupt change of phase of half a period (Fig. 56). Hence if a disturbance of this character were to act on a system of which the natural period corresponds to that of the yellow light there would be little response, because the disturbance produced during the interval between two consecutive minima would be neutralised by the disturbance produced during the next interval. An illustration will make the point clear. If a succession of correctly timed small impulses be given to a heavy pendulum the pendulum will in time be set in violent motion. If, however, after giving, e.g. four of these timed impulses the timing is suddenly altered by half a period the new impulses will tend to check the motion already communicated, and after four impulses the pendulum will be brought to rest^. Fig. 58. Combination of two sine curves. In 1801, in the Bakerian Lecture before the Royal Society, Thomas Young2 propounded the theory which now commonly bears the name of the Young-Helmholtz Theory. His third hypothesis of " the theory of light and colours " was that the sensations of difierent colours depend upon the different frequencies of the vibrations of the light which falls upon the retina. Since it is well nigh inconceivable that each sensitive point of the retina possesses innumerable parts capable of responding to the individual vibrations, he concluded from Newton's researches that three parts would suffice to account for all the colour sensations, including white. He chose three chief colours, red, yellow, and blue, the wave-lengths of which were in the ratios 8:7:6. Each afiected a corresponding part of the sensitive mechanism. In 1802, after Wollas- ton's description of the spectrum, he chose red, green, and violet, which he confirmed on experimental grounds in 1807. He considered that the simplest conception of the fundamental " parts " was three kinds of nerve fibre. Excitation of the first aroused the sensation of red, of the second green, and of the third violet. The first type of fibre was most excited by lights of long wave-length, the third by those of short wave-length, and the second by those of intermediate wave-length, but 1 I am indebted to Prof. W. Watson, F.R.S., for this explanation. 2 Sec Roxj. Lond. Ophth. Hosp. Rep. xix. i. 1, 1913. 200 COLOUR VISION light of all kinds excited all three kinds of fibres, though in varying degrees. Young's hypothesis received scant acknowledgment, largely owing to his resuscitation of the undulatory theory of light, which was regarded as a grave heresy by the majority of scientists, who firmly adhered to Newton's corpuscular theory. It remained unheeded until the middle of the nineteenth century, when v. Helmholtz^ and Clerk-Maxwell^ adopted it as the explanation of their experiments on the mixture of colours. Grassmann^ simultaneously enunciated the laws of colour mixture which bear his name. An essential feature of the Young-Helmholtz theory is the additive process, especially as applied to colourless light sensations. Addition of physical stimuli entails additive physiological reactions resulting in compounded psychical end results. Goethe^ approached the subject from the psychologist's point of view and quickly accumulated facts which were difficult to explain on the Newtonian basis. Such are particularly the phenomena of after-images and simultaneous contrast. The addition theory seemed incapable of supplying a satisfactory explanation^. The claims of four fundamental colours as opposed to three were emphasised by Aubert^, and this psychological point of view became crystallised into the theory of opponent colours which bears the name of Hering'. The clash of the contending theories led to modifications of each. The psychophysical difficulties led Bonders^, Ad. Fick^ and v. Kries to limit what the latter calls the components theory (the Young-Helmholtz theory) to the peripheral link in the visual path. v. Kries has elaborated this view in the theory of zones. Hering, regarding the question from the psychological point of view, and endeavouring to bring the facts into a universal common relation- ship with other physiological processes, emphasised the complementary and opponent characters of certain fundamental colour sensations. From the psychological point of view black, white, red, green, yellow, and blue are fundamental sensations (Mach^°). All other hues appear to 1 Miiller's Arch. f. Anat. 1852. * Trans. Boy. Soc. Edin. xxi. 1855. ' Ann. d. Physik Lxxxix. 1853. * Zur Farbenlehre, 1810. * See, however, McDougall, iyifra. ' Physiologie dcr Netzhaut, Breslau, 1865; Graefe-Saemisch, Handb. d. ges. Augenheil- kunde, 1st ed. n. 1876. ' Sitz. d. Wiener Akad, 1872-1874 ; Zur Lehre vom Lichtsinne, 1876. 8 Arch. f. Ophth. xxvii. 2, 155, 1881 ; xxx. 1, 15, 1884. » Arch. f. d. gei. Physiol, xvn. 152, 1878 ; XLin. 441, 1888. " Sitz. d. Wiener Akad. loi. 2, 320, 1865. HISTORICAL REVIEW OF MODERN THEORIES 201 our consciousness as mixtures of these sensations. They may be divided into two groups, the tone-free or colourless (black and white) and the toned or coloured. The tone-free are simple sensations, neither contains any element of the other. The toned are also, inter se, simple sensations in the same meaning of the term. Both groups may be divided into three opponent pairs, black and white, red and green, yellow and blue. So much for the fundamental psychological conceptions. From the physiological point of view these opponent sensations can be hypothetically brought into line with other physiological processes. We can imagine three substances of unstable chemical constitution : when white is seen the black-white substance undergoes dissimilation {D) or katabolic change, when black the same substance undergoes assimilation (^4) or anabolic change. The sensation is dependent upon the ratio oi D : A, not upon the absolute intensity of either. Similarly red is seen when the red-green substance undergoes the Z)-process, green when it undergoes the ^-process ; and yellow when the yellow- blue substance undergoes the Z)-process, blue when it undergoes the ^-process. Several difficulties are at once apparent. First, while we readily accept the proposition that katabolic changes arouse sensations it will not be so readily admitted that anabolic changes can produce opponent sensations, or indeed any sensations at all. Though this may affect the physiological analogies of the theory it is not of fundamental importance to the theory as a theory of colour vision. Opponent chemical processes on the analogy of oxidation and reduction, or electrical on the analogy of positive and negative electricity, and so on, would serve the purpose equally well and avoid the difficulty. Second, — what must, however, be conceded as a great difficulty of all theories of colour vision, — black and white occupy a peculiar relation- ship as compared with the others. Black is on this theory the result of a physiological process excited by an internal stimulus and not the result of mere absence of stimulation. Moreover, whereas we can pass continuously from black to white, or vice versa, through innumerable gradations of grey, there is no such uniform passage from red to green or from yellow to blue. We can indeed pass by innumerable gradations from red to yellow, yellow to green, green to blue, and blue to red, respectively, but there are no intermediaries between red and green, and yellow and blue, corresponding to grey between black and white. Further, the fundamental toned colours differ in brightness and darkness, and this may be regarded as an inherent black-white element in the 202 COLOUR VISION colours. Hering has therefore found it necessary to hypothecate a white valency for the toned colours, thereby introducing a complication which mars the symmetry of his design. The peculiar position of black and white in the gamut of colour sensations early made itself felt. Woinow^ distinguished between the white which is a mere sensation of light and is due to stimulation of the rods, the binary red-green and yellow-blue whites, and the quaternary white which is a combination of the two binaries. Preyer^ also dis- tinguished between the whites due to stimulation of the chromatogenous elements, the cones, and that due to the photogenous, the rods. Konig^ distinguished between the simple white of rod stimulation and the trichromatic white of the cones, v. Kries expressly adopts this view and thinks that " it would perhaps be more correct not to say that the rods initiate colourless sensations, but that they initiate a sensation- effect which is variable in only one sense."'* There is certainly a peculiarity, which may be described as a bluish greyness, about the scotopic colourless sensation which distinguishes it from photopic " white." Evidence derived from the colour-blind led Young in his sole reference to the subject^ to ascribe Dalton's defect to absence or paralysis of the fibres which were supposed to initiate the red sensation. Konig^ adopted the explanation brought forward by Ad. Fick and Leber'^ that so-called partial colour-blindness (dichromatism) is due to coincidence of two of the elementary-sensation curves, v. Kries^ adopted v. Helm- holtz' reduction theory or congenital absence of one of the fundamental components. Both Konig and v. Kries accepted Liesegang's view^ that total colour-blindness (monochromatism) is due to rod- or visual purple-vision. Hess and Hering^° adduce against this view the complete analogy in the normal and totally colour-blind dark-adapted eyes of the low central and increasing peripheral threshold values. Later years have brought forth a crop of theories, mostly modifica- tions of those already mentioned. The more important will be briefly mentioned in the following pages. 1 Arch.f. Ophth. xxi. 1, 223, 1875- ^ Arch. f. d. ges. Physiol, xxv. 31, 1881. 3 Sitz. d. Akad. d. Wiss. Berlin, 1894. * Ztsch. /. Psychol, u. Physiol, d. Sinnesorg. ix. 87 note, 1896. ^ Lectures on Natural Philosophy, ii. 315, 1807. * Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. iv. 241, 1892. ' Arch. f. Ophth. xv. 3, 26, 1869. 8 Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xiii. 241, 473, 1897. » Photogr. Arch. 1891. " Arch. f. d. ges. Physiol. Lxxi. 105, 1898. SECTION II THE DUPLICITY THEORY The duplicity theory (v. Kries) states that achromatic scotopic vision is carried out through the mediation of the rods alone, the cones being the organ of photopic vision. It is not definitely stated that rod activity is in abeyance in photopic vision, but there is some evidence in favour of this view and in any case the influence of rod vision in the light-adapted eye may be regarded as slight. It has already been necessary for the sake of clearness to point out the great probability of two separate mechanisms being involved in these two very different types of vision {v. p. 57). Broadly speaking vision with the dark-adapted eye, i.e., scotopic vision, is monochromatic or tone-free. Vision with the light-adapted eye, i.e., photopic vision, is polychromatic or toned. In the former the threshold stimulus intensity is low ; in the latter relatively high. We have hitherto stated the facts and attempted to regard them impartially. It will now be instructive to review them briefly in the light of the duplicity theory. So far as anatomical considerations are concerned the distribution of rods and cones in the retina is of prime importance. Rods are absent from the fovea, an area subtending a visual angle of 50' — 70', and also over a surrounding zone, the whole subtending a maximum of 3°. At the same time it should be pointed out that the structural peculi- arities of the rods and cones are not precisely differentiated. The macular cones are elongated and more rod-like than those found in peripheral parts of the retina^. This merging of structural peculiarities is even more marked in the retinae of lower animals^, and while on the whole later researches have confirmed the observations of Max Schultze {v. p. 196), precise anatomical diversity of the two types of neuro- epithelium has been somewhat discounted. Moreover, if we regard the 1 Cf. Greeff, in Graefe-Saemisch, 2nd ed. Bd. i. Abt. 2, Kap. V. - Cf. Putter, in Graefe-Saemisch, Bd. n. Abt. 1, Kap. x ; Franz, in Oppel's Lehrbuch d. vergleich. mikr. Anat. d. Wirbeltiere, vn. Jena, 1913. 204 COLOUR VISION rods as the more primitive type of visual neuroepithelium, as we are probably justified in doing, the persistence of recognisable rod attributes in the cones, even if modified, differentiated, and rendered more complex, might well be expected. Apart therefore from the difficulties of isolating the physiological results of excitation of the rods from those of excita- tion of the cones it may be anticipated that the latter cells will retain some measure of the functions which are in the highest degree charac- teristic of their prototypes. Hence, if it should ever be conclusively proved that the rod-like foveal cones of the human eye possess some trace of visual purple and are endowed with some slight degree of light- adaptation it would not be surprising ; neither, on the other hand, would it militate seriously against the view that the rods and cones have become essentially diverse in function. Beyond the rod-free central area the cones diminish continuously in number and the rods correspondingly augment in passing towards the periphery in all directions. The chief characteristic of central vision in the photopic condition is its great acuity as compared with peripheral vision. This acuity is most marked in the form sense, and the rapid diminution in " visual acuity " in passing from the centre towards the periphery is strong evidence in favour of the view that the cones are the essential retinal end-organs concerned in the discrimination of form. A striking feature of the vision of lower animals not endowed with a specially differentiated fovea is the remarkable acuteness in the perception of the movement of objects. This feature is also prominent in human peripheral vision, though it is certainly equally, and perhaps more highly, developed in foveal vision. Thus Ruppert^ found that while visual acuity and ability to perceive movements both diminish in passing from the centre towards the periphery the former diminishes more rapidly than the latter. On teleological grounds the perception of movement, a function specially associated with the light sense as opposed to the form sense, must be regarded as primitive ; and we have here, if such be needed, an example of the persistence of this primitive attribute in the more highly differentiated cones. When the field of vision of the photopic eye is further investigated it is found that the perception of colours requires an increasingly intense stimulus in passing from the point of fixation towards the periphery. The complicated details of peripheral colour vision have already been ^ Ztsch. f. Sinnesphysiol. xlh. 409, 1908 ; cf. Easier, Munch, med. Woch. p. 1904, 1906. THE DUPLICITY THEORY 205 discussed, and though in photopic vision they afford little incisive evidence in favour of the duplicity theory they cannot be regarded as seriously discounting it. It is certain that the almost, if not complete, colour blindness of the extreme periphery cannot be attributed to mere paucity of cones. The great difference in luminosity values with different lights, in spite of the tone-free perceptions to which they give rise, is against such a view. We must therefore conclude that the scanty cones of the periphery are incapable of arousing colour perceptions, or that there is some block in the conduction of such impulses from this region. The duplicity theory depends therefore for its support chiefly upon the facts of achromatic scotopic in relation to those of photopic vision, and in this relationship the evidence is abundant and confir- matory. On this theory the rods are the end-organ of achromatic scotopic, the cones of photopic vision. In other words, the rods possess a very high degree of adaptation, the cones little or none. The photopic apparatus is isolated in the fovea, but the scotopic apparatus is nowhere completely isolated. The threshold values for the cones, however, are considerably higher than for the rods, and hence in complete dark adaptation the rods become almost, if not wholly, isolated physiologically. In dark adaptation, with low intensities of light, vision is carried out through the rods alone (achromatic scotopia). As the intensities are increased the liminal stimulation values for the cones are exceeded and vision is carried out by both rods and cones (chromatic scotopia). At still higher intensities vision is carried out chiefly or wholly by the cones (photopia), but whether the rod effects are added to the cone effects or are abolished is as yet uncertain. If the rods are the organ of achromatic scotopic vision, the visual purple, which so far as has yet been proved is present in them only, attains a new significance. The early researches of Kiihne and Sewall^ sufficed to show the slight effect of rays of long wave-length upon this substance, and they found that the maximum efl'ect is in the green, not in the part of the spectrum which is brightest to the photopic eye. Kottgen and Abelsdorff^ in more recent researches showed that the curve of absorption values of the visual purple of mammals, birds and amphibia with the interference spectrum of the Auer light has its maximum at about 500 /x^u, of that of fishes at about 540 /x/x. These results do not agree fully with Kiihne's, and further researches on human 1 Untersuchungen, Heidelberg, ni. 221, 1880. 2 Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xii. 161, 1896. 206 COLOUR VISION visual purple are desirable. Judging by the position of the maximum it should approximate more nearly to the type found in fishes. It should be borne in mind that visual purple has not been isolated in a pure state, free from hasmoglobin. Trendelenburg's curve for bleaching values of the frog's visual purple agrees more closely with the achromatic scotopic luminosity curve of the human eye (Fig. 1). There can be no doubt that the visual purple is of fundamental importance in scotopic vision and that its accumulation is the basis of dark adaptation. It follows that the relation of the achromatic scotopic luminosity values of two lights, e.g. a green which is strongly absorbed and an orange which is weakly absorbed by the visual purple, should depend upon the concentration of visual purple in the rods, i.e., it should vary with the degree of dark adaptation. Stegmann^ has shown that such is the case. If a luminosity match is made between green and orange with intensities at which they appear colourless after 5-15 minutes' dark adaptation, and the lights are again compared after much more prolonged adaptation the orange appears much brighter and must be diminished to about three-quarters of its former intensity to restore the match. It is to be noted that the change is in the opposite direction to that of the Purkinje phenomenon. Sachs^ showed that the pupil-reactions vary with the luminosities of the lights, independently of their colour, both in the normal and colour-blind. Abelsdorff^ confirmed these observations and showed further that maximum pupil-constriction occurs in nocturnal animals from green light, whereas in animals of diurnal habits the maximum is obtained from yellow light. Similar, but more reliable and much more striking results, have been obtained from the electrical reactions, especially in day- and night-birds by Himstedt and NageH and Piper ^. Vision of lights of low intensities with the dark-adapted eye is char- acterised by (1) the absence of colour sensations, (2) a greatly increased sensitiveness for lights of low intensity, (3) a relatively increased 1 Dissertation, Freiburg, 190]. 2 Arch. f. d. ges. Physiol. Lii. 79, 1892 ; Arch. f. Oj)hth. xxxix. 3, 108, 1893 ; Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxii. 388, 1900. ' Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxii. 81, 451, 1900 ; Arch. f. Augenhlk. XLi. 155, 1900 ; Arch. f. Anat. u. Physiol. 541, 1900. * Ber. d. naturf. Ges. Freiburg, ii. 1901 ; Ann. d. Physik, iv. 1901 ; in v. Helmholtz, 3rd ed. n. p. 328. * Arch. /. Anat. u. Physiol. 543, 1904. THE DUPLICITY THEORY 207 sensitiveness for rays of short wave-length as compared with those of long wave-length, red showing no appreciable increase. The third characteristic is the cause of the Purkinje phenomenon. The characteristics of rod vision will therefore be (1) total colour blindness, (2) maximum sensitiveness in the situation of the green of the photopic spectrum, and shortening of the red end, (3) a very high degree of adaptation. The theory explains satisfactorily the apparent deviations from Newton's law of colour-mixtures described by the earlier investigators. Of these, Hering^ and v. Kries and Brauneck^ held that colour equations were independent of the intensity of the light. Konig and his pupils^ disagreed with this conclusion, and v. Kries found it necessary later to modify his opinion^. The photopic values of the matches are essentially cone-values. At lower intensities rod- and cone-values are mixed, and at the lowest only rod- values persist. The results of Konig and his pupils are partly vitiated by lack of attention to adaptation, but more to wandering fixation and the use of objects whose retinal images surpass the rod-free limits. The striking examples brought forward by Ebbinghaus and Mrs Ladd-Franklin {v. p. 60) are easily explained by the duplicity theory, and indeed strongly support it. v. Kries^ extended these observations and explained the anomalous results of other observers. He showed that it was theoretically possible to obtain a photopic " white " equivalent to a scotopic " white," i.e., a " white " which would lose only in subjective intensity on diminution of the physical intensity. The difference in rod-values between a homogeneous scotopic white and a mixture- white are so slight in normal colour-vision that the results are ambiguous ; they are much more conclusive in dichromats. The theory also accounts for the " wandering of the neutral point " in dichromats on altering the physical intensity of the light. In most of the observations referred to above macular pigmentation is a disturb- ing influence (cf. Hering). In general form it may be stated that matches valid for high intensities become invalid for low intensities and dark-adaptation in the sense that the mixture which possesses the greater rod-value exhibits the greater luminosity in tone-free scotopic vision (v. Kries). The regional peculiarities of scotopic as compared with photopic vision afford very strong evidence in favour of the 1 Lotos, vu. 1886 ; Arch. f. d. ges. Physiol, liv. 277, 1893. 2 Arch.f. Physiol. 79, 1885. " Brodhun, Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. v. 323, 1893 ; Tonn, op. cit. vn. 279, 1894. « V. Kries, op cit. ix. 81, 1896, etc 6 /jj^^ j^. gl, 1896 208 COLOUR VISION duplicity theory. The " night blindness " of the fovea, with its almost if not quite complete absence of dark adaptation, the absence of Purkinje's phenomenon on direct fixation with sufficiently small areas of stimulation, the almost if not quite complete absence of a photo- chromatic interval at the fovea, and so on have already been sufficiently discussed, but should be referred to again in the light of the theory. The difference between the peripheral and the achromatic scotopic luminosity curves, the former having the same character as the foveal luminosity curve, shows, as has already been mentioned, that the duplicity theory cannot be held to account for the peripheral total colour blindness of the photopic eye. Temporal effects, such as those of recurrent vision, are so compli- cated that the support they give to the theory is equivocal. MacDougall, as has been seen, brought forward evidence to show that the " ghost " is not necessarily, as was thought, a purely scotopic image, though it is so in its typical form. The change in the fusion frequency in passing from lights of low to lights of high intensity, as shown particularly in T. C. Porter's experiments {v. p. 95), and confirmed in those of Ives {v. p. 96), Dow, Schaternikof!^ and v. Kries^, demonstrate the change over from the scotopic to the photopic mechanism. Particularly noteworthy is the sudden change in value of the constant in Porter's logarithmic equation, which is exactly parallel to the change in the constant for Konig's logarithmic equation for visual acuity^ and is attributed to the same cause. The agreement in the intensity, about 0"25 metre-candle, at which the sudden change in the curve takes place, is striking. There is also some evidence to show that the latent period of excitation of the scotopic apparatus is appreciably longer than that of the photopic apparatus. The experiments of Piper, Loeser, and Henius and Fujita {v. p. 123) show that in the areal effects of the periphery as compared with the fovea at low intensities summation of stimuli plays a much greater role in the scotopic than in the photopic apparatus. Certain cases of abnormal vision afford valuable evidence in support of the duplicity theory. Of these, total colour blindness or mono- chromatic vision may be explained as due to functional abeyance of the cones, and night blindness to functional abeyance of the rods. 1 Ztsch. f. Psychol, u. Physiol, d. Sinnesorg. xxix. 242, 1902, 2 Ibid. XXXII. 113, 1903. 3 Koiiig, p. 378 ; see Parsons, Roy. Lond. Ophth. Hosp. Rep. xix. 2, 283, 1914 ; Cf. Bloom and Garten, Arch. f. d. ges. Physiol. Lxxn. 372, 1898. THE DUPLICITY THEORY 209 Total colour blindness, as we have seen, is characterised by luminosity values which correspond with those of the normal achromatic scotopic eye. Comparison of the curves (Figs. 54, 55) shows their almost complete identity. Whereas, however, the normal achromatic scotopic curve is valid only for low intensities of objective lights the achromatopic curve is valid for all intensities, though with the higher intensities photophobia occurs. The process of dark adaptation pursues the normal course in the totally colour-blind. Exposure to bright light causes prolonged diminution of vision, and recovery follows the course which would be expected if it were dependent upon restoration of the visual purple. All totally colour-blind people have marked diminution of central visual acuity, and in many cases there is undoubtedly a central scotoma. The curve of visual acuity with gradually increasing light-intensities shows no sudden bend, as in normal-sighted people, but continues in the same direction as for lower intensities until photophobia sets in (Konig). The reduction in central vision tends to absence of binocular fixation and divergence of the optic axes. We may conclude that the foveal neuroepithelium, if it retains its function at all, approximates in character to that of the normal rods. Uhthoff and v. Kries^ have shown that the flicker phenomenon in the totally colour-blind follows the same laws as in normal achromatic scotopia. Nagel^ has tested a monochromat with matches between orange- yellow (600 nfi) and greenish-blue (490 fjufx) {vide supra). A monochromat can make such a luminosity match at ordinary intensities of light, where- as a trichromat must be dark-adapted and the intensities must be very low. With the photopic eye the intensity {i.e., slit- width of the spectro- photometer in millimetres) was 0-44. After one hour's dark adaptation it was 0-8. Compared with the achromatic scotopic trichromatic, the achromatopic photopic retina contains less visual purple : hence the trichromatic achromatic scotopic value for greenish-blue light is lowered by absorption relatively to the achromatopic. Night-blindness is found in several diseases of the eye, most commonly in retinitis pigmentosa and allied conditions, least adulterated by other pathological symptoms and signs in congenital night blindness. The last-named is congenital and hereditary^, ^ V. Kries, Ztsch. f. Psychol, u. Physiol, d Sinnesorg. xxxii. 113, 1903. 2 V, Helmholtz, 3rd. ed. n. p. 328. ^ Nettleship, Trails. Ojthth. Soc. xxvii. 1907 ; Parsons, Pathology of the Eye, iv. 1400, 1908. P. c. v. 14 210 COLOUR VISION Parinaud^ regarded night blindness as strongly supporting the views which he propounded and which have since become incorporated in the duplicity theory. Night blindness may be looked upon as the obverse of total colour blindness. In the latter rod-vision is isolated, in the former rod-vision is seriously disturbed or in abeyance. In night blindness dark adaptation may be almost abolished or much slower than normal ; in high degrees it is both slowed, and diminished quantitatively, i.e., the higher degrees of sensitiveness of the retina are never reached and the highest possible are reached only after prolonged exclusion of light from the eye^. In low degrees central vision is normal in the light-adapted condition, and may be so in high degrees of the disease, e.g. in retinitis pigmentosa. Colour vision is normal with the exception of occasional diminution of sensibility for blue lights. In retinitis pigmentosa the field of vision is diminished, often almost down to the fixation point, and dark adaptation may be completely absent. Purkinje's phenomenon is much less marked in the night-blind. If a red and a green are chosen, which are equal in luminosity for the normal and night-blind in light adaptation, and the room is then darkened the green becomes much brighter than the red in the course of a few minutes for the normal, but only after a long time, if at all, for the night-blind. Quantitative experiments show that for lights of short wave-length the increase in luminosity after half-an-hour's dark adaptation is 10 to 100 times as great for the normal as for the night- blind. With stimulation areas exceeding the foveal limits the sensi- bility for mixed white lights increases much more rapidly for the normal than for the night-blind on dark adaptation, but if both are tested with red lights there is very little difference. This is further proof of the diminution in the appreciation of Purkinje's phenomenon by the night- blind. The temporal effects in the night-blind have not been investigated as thoroughly as could be wished, and the results are not concordant. The condition is often, in fact generally, partial. It is not sur- prising therefore that cases occur in which Purkinje's phenomenon occurs and which show some degree of dark adaptation, colourless interval for red pigments and so on. Hess^ has laid great stress upon ^ Arch. gen. de mM. 1881 ; C. r. acad. des sci. 1881 ; La Vision, Paris, 1898. ^ Heinrichsdorff, Arch. f. Ophth. lx. 405, 1905; Messmer, Ztsch.f. Sinnesphy^iol. xixi. 83, 1907; Lohmann, Arch. f. Ophth. lxv. 3, 1907; Stargardt, op. cit. Lxxni. 1, 77, 1909; Behr, op. cit. Lxxv. 201, 1910 ; Wolfflin, op. cit. lxxvi. 464, 1910. 3 Arch. J. Avgenhlk. lxii. 50, 1908 ; LXix. 205, 1911. THE DUPLICITY THEORY 211 these cases and regards them as seriously opposed to the duplicity theory. Taken alone the phenomena associated with night blindness are by no means unadulterated evidence in favour of the theory, but viewed in conjunction with the other facts already discussed they are on the whole confirmatory and highly suggestive. Moreover it is by no means certain that in diseases accompanied by night blindness the cones remain unaffected. Except in congenital night blindness it is almost certain that they do not. The pathology of retinitis pigmentosa^ shows that the condition is primarily due to vascular degeneration in the choroid, whereby the chorio-capillaris is destroyed. The pigment epithelium and the outer layers of the retina are dependent for their nutrition upon the integrity of the chorio-capillaris. From its proximity the pigment epithelium must suffer first. The production of visual purple in the rods is un- doubtedly bound up in the integrity of the pigmented epithelial cells, so that primary disorder of the functions of the rods might be confidently anticipated on pathological-anatomical grounds. It is the rule in most nutritional disorders for the most highly differentiated and complex structures to suffer first. That there is good evidence of a survival of the cone-functions after destruction of rod-functions is evidence of an intermediary process such as that described. It must be admitted that the researches of Hess on the comparative physiology of vision show that little support to the duplicity theory can be derived from lower animals. He found {v. p. 132) that all classes of vertebrates possess good powers of dark adaptation, including even tortoises, which possess neither rods nor visual purple. The arguments against the duplicity theory have been collected by Siven^. He lays stress upon the peculiarity of the achromatic scotopic " white " or grey, which he calls blue or violet. He thinks that the rods are chiefly concerned in the perception of lights of short wave- length. He finds the strongest support of his view in the colour-fields, which are admittedly difficult to reconcile with the duplicity theory. He admits the absence of Purkinje's phenomenon at the fovea, and says that when colour is no longer perceived centrally it can still be elicited from the periphery, presumably however only by lights of short wave- length. His experiments in conjunction with Wendt on the effects of santonin poisoning and the yellow vision of jaundice afford him additional 1 Parsons, Pathology of the Eye, n. 602, 1905. 2 Siven and Wcndt, Skand. Arch. f. Physiol, xiv. 196, 1903 ; Sivcn, op. cit. xvn. 306, 1905 ; XIX. 1907 ; Zisch.f. Sinnesphysiol. xlh. 224, 1907 ; Arch, of Ophth. xlh. 2, 1913. 14—2 212 COLOUE VISION arguments, and he regards Hess's observations on birds^ as also con- firmatory. Baiier^ has brought forward arguments to show that the role of the visual purple is not limited to the condition of dark adaptation. Since Klihne's researches it has been known that it is difficult to bleach the visual purple completely so long as the retina is in contact with the pigment epithelium. Bauer found that though frog's visual purple was rapidly bleached when exposed to bright sunlight, the retina again became red after several hours in spite of continuous exposure. He concluded that increased destruction of the substance is associated under physiological conditions of exposure to bright light by correspondingly increased production, and that therefore the visual purple plays an important part in photopic as well as in scotopic vision. The clear explanation which the duplicity theory affords of the chief differences between human photopic and achromatic scotopic vision and of the peculiarities of the intermediate stage, scotopia with only moderate dark adaptation (chromatic scotopia), tempts one to carry the theoretical considerations farther than is perhaps warranted. Dark adaptation appears to be directly associated with the visual purple. On teleological grounds extreme sensitiveness to light and shade are most needed in the dusk, and while the loss of visual acuity for form is a serious disadvantage the gain in concentration is a counter- balancing merit. The rods are more intimately connected with each other by nervous paths than the cones (Ramon y Cajal), so that relatively widespread impulses become summated in the resultant psychological impressions. An even less complete insulation of the rods may be brought about by dark adaptation, possibly through the retraction of pigment in the retinal pigment cells. On the other hand the photopic position of the pigment tends to protect the more sensitive rods with their contained visual purple and to isolate the less sensitive cones. The role of the pigment cells in dark adaptation has already been the subject of conjecture^, but it must be remembered that the wandering of the pigment has not been conclusively proved to occur in warm- blooded animals. 1 Arch.f. Angenhlk. Lvn. 298, 317, 1907 ; lix. 143, 1908. * Arch. f. d. ges. Physiol, cxli. 479, 1911. 3 Exner, Sitz. d. Wiener Akad. xcvm. 3, 1889 ; Nagel, in v. Helmholtz, 3rd ed. n. 331. SECTION III THE THREE-COMPONENTS THEORY (YOUNG-HELMHOLTZ) CHAPTER I STATEMENT OF THE THEORY The Yoiing-Helmlioltz theory of colour sensations is based upon the facts of the mixture of pure-colour stimuli (see Part I, Sec. II, Chap. iii). We found that within a certain range, which includes all ordinary conditions of colour- vision, and with certain well-defined exceptions {v. p. 36), every conceivable light or light-mixture gives rise to a sensa- tion which can be accurately matched by the sensation produced by a suitable mixture of only three lights. If we choose three actual wave- lengths, R, G, and F, on the colour-diagram and join them by straight lines, we find that part of the curve is outside the triangle thus obtained. This means that there are certain spectral colours which cannot be obtained in their full spectral saturation by the mixture of R, G, and V. If we choose a point on the curve outside the triangle, e.g., GBl, we can by experiment obtain an equation of the following form : aR + ^GBl ^yG + eV. We thus find an expression for GBl ^GBl = yG + eV - aR which represents the unmixable colour in terms of R, G, and V. In order to avoid negative quantities we must assume the existence of colour-sensations which lie outside the colour-diagram. The fact that, by previously stimulating the eye with the complementary colour, we can obtain a colour-sensation which in saturation far exceeds that obtained by viewing the spectral colour without such previous excitation {v. p. 109), affords some evidence that this assumption is not unwarranted. 214 COLOUR VISION The assumption is minimised by describing a triangle around the diagram which will only just succeed in including every part of the diagram. Such a triangle is seen in Figs. 8 and 9 (pp. 39, 40). We can then obtain a universal equation for any spectral colour, F, viz. in terms of the three theoretical sensations, R, G, and V, in which X, y, z are all real positive quantities. The objective quantities of light in this equation come into con- sideration here only as sensation-stimuli, and as such have a physically measurable value. If we make a further assumption, viz., that the physiological processes which accompany sensations have a definite quantitative relationship to the physical stimuli, we may say that X, y, and z are respectively the red, green, and violet values of the light F in terms of the fundamental colours R, G, and V ^. It also follows that in the sensory process there must be three corresponding kinds of activity, which coexist side by side without interfering with each other, and upon which every variety of colour sensation depends. Let us suppose that there is some method whereby we can determine three measurable quantities, (f), x^ ^ representing three physiological processes, which taken together completely define the character of the visual sensation. We should then be able in every case to find by observation the relationship between cf), x, »A ^^^ *^® values x, y, z of the incident light. We should then have cf), x? ^ represented as three functions of x, y, z and conversely x, y, z as three functions of cf), x, ^^ That is, , x, "A, therefore x, y, z can be expressed solely by (/), X, ^- These values of the x, y, z functions of <^, x, ^ {'^'-^-ifi^f^ifz') are therefore quantities which depend only upon the character of the sensation, and moreover possess a certain individuality, since each can be aroused, exist and again disappear in the nervous apparatus inde- pendently of the other two and unaffected by them. This independent existence is, however, exactly what we are in search of when we speak of elements, ingredients, or components of the sensation. If we therefore 1 See V. Helmholtz, 2nd ed. p. 341. STATEMENT OF THE THEORY 215 denote the function of , x^ 'P which represents x by R, and the correspond- ing other two functions by G and V, then these quantities, R, G, V, are to be denoted as components of the colour sensation, and similarly any linear function of them (aR + 6G + cV) may also be thus denoted. From the mathematical elaboration of these assumptions v. Helm- holtz stated the Young-theory in the following form : " (1) In some part of the conducting nerve substance, under the influence of coloured light, three different, independent, and mutually unopposed elementary activities arise ; we will them call the elementary stimulations. Their amount is directly proportional to the correspond- ing colour-values, x, y, z of the objective light ; they correspond to the R, G, V of the above description. " (2) All activities passing further towards the brain, as well as the sensations actually entering into consciousness under the given conditions of the reacting brain, are only actions of the three elementary stimulations, R, G, V, and in amount are functions, cf), x^ ^> of those elementary stimulations. " (3) Either the elementary stimulations themselves or three mutually unopposed actions dependent upon them are conducted inde- pendently to the central organ." Putting these conclusions in simpler language and in their most general form, R, G, V, are any three points so situated that, when joined, the triangle thus constructed completely encloses the colour-diagram of the given spectrum. In this manner positive values are ensured. From observations on colour-mixtures with the given spectrum we can construct valency curves which represent the stimulation values of any spectral light for each of the three components, R, G, V, of the resultant sensation. Thus F = xR + yG + zV means that the light F is matched by a mixture of x parts of R light, y parts of G light, and z parts of V light, R, G, V being the physical stimuli in the mixture. If these physical stimuli act respectively upon the physiological counterparts or elements of sensation, R, G, V, then xR, yG, zV clearly represent the strengths with which the light F acts upon the R, G, V elements, i.e., they are its R, G, V valencies or values. The valency curves are therefore nothing more than the gauging curves of the spectrum to which reference has already been made {v. p. 39). But it has been shown that every gauging value belonging to one such group of curves must always be a linear function of the three 216 COLOUR VISION gauging values belonging to any other group of curves. In any given spectrum, therefore, the R, G, V values must be some linear function of the three empirically observed gauging values {R, G, V). The three gauging curves therefore represent the relative values of the three sensations for each light throughout the given spectrum. The simplest concrete conception of the bases of the three sensation elements or ingredients is that there are three components which are counterparts of the physical stimuli. We shall at present use the term component in the broadest sense : it may represent a chemical, an electrical or some other process acting upon different substances or nerve-fibres and giving rise to nervous activity, but the general state- ment of the theory necessitates no such concrete conception. In this, its most generalised form, the Young-Helmholtz theory explains satisfactorily the sensations resulting from colour-mixtures in the normal visual system, since it is indeed founded upon them. It also explains satisfactorily the gross divergences from the normal system. Thus dichromatic vision is due to absence of one of the theo- retical components ; in the protanopes this is the R component, in the deuteranopes the G component, and in the tritanopes the V component. Indeed, the determination of the absent component follows mathe- matically from the correlation of the facts of colour-mixtures in dichro- mats. In their colour-diagram all the colours which appear to them to match must lie upon a straight line, since the line joining any two points representing homogeneous colours contains all the points representing the colours which can be mixed from those homogeneous colours. Similarly the mixtures of any of these colours with any other colours lies on a series of straight lines, v. Helmholtz has shown that all these lines either meet in a point or are parallel^. The point of intersection corresponds to a colour which has no stimulus value for the dichromatic eye. It is generally called the Null-point (Fehlpunkt) of the system. Conversely, the normal colour diagram can be constructed from the combined protanopic and deuteranopic observations (Fig. 57). It cannot be expected to coincide precisely with that of a single individual on account of differences in macular pigmentation, variations in the spectra used, and so on, but it shows a remarkable similarity^. Anomalous trichromatic vision may be regarded in various ways on this theory. The simplest explanation is that it is a reduction system ^ V. Helmholtz, 3rd ed. p. 123 ; Greenwood, Physiology of the Special Senses, p. 153. 2 Greenwood, loc. cit. p. 150 ; v. Kries, in Nagel's Handb. d. Physiol, d. Menschen, in. p, 161. STATEMENT OF THE THEORY 217 in which the stimulus values for any light of one of the components is uniformly less than normal. Partial protanopia is then a diminution of sensitiveness of the R component, partial deuteranopia of the G component. Other modifications are theoretically possible and are included in the general form that if in the normal system ^ = /i {x, y, z) and so on, then in the anomalous trichromatic system (f)' =/i' {x, y, z) and so on. This generalisation is, however, too vague to be of much practical value, and there are other possibilities arising directly out of Bx V.515 AX Fig. 57. Normal trichromatic colour diagram constructed from combined protanopia and deuteranopic observations. A, null-point of protanopes; B, null-point of deuteranopes. (v. Kries.) the Young-Helmholtz theory, such as a shift of one of the curves {vide infra), which account for at least one group of the anomalous trichromats. Monochromatic vision may be explained on the Young-Helmholtz theory by the identity of all three components. In this case 4> -fi (»' 2/> 2) = X =/2 i^> y,^) = ^ =h {^. y^ 2)- The facts do not support this view, and are much more satisfactorily explained by the duplicity theory. At the same time there are some 218 COLOUR VISION observers, e.g., Schenck, who regard monochromatic vision as divided into various types, some of which are due to modifications in the cones and must therefore be correlated in some such manner as that suggested with the three-components theory. Difficulty is experienced when the attempt is made to explain the phenomena of peripheral vision and of induction by the theory. It is at once clear that neither can be explained as evidence of any reduction system. Fick^ suggested that the reaction-values of the components were not the same in all parts of the retina. The suggestion is inadequate ; on the one hand because it implies a simple reduction system and this fails to account for the phenomena, or on the other hand because it is too general to be of practical value. Schenck's attempt to correlate the facts with a modification of the trichromatic theory will be described later. The relationship of normal and deuteranopic peripheral colour vision to central colour vision approximates more closely to a reduction system than the corresponding relationship in the protanopic system. In all, however, the modifications produced by areal effects negative so simple an explanation and afford some evidence in favour of v. Kries' theory of zones. The greatest difficulty, however, is experienced in explaining the facts of induction. In general terms it is not difficult to conceive a diminution in response of the components in one direction, associated with an increase in another, after previous stimulation. We might thus account for the increased response to the complementary after stimula- tion with a given light. Indeed, such a view falls in well with other physiological findings, so admirably elaborated in Sherrington's work. If all spectral lights act upon all three components, then the increased response to the complementary after previous stimulation with a colour can be explained, and this was the view adopted by v. Helmholtz. It lands us, however, on the horns of a dilemma, for the facts of dichro- matic vision — and the same is true of trichromatic {vide infra, Chap, ii) — show that lights of greater wave-length than about 550 /x/x do not act at all upon the V component, since no standard blue has to be mixed with the standard red in order to match colours in this part of the spectrum. Yet the saturation of yellow (589 /x/x) is undoubtedly increased by previous stimulation with the complementary blue. v. Kries has shown that this effect is not to be explained by any alteration of the intrinsic light of the eye, but is caused by a quantitative change in the response to the secondary light. If yellow light acts only on the red and green 1 Arch. f. d. ges. Physiol XLvn. 274, 1890. STATEMENT OF THE THEORY 219 components we cannot explain on the theory the increase in saturation which follows previous activity in the blue components. Hess refuses to admit that the facts can be brought into line with the theory. We must therefore accept the theory as explaining satisfactorily either the phenomena of after-images or those of dichromatic vision, but not both. The large mass of accurate evidence which has now accumulated, chiefly from the observations of Abney and v. Kries and their fellow-workers, showing that all spectral lights do not act upon all the components makes it impossible to accept the explanation as it stands for successive induction. Moreover, successive induction is a complex condition allied to simultaneous induction, and it must be admitted that the facts of simultaneous contrast cannot be explained directly by the trichromatic theory alone. With regard to spatial induction v. Helmholtz made no serious attempt to correlate the facts with the theory. In his opinion the phenomena of simultaneous contrast are purely psychical and are explained as "illusions of judgment." The arguments which can be advanced in favour of this view do not concern the three-components theory, and it is only by modifications of the theory, which will be dealt with at a later stage, that any more purely physiological explanation, on the three-components' basis, of spatial induction can be advanced. Finally, it may well be asked whether there is any direct positive evidence of the existence of independent visual components answering to the theoretical components. It is in this respect that the electrical researches of Gotch, Himstedt and Nagel, and others, and the fatigue experiments of Burch are so valuable. They offer no crucial proof, but they are suggestively concordant with the view that such com- ponents exist. It has already been mentioned that Burch is inclined to predicate four. Theoretically this is of no importance, though it complicates matters and is unnecessary for the explanation of the facts of colour-mixtures, etc. It does not in any way detract from the support which Burch's experiments afford of an additive theory de- pendent upon processes occurring in certain components. As is usually the case, the three-components' theory becomes less plausible the more concrete the form it takes. We have seen that Young suggested three different types of nerve-fibre, others have suggested three substances which undergo chemical change, others again have suggested electrical changes. Each such hypothetical fibre or substance, etc., leads inevitably to further hypothetical con- ceptions of the details of the processes. These become the subject 220 COLOUR VISION of controversy and the whole theory is thrown into disrepute, the cardinal fact being overlooked that they are in no integral sense germane to the theory. It may, however, be interesting to record some of the suggestions which have been made. As regards the three fundamental component sensations Konig's experiments led him to regard them as a red just outside the spectrum and rather more purple than the spectral colour, and a blue at 470 fifji slightly more saturated than the spectral blue. v. Helmholtz, on the basis of Konig and Brodhun's observations on the sensitiveness for discrimination of hues in the spectrum {v. p. 30), chose a carmine-red, bluer than spectral red, a yellowish-green between 540 /x/a and 560 /x/x, and an ultramarine blue, all much more saturated than the spectral colours. As regards the retinal bases for the theory, Konig as already mentioned regarded the visual purple as the basis of colourless scotopic vision. With greater stimulus intensities the visual purple is trans- formed into visual yellow, which forms the basis of the fundamental blue sensation. He placed the substrata of the red and green sensations in the pigment epithelium. If this were true, foveal vision should be monochromatic in protanopes and deuteranopes. He regarded the cones as a purely dioptric mechanism. CHAPTER II RESEARCHES BASED UPON THE THEORY I, Normal Colour Vision The first researches based on Young's theory were made by v. Helm- holtz (1852-3)^. From them he elaborated the three-components or Young-Helmholtz theory. He was followed by Clerk-Maxwell (1855-6)^. Clerk-Maxwell placed three slits in the spectrum of daylight at the following places : (1) in the red {R) between the Fraunhofer lines C and D, twice as far from the latter as from the former ; (2) in the green (G) near E ; (3) in the blue (J5) between F and G, twice as far from the latter as from the former. Lights from these slits were mixed in the proportions to match a comparison white {W) derived ^ References in v. Helmholtz, 3rd ed., ii. p. 137. - Scientific Papers, Cambridge, 1890. RESEARCHES BASED UPON THE THEORY 221 from the same source of light. From the sUt-widths he thus obtained a standard equation : He then obtained 14 other equations for W, in each case mixing three Hghts from different parts of the spectrum in suitable proportions. By- eliminating W from each of these equations he obtained the sensation- values of 14 positions in the spectrum in terms of the sensation- values of the three standard colours, R, G, B. Fig. 58 shows the curves plotted from these values. It will be noticed that the R and B curves pass below the base line. His standard colours were therefore not chosen so as to eliminate negative values. Red. Blu.e. Fig. 58. Clerk-Maxwell's sensation curves. The dotted line is the algebraical sum of the ordinates at each point ; it is not a true luminosity curve. Abscissae, wave- lengths of the prismatic spectrum of sunlight ; ordinates, arbitrary scale. (Clerk- Maxwell.) Papers followed by J. J. Miiller^, Preyer^ and Bonders^, but no other researches on the three " sensation curves " were made until Konig embarked upon his observations in 1883, continuing them until his death in 1901. Almost simultaneously, Abney commenced the work which he has recorded in a series of papers, culminating in his Researches in Colour Vision (London, 1913), and which is happily still in progress. Certain points on the sensation curves have also been worked out inde- pendently by F. Exner^. 1 Arch.f. Ophth. xv. 2, 208, 1869. " Arch.f. d. ges. Physiol, i. 299, 1869. 3 Arch. f. Ophth. xxiii. 4.. 282, 1877 ; xxvn. 1. 155, 1881 ; xxx. 1, 15, 1884 ; Onderzoek. i. Lab. Utrecht. 1882. * Sitz. d. Wiener Akad. cxi. i ; a, 1902. 222 COLOUR VISION Konig's experiments on normal colour vision were carried out in conjunction with Dieterici^ and an abstract was communicated to the British Association in Birmingham in 1886. The curves were calculated from a vast number of matches made with Helmholtz' spectrophoto- meter. An objection to this instrument is that the variations in in- tensity are made by a polarisation method and account is not always taken of the appreciable polarisation of the light by the prism of the spectroscope itself. The observations and calculations were of the same kind as those made for dichromatic systems (v. p. 163), but were necessarily of a more complicated nature. Fig. 59 shows the curves 720 700 680 660 640 620 600 580 560 540 5 20 500 480 460 440 420 400 380 a B C P E b F G H Fig. 59. R, G, and V. sensation curves. K, Konig's G-curve; D, Dieterici's G-curve; G-curve of an anomalous trichromat. Abscissae, wave-lengths of the inter- ference spectrum of sunlight; ordinates, arbitrary scale. (Konig and Dieterici.) referred to the interference spectrum of sunlight. Konig's and Die- terici's R and F curves coincide ; the G curves are slightly different. The distortion of the G and F curves between 535 /x/x and 475 ju,^ is due to absorption by the macular pigment. By far the most exhaustive experiments have been made by Abney, and they will be discussed in greater detail. He has adopted another method of determining the three sensation curves, dependent upon the luminosities of the colours. His results confirm and correct in detail the curves obtained by Konig. The principle of the experiments is as follows^. ^ Konig and Dieterici, Sitz. d. Akad. d, Wiss. Berlin, 1886 ; Ztsch. f. Psychol, v. Physiol. d. Sinnesorg. iv. 241, 1892 ; m Konig, pp. 60, 214. 2 Watson, Proc. Roy. Soc. Lond. A, lxxxyui. 404, 1913. RESEARCHES BASED UPON THE THEORY 223 "We have already described the methods of obtaining the luniinosity curve of the spectrum. In it the abscissae are wave-lengths of the spectrum, the ordinates luminosities in arbitrary units, the maximum brightness being 100. Suppose that when making a set of observations we start with the movable slit at the extreme end of the spectrum and determine the intensity iVi of the white which appears of the same brightness as the colour and then move the slit towards the blue through a distance equal to its width and again determine the intensity Wg of the white, and so on throughout the spectrum. We should in this way determine piece by piece the brightness of the whole spectrum, and the sum Wj^ + w.2 + w^ -}- , etc., could be taken to represent the total brightness of the whole B 700 600 400 Fig. 60. R, G, and B, sensation curves. These are Konig and Dieteriai's curves cor- rected to new determinations of the points of section, n, h, c, d. Abscissae, wave- lengths of the interference spectrum of the arc light ; ordinates, arbitrary scale. (F. Exner.) spectrum. If now the slide carrying the slit were removed, so that the light corresponding to the whole spectrum were allowed to fall on the same portion of the screen, thus forming white, and we now determine the intensity W of the comparison white which is equal in brightness to this recombined spectrum, then it has been shown by Abney, Tufts, Ives and others that y^ ^ w^ + W2 + w^+ , etc. That is, the luminosity of the recombined spectrum is equal to the sum of the luminosities of its parts. Now the sum w-^^ + Wo-\- w^-\- , etc., is proportional to the area 224 COLOUR VISION enclosed by the luminosity curve. Hence the area of the luminosity curve represents the total brightness of the light which is formed into the spectrum and is, therefore, a constant whatever the condition of the vision of the person who makes the observation. This is at once apparent if we remember that if the brightness of the whole recombined spectrum is compared with the comparison white, since these whites are derived from the same source and must, therefore, have exactly the same composition, a setting which appears correct to one person must also appear correct to any other, whatever the differences which may exist between their vision. On the Young-Helmholtz theory the sensation produced by light of any colour is the sum of the sensations due to stimulation of the three component mechanisms. Konig and others obtained their curves for the three components by matching light of one wave-length with a mixture of lights of two other wave-lengths. Abney, using a similar method, adopted the principle that the sum of the ordinates of the luminosity-values of each component for any given wave-length is equal to the ordinate of the luminosity curve of the spectrum for that wave-length. He then proceeded thus^ : The red sensation can be perceived in purity at one end of the spectrum. From the darkest red to a point near the C line, a little above the red lithium line, the colour is the same, though, of course the brightness varies, but the brighter red colour can be reduced so as to form an exact match with the dark red, and no mixture of any colours will give a red of the description we find at the end of the spectrum. At the violet end of the spectrum we also find that the colour is the same throughout, from the extreme visible limit to a point not far removed from G, but it is not for this reason to be accepted that the colour is due to only one sensation. It might be due to two or three sensations if they were stimulated in the same proportions along that region, and if the identical colour could be produced by the combination of other colours. Experiment shows that a combination of two colours will under certain conditions make violet, and that instead of a simple sensation of violet we have in this region a blue sensation combined with a large proportion of red sensation. If we know the percentage composition of the violet mixture we may provisionally use this part of the spectrum as if it excited but one sensation, and subsequently convert the results obtained with it into the true sensations. Thus in calculating the percentage of red in any colour, that existing in 1 Abney, Phil. Trans. Boy. Soc. Land, cxcin. 259, 1899 ; ccv. 333, 1905. RESEARCHES BASED UPON THE THEORY 225 the provisional violet sensation would have to be added to it, and the same amount be abstracted from the violet to arrive at the true blue sensation. The green sensation would remain unaltered. Having at one end of the spectrum a pure red sensation, and at the other mixed sensations, due to the stimulation of the red and a blue sensation, it remains to isolate the green sensation. Owing to the over- lapping of the curves in the green of the spectrum, due to the fact that this region stimulates all three of the sensations, the effect of the pure green sensation is never experienced by a normal eye. In any colour where the stimulation of all three sensations occurs there must be always an admixture of white light, and we have to search for that point in the spectrum where white alone is added to the green sensation. The following diagram, Fig. 61, will show some variations in com- position of a colour that may be met with. The provisional use of a d. e a 6 c R G A G B G C Eig. 61. Diagrams illustrating Abney's method of determining the normal sensation curves. violet sensation will not alter the argument, since, as before said, we may replace it by blue and red sensations. The different figures are purely diagrammatic. They are constructed on the supposition that equal heights of line above the base line show the stimulation necessary to give the effect of white light. The scale applicable to each of the three lines is necessarily quite different in the scale of luminosity ; that of the violet in particular is very greatly exaggerated. A, B, and C represent colours each containing a sensation of white. Let the stimulation of the sensations be represented by vertical lines. In A we have the red and green sensations of equal heights, but V is less. Drawing a horizontal line through c, aR, bG, and cV, represent equal stimulations, which make white, leaving da and eb equal. We thus have a colour which is made up of a mixture of R and G sensations {RS, and GS), together with white. Now equal stimulations of RS and 15 p. c. V. 226 COLOUR VISION GS, we shall see later, give the sensation of yellow. If we place a slit in the violet and move another slit along the less refrangible part of the spectrum, we shall find a place where this colour and violet together make a white, the slits being opened or closed to make the match. This position, then, is that in which the red and green sensations are equally- stimulated, and answers to ^. In 5 we have a green and violet with equal ordinates and a deficiency of red. If we place a slit in the red and move another about in the green, we shall find a colour which with the red makes white. This position, then, will have an equal stimulation of green and violet. This gives another fixed point. The next point to determine is shown diagrammatically by C, which illustrates the green we have to look for, mixed only with white. This is more difficult to find, as it would require a purple to be added to make a match with the white, and this does not exist in the spectrum. Suppose we mix A with B, we get a diagram of the kind shown in the fourth diagram. There are equal reds and violets stimulated, but a larger stimulation of green sensation. This gives a colour paler than the spectrum colour, but still a green which can be matched. There are also other plans, dependent on trial and error, for fixing this point. When the observations for obtaining the fixed points have been made it will be found that the complementary of the violet is at scale Qumber {SSN) 48-7 (577-2 /x^u), that of the red at SSN 34-6 (500 /x/x), and that where green is mixed only with white is at SSN 37-5 (515 /x/x). These are therefore the points of intersection of the three sensation curves (Fig. 64). It is next necessary to ascertain the amount of white in the green at SSN 37-5. One slit is placed in this situation and another at SSN 59-8, the position of the red lithium fine. The luminosities of these colours, with equally wide slits, are taken by comparison with another light, such as yellow or white. They are found to be 39-2 and 9-4 respectively. A patch of yellow light from a second spectrum, derived from the same source of light, is placed beside the mixture of red and green. In order to obtain a colour match it is necessary to add white to the homogeneous yellow. From the slit widths and the luminosity of the added white an equation in terms of luminosity is obtained : a (yellow) + b (white) = c (red) + d (green) or a (yellow) = c (red) + d (green) — b (white). Since the red contains no white the percentage of white in the green RESEARCHES BASED UPON THE THEORY 227 is ^xlOO. It was found to be 69 per cent. The percentage of green sensation is ^ x TOO. a Having obtained the percentage of white in the green the percentage sensation composition of other colours in terms of luminosity can be readily found ^. Fig. 62 shows the percentage sensation curves. From the data obtained by these experiments the analysis of the sensation-values can be carried further by calculation. Having deter- mined the luminosity values of the different parts of the spectrum by heterochromatic photometry (585/zju. = 100) the luminosity-values of 100 90 60 70 CO 30 40 30 20 10 ^ ^ .^"^ ^y ___ \ y ' x^ \ y \ ^ ^ t--^. 7^ ____—- vH ^^•Jt / \ ^v^^ — -^^ / v. \ K ,>^ ^ ^ '"^"r.. ..^-^^ ^ Li 1 ^^ y ] i ^^iSiia ^S^SCi/ 1. . . . MY f IS 20 25 30 35 40 46 50 Scale of Prismatic SpecfruTn. 55 60 fig. 62. Percentage of the B. G, and B sensations in the spectrum colours in terms of luminosity. Abscissae, wave-lengths of the prismatic spectrum of the arc light ordinates, arbitrary scale. (Abney.) the R, G and V sensation can be readily obtained from their percentage- values. Further it is found that the violet, which has hitherto been used as a simple sensation, is made up of 72 per cent, of red sensation and 28 per cent, of blue sensation. The R, G, and V values can there- fore be transformed into fundamental R, G, and B sensation-values, B being the fundamental blue sensation. We thus obtain the R, G, and B percentage-values, and from them, by calculation, their luminosity- values. The luminosity curves of the spectrum and of the three components are shown in Fig. 63. For any wave-length the sum of the ordinates of the three sensation curves is equal to the ordinate of the total luminosity curve at that point. When white light, i.e., light of all wave-lengths, enters the eye the effect produced on the three components will be 1 Abney, p. 235. 15—2 228 COLOUR VISION proportional to the areas of the three sensation curves. The relative areas of these curves for the electric arc are RS 579, GS 248, BS 3-26. When considering the matches made between different lights it is often convenient to adopt a different scale, viz., one in which the areas of the three sensation curves are equal to each other. With this scale equal ordinates of the three sensation curves correspond to a mixture which will appear white to the normal eye. If we multiply the ordinates of the green sensation curve by 579/248 = 2"21, and those of the blue sensation curve by 579/3-26 = 117, the green and blue sensation curves will have the same area as the red sensation curve. 100 90 80 70 60 60 40 30 20 10 / ^ N, /. s \ / / "N \ \ \< / \ ^o 1 K f \ %. \c C- ) \ f^-""- ^: / r>% \Ofj_ ^ 7^ V \ / ^7 ^ ^ \ \ / .4 y ^ N \ N ^ / ^ -^ 7 Bi UE SEN SAT ro/v ^i;::: ^2 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 26 24 7o|oo" ' I lesjool I ' I 6o|oo • ' ' ' 5o[oo I I i I 5o[oo*u ' • I '" Fig. 63. Abney's E, G, and B sensation curves in terras of luminosity. The sums of the ordinates at any point are equal to the ordinate of tlie photo pic luminosity curve. (Watson.) In the curves in Fig. 63 the source of light was the crater of the arc light with a horizontal positive pole. The source of light in Abney's original investigations was the crater of the arc light with sloping carbons, and the corresponding factors are 2-3 and 190. The luminosity curves are therefore brought to equal areas by multiplying the GS and BS luminosity- values by 2-3 and 190 respectively, thus giving the curves shown in Fig. 64, where equal stimulation of all three components, i.e., equal ordinates, give the sensation of white. Sir William Abney was the first to work out the colour sensations of the normal trichromat exhaustively in this manner. RESEARCHES BASED UPON THE THEORY 229 These results afford an explanation of the change in hue which occurs when white light is added to the spectral colours. When this is done the red becomes pinker, the scarlet orange, the orange yellow, the yellow green. At SSN 48-7 (577-2 /x/a) in the yellowish-green, no change in hue occurs. Beyond this point, the green becomes yellowish, the blue shows little change, and the violet becomes nearly salmon-pink. Now SSN 48-7 is the point where the red and green sensation- values are equal to their sensation-values in the white light. It therefore seemed probable that the change in hue in lights of medium and long wave-length was due to the addition of the red and green sensation- values of the white light, the value of the blue sensation being so small ^ \ ./ \ -1 { I \ / ^ \ i '^ / / v^ 1/ \ K 1 ' / / x ^ f ^ t ^r D 1, ... "^ k 80 70 60 50 40 30 20 10 20 25 50 35 40 55 60 65 Fig. 64. Abney's R, G, and B equal-area sensation curves. The sums of equal ordinates of the three curves at any point represent the sensation of the unanalysed white light. (Abney.) as to be negligible. This conjecture was fully borne out by experiment, the matches being identical with those calculated, within the range of experimental error. SSN 48-7 is easily found, since it is. the comple- mentary colour of the pure blue of the spectrum. In the colour triangle (Fig. 9) a mixture of white with any spectral colour is represented by a point (a) on the line joining W with the given point on the curve, the position of a being determined by the relative amounts of the colour and white. The change in colour due to the admixture with white is found by joining the points B and a and producing Ba to meet the curve. The mixture will match the spectral a 6 c d 573 503 496 450 577-2 515 500 577 508 494 475 230 COLOUR VISION colour which is represented by the point of intersection with the curve. That this should be the case shows that the blue element is practically in abeyance. In experiments in matching white with mixtures of spectral rays it was found that about 2-8 per cent, of red or green could be added without being perceived. If we examine the sensation curves shown in the previous diagrams we shall find that, besides the maxima of each sensation, there are important points where the curves intersect. We will call the point where the R and G curves intersect a, R and B, b, and B and G, c. These points have been determined by Konig, Abney, and Exner^, and their values are as follows : Konig (sunlight) Abney (arc light) Exner (arc light) d is the complementary colour of a. From these data Konig and Exner deduced the fundamental colour sensations : R, a purplish red (com- plementary to 494 /XjU. (Exner); G, 505 fx/x (Konig), 508/z/la (Exner); B, 470 fxfjL (Konig), 475 />t/x (Exner) ; all, however, more saturated than the spectral colours. We have here the explanation of the variations in the discrimination sensibility for hues in the spectrum {v. p. 30). In the equal area curves, e.g. Fig. 64, equal ordinates at any point make white. If we subtract the white in those parts of the spectrum in which all three sensations are stimulated, we can obtain the ratios of the ordinates of the sensation curves to each other throughout the spectrum. These ratios will vary in different parts. In regions where the variation occurs most rapidly we should expect the hues to change most rapidly. Calculations made by Steindler on this basis from Exner's values show that the positions of maximum change are at 500 /i^u., 570 — bdOfxfx, and 635 fifi, of minimum change at 470 /x/x, 530 fif-t, and 625 /x/x. These calculations agree remark- ably well with the results of direct observation, which show that the maxima of discrimination sensibility for hues were at II, 492 /x/x, III, 581 /x/x, and IV, 635-5 /x/x, the minima being at 458 /x/x, 533 /x/x, and 627 /x/x {v. p. 31). This very striking confirmation of the Young-Helmholtz theory was first pointed out by Konig^. The curves in Fig. 65 are derived by calculation from the same 1 Sitz. d. Wiener Akad. cxi. iia, 857 1902. ^ Konig, p. 106. RESEARCHES BASED UPON THE THEORY 231 numbers which gave the curves in Fig. 62 1. Thev are useful for fore- telling the results of fatiguing the eye for different colours. They are constructed on the basis that equal stimulations of all three components give rise to the sensation of white. If the eye is fatigued with white light the effect will be to tire each equally, and therefore the same white seen by such an eye will appear darker. H the eye is fatigued for SSN 48-6, RS and GS are equally fatigued. Suppose they are fatigued so that the ordinates of the curves are reduced to one-half, and then a colour, about SSN 42, of which the normal composition is IRS to 2GS, is observed. The effect will be to make the new relationship ^RS to IGS. Since the proportion remains the same no change in hue ^ ^ ^ ^^ ^ ^\^ ^ , A f % Gret n Se, ^ff/S 60 _> 0 SO 3 ao S 50 I 20 "^ 10 20 22 24 26 26 30 32 34 36 36 40 4! 44 46 *8 SO 52 S4 S6 Se 60 Scale of Spectrum SSN. Fig. 65. Percentage of the R, G, and B sensations in the spectrum colours when e(|ual stimulations of the three sensations give rise to the sensation of white. (Abney.) will be observed. Suppose, however, a colour about SSN 60, where there is no GS is observed. The only effective fatigue will be for RS, and the red will merely appear darker. One more example. Suppose the fatiguing colour is about 42, where the ordinate of GS is twice that of RS, and suppose the fatigue reduces the ordinates to one half. They are then RS = \ and GS = |. The eye then observes SSN 50, where RS is 54-5 and GS 42-5. As observed RS will be ^ = 13-6, and GS will be -^ = 21-25, i.e., as 1 : 0-64. This is about the ratio found at SSN 44. Therefore the yellow at SSN 50 will appear to the fatigued eye as a green of SSN 44, 1 Abney, p. 240. 232 COLOUR VISION Experiments made by Abney support these conclusions. He has shown how the factor of fatigue can be arrived at and how luminosity curves can be obtained for the fatigued eye. He did not employ the high intensities used by Burch. II. Dichromatic Vision Young himself^ suggested that Dalton's colour blindness was due to absence of the red component, v. Helmholtz, Donders, Konig, and others found that this hypothesis accounted well for the facts, and further, that the second great type of colour blindness, now known as deuteranopia, could be attributed to absence of the green component. The rare cases of tritanopia were attributed to absence of the blue component. If the red component is absent it will be seen from Fig. 64 that from SSN 65 to 60 there will be no sensation of colour, nor indeed of light. The red end of the spectrum will be shortened. From SSN 60 to 50 the green component only is present. It will give rise to a sensation much exceeding in purity that of the normal unfatigued eye. The sensation will differ only in intensity, not in hue, just as in the red of the normal eye between SSN 65 and 60. As has already been pointed out we have no certain knowledge of the actual sensation which the protanope is conscious of, and there can be little doubt that it is very different from what we call " green," though he may call it green and often does ; yet he often, too, calls it red. At SSN 50 the third component is stimulated. At SSN 34-6 (500 /z/x) the green and blue sensation curves intersect. As the ordinates are equal he will have a sensation similar to that of the trichromat when all three components are equally stimulated, i.e., he will have a sensation of the same nature as that which he obtains from white light, and he will therefore probably call it white. At any rate he will match it with the white of the combined spectrum if the intensity is suitably arranged for him. In other words SSN 34-6 is his neutral point {v. p. 163). Between SSN 50 and 34-6, then, his colour-sensation, whatever it is, is becoming less saturated. Beyond SSN 34-6 his colour sensation becomes more and more saturated, until at SSN 16, where the green component ceases to be stimulated, it reaches its full spectral saturation, and remains the same, dift'ering only in intensity, to the end of the spectrum. 1 Lectures, ii. 315, 1807. RESEARCHES BASED UPON THE THEORY 233 If the green component is absent the visible spectrum will be the same length as to the trichromat. From SSN 65 to 50, and from SSN 16 to the end of the spectrum he will have isolated pure sensations, "red" and "blue" respectively. At SSN 37-5 (515 /^^la), where the two equal-area curves intersect, he will see " white " : this is his neutral point, and we see that it is rather nearer the violet end than in the protanope. In some parts of the spectrum it is easier for the dichromat to measure the luminosities of spectral lights than for the trichromat. In fact near the neutral points it is a simple matter of homo- as opposed to hetero- chromatic photometry. Theoretically, the protanopic plus the deuter- anopic luminosity curves should equal the trichromatic plus one blue luminosity curve. The extra blue curve makes very little difference, as it is relatively very small compared with the other two {v. Fig. 63). Moreover, its effect is further diminished by macular absorption. As already mentioned, we have no means of making absolute measure- ments, but the relative values hold good. We should expect the discrimination sensibility for hues in the spectrum to be very defective in dichromats. The first experiments of this nature were made by Brodhun^, himself a deuteranope. His curve is given in Fig. 4. We see that he has only one maximum, at 500 /x/Li, near the F line, but here the curve is very sharp and the value of SA is very low, so that his discrimination sensibility is very great at this point. Steindler^ examined three protanopes and five deuteranopes (Figs. 66, 67). She found that SA was much larger than for the normal in all cases. The deuteranopes showed only one maximum, the average being at 503 [xfi. The protanopes showed two maxima, at 500 /x/a and 598 /x/x. On the same basis of calculation as for trichromats {v. p. 230) the maxima for deuteranopes should be at 500 /a/x and 635 /x/x. The first position agrees well with that found experimentally ; at 635 jjifi the luminosity was too low for deuteranopes for accurate obser- vations to be made. The calculated maxima for protanopes were 500jLt)M and 600 /Xju, which agree excellently with the observations. Tritanopes are so rare that they could not be examined ; their theoreti- cal maxima are at 540 /x/x and at 620 /x/x. Konig^ calculated that the number of hues which a deuteranope could discriminate in the spectrum was 140, as compared with the normal 1 Ztsch. f. Psuchol. u. Physiol, d. Sinnesorg. in. 89, 1892. 2 Sitz. d. Wiener Akad. cxv. ii a. 115, 1906. 3 Konig, p. 368. 234 COLOUR VISION 160-165. The number seems high, especially when one considers that the integration interval is much less. It is explained by the fact that between the lines E and F deuteranopes possess a much greater dis- crimination sensibility than normal trichromats {v. Fig. 4). Konig thought that the number for protanopes was probably about the same as for deuteranopes. Fig. 66. Curve of discrimination sensibility for hues of tlie interference spectrum of the arc light for a protanope. # Neutral point. Abscissae, wave-lengths ; ordinates, differences of wave-length (5X) capable of being discriminated. (Steindler.) Liebermann and Marx^ have compared the discriminative sensibility for hue of a protanope and trichromat by the method of mean error. They found that of the protanope much inferior. At 509-6 /x/x, near the neutral point, the mean of 50 observations gave 8A = 3-6 /xju, for the normal, 16-2 ju,^ for the protanope. ^ Ztsch. J. Sinnesphysiol. xlv. 103, 1911. RESEARCHES BASED UPON THE THEORY 235 Bruckner and Kirsch^ found that the chromatic action time (v. p. 91) for deuteranopes, especially for red and green, is considerably greater than for normal trichromats. No experiments of this nature have been made upon protanopes. ^00 50 1 , , n-o \ '■ ' to-o li I 80 ji / C-0 \ ^ / '4~'^^M^^—\>^ ^ /v^55||^In ^ 24^5^7 li&^ S: Jtll^jj ^^^\ \ IWIZTI S^^X^^ JM^-tt Sss^ S ImriJ- ^^t%\K timJ^2^ twi\-^ jfm / // 9'o;i^.\\ \ ^jfjkk k |\\^.\-!:A jkmi^J^ ^mxlx Mtt^ 7 ^l^vX -Mt-J t ^kW\^ IjWrr 1 ^M^ -iWT t m%S m^ J- S|^\ ^// -/ ^§ii ^i£^ y ^§1 ^ 2 ^ ^--? )i I I lecjoo' I I I op I I I I SOpOAU Fig. 69. Luminosity curves of a normal trichromat with excessive macular pigmentation. + + + foveal ; O O O Q parafoveal. (Watson.) of measuring directly the amount of deficiency of red or green sensation in any protanope or deuteranope, whether partial or complete. All that is necessary is to obtain the examinee's flicker luminosity curve with the same source of light and under as nearly identical conditions as possible. The fact that all the curves intersect at a particular wave-length affords one means of deciding that the case belongs to the group of approximate dichromats. His average readings at this point should agree with the average readings of the normal trichromat, and in every case the normal readings should be taken at the same time. For purposes of diagnosis RESEARCHES BASED UPON THE THEORY 241 it is not necessary to take readings at all points of the curve. Readings at SSN 48-8 (5770 A.U.), the point of intersection, and at SSN 42*8 (5410 A.U.), in the green, and SSN 53-2 (6090 A.U.), in the red, usually suffice. We have now to consider some divergences from the calculated curves. In most cases these are slight and are limited to the green- blue, where differences may be expected, since the effect of the blue sensation has been neglected in the calculated curves ; and further, slight differences in macular pigmentation begin to be important in this region. These divergences do not vitiate the method for purposes izu — / . .^Y \ -/Zc^m^Z\ "^ qn LCZa%Z%^Z\ K -^J^it '"^^^^tK ^ 80 ttlLt%l 1^^^"^ -StA-Tt SgS^^^ 70 IWAriJ- ^%.\^z^ tiW--ri- xiSv^^^- 60 -M-'T- t ±-Sg^\ ^ ®° mU q^o ^5^S^S^^^ 50 mW-I: t^ twi^x ^° j!LL^tri$i ^ty^^ 40 'iiM'J-^ J"^ _s^\mv\v ^° mt-r 7 ^^.^^^ in '/LT^ 1 o\k\^\ ^° Mtj' 4 5Slx^^ 20 Jm-t t §SSS '° M^ 4 T Si§^^ ,0 Mv -y^ §11^^^ 10 'My'/ y ■ ■ *$^ * ^^ ^^ 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 Fig. 70. Luminosity curves of a normal trichromat with subnormal macular pigmentation. + + + foveal ; O O O parafoveal. (Watson.) of diagnosis of protanopes and deuteranopes, since they occur in a part of the spectrum where their characteristic colour defects are least important from the practical point of view. Cases, however, occur of marked excess or deficiency of macular pigmentation, and in these the luminosity curves show marked differ- ences from the normal (Figs. 69, 70). Fig. 69 shows excess of pigmenta- tion. Fig. 70 deficiency. By using a central fixation dot and a white ring, the inner and outer edges of which subtend visual angles of 3° 22' and 5°, Watson was able to obtain luminosity curves of a parafoveal p. c. V. 16 242 COLOUE VISION area. Fig. 71 shows his foveal and parafoveal curves. Unusual macular pigmentation may be combined with approximate dichromatism, as in Fig. 72, where partial deuteranopia coexists with excessive macular pigmentation. Other methods can be employed for diagnosing approximate dichro- matism. In addition to taking the ordinary equation, i.e., matching the white of the source of light with a mixture of homogeneous red, green and violet, a pure colour may be matched with a mixture of two pure colours, as in the Rayleigh equation ; or a mixed colour, such as that obtained by passing the white light of the re-combined spectrum 120 110 100 90 80 70 60 50 40 30 20 10 64 62 60 56 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 70|00' ' I ' 65[nol ' I ' ..o|on I t 1 t ..joo till jojooAU ' ' Fig. 71. Luminosity curves of a normal trichromat with average macular pigmentation. + + + foveal ; 0 O O parafoveal. (Watson.) through a saturated solution of potassium chromate, may be matched with a spectral colour. A saturated solution of this salt, | inch thick, will filter off nearly all the blue in the white light, so that the light appears yellow. Now, from about SSN 50 to the red end of the spectrum there is no measurable amount of blue sensation present. If a slit is caused to traverse the spectrum a position will be found where the light exactly matches that transmitted through the chromate. For the normal eye it is at SSN 49-6 (583)Lt/i). If a partial deuteranope makes the match the /— »^_r)y'^ "N / ' ^i— ■ ^ /7lc<^n%%Z\ ^ rrztvK%%Z\ ^ vsz^s^" SSS^^ V wJirn W^%^t- JWAtt W^- V rJn& ^^^^ ^ ° ^23^^ k ^^^ % VM-J ^ ^S&s^^^ it Mau- It- ^s^§$vS dLt t P °^^\ N MW1-^J~ ^ix\ mti- t ^SSo^ -M-M- t ^%Wk wr/7 / ^^\K M^^ 7" ^iyis>- M/ /- i|3^-.GL '©-' y^ ■^ RESEARCHES BASED UPON THE THEORY 243 slit must be moved towards the red, if a partial protanope towards the green. There is a considerable band of the spectrum over which the slit may be moved without vitiating the approximate dichromat's match, but Watson^ has pointed out that the match which satisfies the normal person is always one limit of the band, though the band lies sometimes on one side and sometimes on the other of this position. There is at present no explanation of this fact. We shall find that the izu — ■i ^ r\ /^ — ■^ f r N \ 1 r\r\ / •^ fS- L s \ ^ '/, / K ^ ^ S, \ on /t /^ ^\ '/ * ^ s s. \^ r> J / / /, vN v\ V s. V' \ / / \l V ^^ \ \ o IX V / ' \\ \S \. ^ \ "7ri i . / / i \ \ ^: S^ \ 70 u ^ / \' ^^ r*'' S-f. \ p.n / ' if ' / "if N^ § v\ % IT \ I'd ^o^ l\ V] N FLA Ijk / ^ 1 5 \ C" y> \ r\ h I'l / lS ^ P- y ' N- 64 62 60 58 56 54 52 50 48 46 44 42 40 38 36 34 32 30 28 I r 1 I I I I I I I I I I I I „, 7000 65 00 6000 55 |oo I I 50 00 AD Fig. 72. Foveal luminosity curve of a deuteranope with excessive macular pigmentation. (Watson.) anomalous trichromat whose peculiarity is due to a shift of one of the sensation curves makes a much more accurate setting by this method. IV. Anomalous Trichromatic Vision Shift of a Sensation Curve Abney and Watson^ have pubHshed an exhaustive account of a case of anomalous trichromatic vision in which the green sensation curve is of normal shape and size but is displaced towards the red end of the 1 Proc. Boy. Soc. Lond. May 28, 1914. « Ibid. A. LXXXIX, 232, 1913. 16—2 244 COLOUR VISION spectrum. There is good reason to think that cases of the same nature are not uncommon, but in the cases hitherto examined the shift has always affected the green sensation curve and in the same manner. Figs. 73 and 74 show the normal sensation curves in terms of luminosity, Fig. 73 being the equal areas curve. If an observer is an " approximate dichromat," e.g., a partial ^ 22 24 26 ze 50 32 M 36 58 40 42 44464650525456566062 ^ 4^0 5/000 5500 6,000 6i500 zoooA.U Fig. 73. Normal sensation curves in terms of luminosity. (Abney and Watson.) ;0I2I4.I6ie2O222426 28 50 52>t3636 4042444. 23 1911. OTHER THEORIES 299 infant. If, however, this view is erroneous, the evidence is in favour of delayed development of perception of blue. Further, if congenital colour blindness is atavistic, then such evidence, again uncertain and little trustworthy, as we have of the nature of the sensations experienced by gross cases of dichromatism — and most of these presumably belong to Edridge-Green's class of dichromics — is that the two colours they perceive resemble the blue and yellow of normal vision. " Anomalous trichromats " certainly perceive more than two colours, but we have adduced reasons for be- lieving that this group includes different types, and it is by no means certain that any of these are due to atavism. If they are, the problem of discovering the nature of their colour sensations as compared with the normal is even more difficult than in the case of the dichromats, and certainly no dogmatic statements can be rightly made in the present state of knowledge. ^ INDEX OF SUBJECTS {The numbers in heavy type indicate the principal references to the subjects.) Absorption, 6 by visual purple, 13, 205 Achromatic acotopia, 14, 53 Achromatopia, 186 Action-time, 90 chromatic, 91, 235 Adaptation, 17, 49, 203, 254 in dichromats, 174 Adequate stimuli, 19 After-images, 101, 218, 254, 272, 280 fading of, 110 Akyanoblepsia, 180 Allonomous equilibrium, 252 Amphibia, anatomy, 11, 15 colour vision of, 142 Amphioxus, 145 Anatomy, 7 comparative, 10, 142, 196 o Angstrom unit, 4 Animals, vision of, 131, 196, 203, 298 Annulus, 5 Anomalous trichromatic vision, 159, 182, 216, 235, 243, 266, 288 Apperception, 24 Approximate dichromatism, 235 Area, light, 119 rod-free, 10, 203 sensibility, 119 Areal effects, 51, 79, 95, 97, 117, 180, 250 in anomalous trichromats, 185 in dichromats, 180, 250 Autonomous equilibrium, 252. Bidwell's ghost, 86, 88 Binocular contrast, 129, 262, 276 Binocular struggle, 276 Birds, anatomy, 11, 15 colour vision of, 140 Black, 25, 36, 103, 201, 202, 251, 276 Bleaching of visual purple 13, 55, 206, 212 Blind spot, 8, 129 Border contrast, 127 Brightness, 28, 42, 96, 100, 223, 236, 294 specific, 256, 267 Brown, 25, 36 Cerebral centres, 277, 292 Charpentier's bands, 87 Chemical changes due to light, 12 Choroid, 8 Chromate test, 242 Chromatic scotopia, 53 Chromatic threshold, 19, 60 Coefficients, law of, 107 Coefficient of vai'iation, 149 Colour blindness, acquired, 160 facts, 158 in primitive races, 147, 148, 150 partial, 158 , theories, 193 total, 159 unilateral congenital, 180 Colour diagram, 37, 213, 216, 229 Colour equations, 38, 213, 226 Colour, extinction of, 61, 121 Colour ignorance, 161 Colour mixtures, 33, 162 Colour names, 160, 296 Colour sensations of dichromats, 161, 180 Colour triangle, >7, 213, 216, 229 Colour vision, evolution of, 130. 282. 285, 294, 298 of primitive races, 145 Colours, fundamental, 214, 220, 230, 258 invariable, 70 memory, 22 pigment, 6, 78 simple, 25, 27, 104, 220, 224, 251, 276 Comparative anatomy, 10, 142, 196 Comparative psychology, 131, 211, 298 Complementary after-image, 101 Complementary colours, 35, 101, 128, 226, 254 Complete fading, 261, 275 Conditional reflexes, 134 Cones, 9, 196, 203 Contrast, binocular, 129, 262 border, 127 luminosity, 127 simultaneous, 17, 125, 219, 255, 264, 279. successive, 101, 219, 254, 279 surface, 127 Critical frequency method, 96 Curve, equal-area, 228 luminosity, 42, 172, 189, 223, 240 minimal field, 72 minimal tinu;, 74 peripheral luminosity, 71, 173 persistency, 62 sensation, 232 valency, 215 302 INDEX OF SUBJECTS Ddmmertingssehen {see Scotopia), 17 Deuteranopic vision, 159, 232, 287 Deviation, standard, 149 Dichromatic vision, 158, 162, 216, 232, 267 adaptation in, 174 Dichromatism, approximate, 235 Dichromic, 294 Differential threshold or Kminal discrimina- tion value, 19, 30, 230, 233, 294 Diffraction spectrum, 3 Discrimination of hue, 30, 230, 233, 294 Dispersion spectrum, 3 Double-room, Hering's, 58, 74, 126 Duplicity theory, 198, 203 Electrical changes due to light, 15, 139, 206, 219 Energy of spcctrunij 28 Episcotister, 5, 92 Epithelium, pigment. 11, 145, 197, 211. 212 Equal-area curves, 228 Equality of brightness method. 44, 98, 107 Evolution of colour vision, 130, 282, 288, 294, 298 Extinction of colour, 61, 121 of Hght, 62. 121 Fading, complete, 261, 275 of after-images, 110 Fatigue, 112, 219, 231, 281 Feclmer-Helmholtz law, 107 Fechner's law, 20 Field of vision, 66 Fishes, anatomy, 11, 13 colour vision of, 143 Flicker method, 45, 55, 173, 191 Form sense, 204 Fovea centrafis, 9, 16. 81, 105, 117, 208 night blindness of, 50, 83, 208 Fraunhofer lines, 3, 29, 47 Fundamental colours, 214, 220, 230, 258 Gauging the spectrum, 39, 164, 215 General threshold or liminal value, 19, 60, 80, 117 Grassmann's laws, 33 Grey, 271, 282 mean, 251, 260 "Hemeralopia" (night-blindness), 49, 197. 209 Heptachromic. 294 Heredity, 159, 186 Hering's "double-room," 58. 74, 126 Hering's theory, 200, 251 Heterochromatic photometry, 43. 96 Hexachromic. 294 Hue, 27 discrimination of, 30, 230, 233, 294 Inadequate stimuli, 19 Induction. 17, 49, 101, 125, 218, 254, 279 temporal, 17, 49, 203, 255 Induction, simultaneous (Hering), 125, 127. 219, 255, 279 spatial, 17, 125, 219, 255, 264 successive, 101, 219, 254, 279 Infant vision, 131, 152 Inhibition. 93, 259, 275 Intensity, 28, 57, 78, 91, 95 Interference spectrum, 3 Intrinsic light, 104, 251, 276 ^ Invariable colours, 70 Law, Fechner-Helmholtz, 107 Fechner's, 20 Grassmann's, 33 MiiUer's, 18 Porter's, 96, 208 Talbot-Plateau, 92 Weber's, 19 of Coefficients, 107 of Specific energies of the senses, 18 Least perceptible differences, method of, 19 Lenticular pigmentation, 8, 41. 181 Light area, 119 chaos, 104, 251 extinction of, 62, 121 intrinsic. 104, 251, 276 sense, 204 Luminosity, 28, 42. 96, 100, 223. 236, 294 Luminosity contrast, 127 Luminosity curves of dichromats, 172, 240 of monochromats, 189 of trichromats, 42, 223 Macula lutea, 9, 81 Macular pigmentation, 41, 169, 181, 183. 241, 259 Mammals, anatomy, 11. 14 colour vision of, 132 Mean error, method of, 19 Mean grey, 251, 260 Mean variation, 149 Memory colours, 22 Method, constant, 19 critical frequency, 96 equality of brightness. 44, 98, 107 flicker, 45, 55, 173, 191 of least perceptible differences, 19 limiting, 19 of near error, 19 preference, 132, 137 of production, 19 of right and wTong answers, 19 visual acuity, 44, 99 Minimal field luminosity curve, 72 Minimal time luminosity curve, 74 Mixture of pure-colour stimuli. 33, 162 Monochromatic vision, 159, 186, 209. 217, 267. 288 MiiUer's law, 18 Negative after-image, 101 Neutral point, 163, 169, 207, 233 INDEX OF SUBJECTS 303 Night blindness, 49, 197, 209 of fovea, 50, 83, 208 Null-point, 216 Nystagmus, 187 Olive-green, 25, 36 Opponent colours theory, 200, 251 Optogram, 11, 13 Pentachromic. 294 Peripheral luminosity curve, 71, 173 Peripheral vision, 17, 50, 66, 123, 204, 208, 218 283 in diciiromats, 173, 176, 179 Persistency curve, 62 Phenomenon, flicker, 93 Purkinje, 57, 84, 96, 175, 177, 206, 210, 257 PhUology, 145 Photerythrous, 168 Photochemical theory, 289, 291 Photochromatic interval, 60. 80, 81, 120 Photometry, heterochromatic, 43, 96 Photopia, l7, 27, 203 Phototrophism, 130, 142 Pigmentation, lenticular, 8, 41, 181 macular, 41, 169, 181, 183, 241, 259 Pigment epithelium, 11, 145. 197, 211, 212 Pigments. 6, 78 Porter's law, 96, 208 Preference method, 132, 137 Primitive races, colour blindness in. 147, 148, 150 colour vision of, 131, 145 Protanopic vision, 159, 232, 287 Psychological process, 274 Psychologv, comparative, of colour vision, 131,*^ 211, 298 Psychology of colour vision, 17 Psycho-physical methods. 19 paraUeHsm, 21, 274, 290 Pupil reactions, 139, 141, 206 Purkinje's phenomenon, 57, 84 96, 206, 210, 257 in dichromats, 175, 177 reversed, 97 Purple, 27 visual, 12, 196, 205, 212, 220, 292 Rayleigh mixture. 182, 242 Reactions, pupil, 139, 141, 206 Recurrent vision, 85 Regional effects, 66, 95, 98, 117, 123, 258, 285 in dichromats, 177 Reptiles, anatomy, 11, 15 colour vision of, 142 Retina, anatomy of, 8 Retinitis pigmentosa, 209, 211 Rhodopsin, 12 (see Visual purple) Right and wrong answers, method of, 19 Rod -free area, 10, 203 Rods and cones, 9, 196, 203 Santonin, 211 "SateUite," 86 Saturation, 29. 36, 39, 59, 229. 250 Scoterythrous. 168 Scotoma, 103 central, 188, 267 Scotopia, 17, 49, 52, 203 achromatic, 14, 53 chromatic, .53 Sectors, rotating, 5,. 92 Sensation curves, 222 Sensation stimulator, 286 Sensibility area, 119 Shift of sensation curve, 243 Shortening of the spectrum, 168 Simple colours, 25, 27, 104, 220, 224, 251, ■ 276 Simultaneous contrast, 17, 125, 219, 255, 264, 279 induction (Hermg,) 125, 127, 219. 255, 279 Spatial induction, 17, 125, 219, 255, 264 Specific brightness of colours, 256, 267 "Specific energies'' of the senses, law of the, 18 Specific threshold or liminal value, 19, 60 Spectrophotometer, 5, 222 Spectrum, 1, 27, 49 diffraction, 3 dispersing, 3 energy of, 28 interference, 3 prismatic, 3 Spectrum, gauging the, 39, 164, 215 Standard deviation, 149 Stimuli, adequate and inadequate, 19 Stimulus receptor, 286 Structural changes due to light, 11 Successive contrast, 101, 219. 254, 279 induction, 101, 219, 254, 279 Surface contrast, 127 Synapses, 274 Talbot-Plateau law 92 Temporal effects, 85, 208 Temporal induction, 17, 49, 203, 255 Tetrachromic, 294 Theories, 193 Theory, 193 Bonders', 270 duplicity, 198, 203 Edridge-Green's, 291 Hermg' s, 200, 251 Ladd-Franklin's, 271 McDougaU's, 274 G. E. Mailer's, 290 opponent colours, 200, 251. physical, 198 Schenck's, 285 three-components, 199, 213 of two retinas, 197 Wundt's, 289 Young- Helmholtz, 199, 213 of zones, 200 304 INDEX OF SUBJECTS Threshold, chromatic, 19, 60 differential, 19, 30, 230, 233, 294 general, 19, 60, 80, 117 specific, 19, 60 Tone, 28 Total colour-blindness, 159, 186, 209, 217, 267, 288 Trichromatic vision, 38, 158, 213, 274 anomalous, 159, 182, 216, 235. 243, 266, 288 Trichromic, 294 Tritanopic vision, 159, 180, 233, 288 Uniocular struggle, 275, 277 Unit of sensation, 20, 61 Valency, 202, 253 curves, 215 white, 44, 202, 257, 264, 268 Variation, mean, 149 Vision of animals, 131, 196, 203, 298 of primitive races, 131, 145 anomalous trichromatic, 159, 182, 216, 235, 243, 266, 288 deuteranopic, 159, 232, 287 dichromatic, 158, 162, 216, 232, 267 Vision, monochromatic, 159, 186, 217 288 protanopic, 159, 232, 287 recurrent, 85 tritanopic, 159 180, 233, 288 trichromatic, 38, 158, 213, 274 Visual acuity. 204 method, 44, 99 Visual purple, 12, 196, 205, 212, 220, 292 absorption by. 13, 205 bleaching of, 13, 55, 206, 212 Visual substances. 12, 201, 219, 253, 270, 271, 280, 285, 291 Visual yellow, 13, 197, 220 Weber's law, 19 White, 4, 25, 201, 202, 207, 211, 223, 226, 283, 285 White valency, 44, 202, 257, 264, 268 Yellow, 115, 198, 260, 283, 285, 298 Yellow spot, 9, 81 {see Fovea centralis) Yellow, visual. 13, 197. 220 Young-Helmholtz theory, 199, 213 Zones, theory of, 200 INDEX OF AUTHORS Abady, 94 Abbott, 134 Abclsdorff, 12, 13, 140, 205 Abney, Sir WUliam, 2, 4, 5, 39, 40, 42, 43, 44, 45, 46, 47, 48, 53, 54, 56. 60, 61, 62, 63, 64, 65, 66, 72, 74. 75, 76, 77, 78, 79, 80, 81, 92, 94, 96, 106, 107, 116, 121, 122, 123, 126, 163, 173, 175, 180, 181, 182, 183, 190, 192, 219, 221, 222, 223, 224, 225, 227, 228, 229. 230, 231. 232, 236, 238, 239, 240, 243, 244, 247, 259, 294 Aircy, 94 Albert, 59 Allen, 94, 96, 98 Allen, Grant, 146 Angelucci. 11, 12, 15 Angier, 35, 36, 179 o Angstrom, 55 Arago, 50 Asher, 118, 119 Aubert, 4, 5, 29, 49, 60, 68, 118, 126, 127, 129, 200 Axenfeld, D., 129 Baird, 68, 70, 258 Baldwin, 152, 153 Bateson, 143 Basler, 204 Bauer, 145, 212 Bcale, 252 Becker, 180, 191, 288 Behr, 210 Bell, 98 Bellarminoff, 95 Bentley 143, 144 Benzaky, 146 Bernstein, 259 Berry, 197 Bidwell, Shelford, 86, 87, 88, 89, 94 Birch-Hirschfeld, 11 Bjerrura, 187, 189, 192 Bloom, 51, 208 Bois-Reymond, du, 15 Boll, 12, 13, 15, 196 Boltunow, 50 Bonnet, 18 Bosscha, 86, 89 Boswell, 66 Brandes, 127 Brauneck, 207 Breuer, 51, 183 Brewster, 129 Brodhun, 30, 59, 92, 173, 175, 207, 220, 233 Briicke, 43, 93, 112, 125, 126, 128 Briickner, 43, 45, 91, 235 Brunner, 259 Buffon, 44 Buhler, 105 Bull, Ole, 70 Burch, 3, 94, 102, 112, 113, 114, 115, 129, 219, 232, 294, 297 Burford, 133, 135 Cajal, Ramon y, 12, 212 Celsius, 44 ChaUis, 198 Charpentier, 45, 50, 51. 6]. (iS. 69. 79, 81, 86, 87, 89, 90, 95, 117, 120, 129, 197, 198 Chauveau, 129 Chevreul, 38, 125, 126 Clerk-MaxweU, 4, 21, 29, 40, 115, 126, 200, 220, 221 Cohn, 181, 191 Colbum, 192 Cole, 134, 142 Collin, 181 Colvin, 133, 135 Crova, 45 Dalton, John, 158, 202, 232 Davis, 134 Davis, A. S., 86 Dewar, 16 Dieterici, 30, 31, 35, 39, 182. 190, 191, 222, 223 Dimmer, 9, 10 Dittler, 12, 84, 90, 105, 108 Dobrowolski, 30 Donders, 50, 69, 120, 159, 163, 168, 181, 182, 183, 185, 191, 200, 221, 232, 268, 270, 271, 273 Dor, 191 Dove, 58, 126 Dow, 94, 98, 208 Draper, 42, 60 Dreher, 70 Drobisch, 198 Ebbinghaus, 36, 60, 127, 128, 129, 207 Edridge-Green, 12, 13, 29, 30, 109. 115, 116, 280, 291, 292, 294, 295, 296, 297, 29S, 299 Eisenmeier, 90 Engelmann, 11, 12, 16 Ewald, 268 Exner, F., 30. 86, 90. 94, 95, 212, 221, 223, 230 306 INDEX OF AUTHORS Fechner, 20, 21, 42, 52, 94, 103, 107, 109, 110, 112, 126, 127, 128, 129, 255, 270, 274, 294 Feilchenfeld, 59, 120 Ferry, 94, 96 Festing, 42. 45, 53, 54, 61, 62, 64 Fick, Ad., 51, 81, 92, 110, 127, 200. 202, 218, 289 Filehne, 95 Fleming, 45 Franz, 145, 203 Frey, v., 35, 41, 124 Frisch, v., 145 Fritsch, 10 Fuchs, 14, 16 Fujita. 123, 124, 208 Fukala, 188, 192 Galezowski, 191 Garbini, 152 Gardner, E. A., 146 Garten, 11, 12, 13, 51, 208 Gates, Elmer, 132 Geiger, 145, 146, 147 Gertz, 187, 188, 189, 192 Gladstone, 145, 146 Glan, 35, 40 Goethe, 21, 126, 128, 158, 180, 200, 251 Golant, 4 Gorham, 45 Gotch. 15, 16, 82, 219 Graber, 132. 143 Grailich, 58, 198 Grassmann, 29. 33, 43, 200 Gray, 60 Greeff, 9, 10, 11. 203 Greenwood, 36, 38, 182, 216. 264 Griinbaum, 0., 92, 94, 95 Grunert, 187, 188, 189, 192 Guillery, 50, 80 GuUstrand, 10 Giirber, 110 Guttmann, 183, 184, 185 Haab, 13, 197 Haas, de, 16 Halben, 183, 184 Hamaker, 86, 90 Hartley, 198 Hartridge. 198 Haycraft, 45, 55, 56. 94, 96, 97 Hayes, 180 Hegg, 70 Heine, 10 Heinrichsdorff, 210 Helmholtz, v., 1, 4, 5. 16, 20, 28. 29, 34. 35, 36, 37, 42, 43, 44, 52, 58, 92, 93, 103, 104, 107, 109, 110, 111. 112. 113. 120. 124, 126, 127. 128, 129, 158, 159, 165, 198, 200, 202, 206, 209, 212, 214, 215, 216, 218, 219, 220, 222, 232, 255, 274, 277 Henius, 123, 124, 208 Hering, H, 21, 22, 23, 24, 25, 41. 44, 53. 57. 58, 61, 68, 74, 84. 90, 101, 106, 107, 108, 110, 118, 124, 125, 126, 127, 128, 129, 159, 181. 182, 183, 188, 189. 190, 191, 192, 193, 197, 200, 202, 207, 251, 252, 253, 254, 255, 256, 257. 258, 259, 260, 263, 264, 266, 267, 269, 270, 273, 277, 279, 280, 290, 291 Hermann, 127, 129, 181 Hersehel, Sir Wm., 44 Hess. 11. 12, 15. 68, 70, 72. 74, 80, 86, 89, 110, 126, 127, 128, 129, 132, 134, 139. 140, 142, 143, 144, 145, 181, 187, 188, 189, 192, 202, 210, 211. 212, 258, 263, 264, 267, 298 Hessberg, 187, 189, 192 Hilbert. 102 HiUebrand, 53, 172, 189, 256, 257, 267 Himstedt, 16, 132, 206, 219 Hippel, V, 180, 187, 188, 191 Hirschberg, 120, 123 Hofler, 38 Holmgren, 16, 147, 148, 180, 181 d'Hombres-Firmas, 191 Homer. 145, 146 Huddart, 158, 191 Hyde, 92 Ives, 48. 94, 96, 97. 98, 99, 100, 208, 223 Javal, 146 Juler, 187, 188, 192 Jurin, 127 KaLischer, 135, 136 Katz, 22 Kinnaman, 133 Kinoshita, 90 Kirsch, 91, 235 Kirschmann, 128 KofEka, 185 Koike, 84 Ivollikcr 0 K6Uner,'l60, 181, 187, 192 Konig, 13, 30, 31, 32, 34, 35, 39. 41, 42, 45, 46, 47, 55, 57, 58 59, 60, 61, 64, 72, 81, 84, 100, 163, 164, 167, 172, 173, 174, 175, 180, 181. 182, 183, 187. 188. 189, 190, 191, 197, 202. 207, 208. 209. 220, 221, 222, 223, 224, 230, 232, 233, 234, 236, 267, 270, 280, 288, 295 Koranyi. 142 Koster, 9, 10, 61, 81, 84 Kottgen, 13, 205 Krarup, 35, 36, 40, 47, 55 Krause, 11, 129 Kressig, 191 Krienes, 181 Kries, v . 20. 21, 31, 34, 36, 37, 38, 39, 41, 43, 50, 51, 53, 58, 61, 70. 71, 81, 83, 84, 86, 89, 93, 94, 105, 106, 107, 108, 109, 124, 125, 159, 163, 164, 166, 167, 168, 169, 170, 171, 172, 173, 174, 176, 177 178, 179, 182, 183. 188, 189, 190, 191, 192, 198, 200, 202, 203. 207, 208, 209, 216, 217, 21 8, 219 259, 264, 268, 274 Kruss, 94 INDEX OF AUTHORS 307 Kiihne, 11. 12. 13. 14, 16, 196, 197, 205, 212, 268 Kuhnt, 9, 129 Kunkel, 90 Ladd-Fraiiklin, Christine, 50, 60, 188, 207, 260, 271, 272, 273, 280, 292 Lambert, 38 Landolt. 66, 68, 69, 187, 191 Leber, 202 Lehmann, 127 Leroy, 118 Levy, 179, 181, 183 Liebermann, 234 Liesegang, 197, 202 Lipps. 20 LleweUyn, 188 Lodato, 12 Loeb, 142 Loeser, 52, 120, 123, 208 Lohmann, 210 Lotze, 182, 183, 184, 194 Lubbock, 132 Luckiest, 98 Lummer, 60, 84, 92 MacDougaU, 21, 24. 85, 86, 87, 88, 89, 90, 91, 93^101, 102, 104, 111, 150, 151, 152, 153, 156, 193, 200, 208, 252, 260, 261, 263, 274, 275, 276, 277, 278, 279. 280, 281, 282, 297 Mach, 21, 119, 127, 200, 251, 270 Maclver, Randall-, 148 McKendrick, 16 Magnus, 146, 191 Maltzew, Fraulein v., 179, 184 Mandelstamm, 30 Marsden, 162, 153 Marshall, Devereux, 12, 115 Martius, 45, 90 Marx, 234 Mast, 130 MaxweU, Clerk-, 4, 21, 29, 40, 115, 126, 200, 220, 221 May, 188, 190 Mayer, 129 Merkel, 9 Messmer, 210 Meyer, H., 126, 127, 129 Millar, 94 Morgan, Lloyd, 140 Miiller, G. E. 16, 20, 129, 274. 277, 290 MiiUer, H., 9, 12, 196 Miiller, Johannes, 18, 126, 200, 221 Miiller, P., 90 Munk, 86 Myers, 19, 151, 152, 154, 156, 157, 195, 298 Nagel, 16, 31. 43, 53, 54, 60. 61, 66, 71, 83, 84, 106, 107, 132, 148, 163, 164, 166, 167, 173, 174, 176, 177, 178, 180, 181, 182, 183, 185, 187, 188, 189, 190, 191, 192, 206, 209, 212. 216, 219 Nahmacher, 12 Nettleship, 159, 186, 187, 191. 192, 209 Newton, Sir Isaac, 3. 33, 34. 42, 43, 58, 84, 198, 199, 200, 207 Nichols, 28 Nicolai, 133 Nutting, 100 Orbeli, 105, 134 Parinaud, 61, 81, 197, 198, 210 Parker, G. H., 142 Parsons, 8, 44, 52, 66, 117, 159. 187, 208, 209. 211 Pauli, 259 Pawlow, 134 Pearse, 142 Peirce, 30 Percival, 94 Pergens, 192, 288 Pertz, 51 Pfluger, 188, 192 Pheophilaktowa, 132 Piper, 11, 14, 52, 123, 124, 139, 181, 192, 206, 208, 288 Plateau, 91, 92, 93, 95, 103, 110, 255 Pole, 180 Polimanti. 45, 48, 72, 94. 173, 176, 177, 183 Porter, A. W., 109, 115, 116 Porter, J. P., 140 Porter, T. C, 94, 95, 96, 97, 191, 208 Pretori, 127, 128, 264 Preyer, 152, 175, 202. 221 Purkinje, 9, 43, 57. 58, 59, 67, 74, 84, 86, 96, 97, 98. 112, 129, 145, 175, 206, 207, 208, 210, 211, 257, 268, 293, 297 Putter, 203 Pyle, 103 Querenghi, 191 Raehlmann, 53, 187, 191 Rayleigh, Lord, 3, 159, 182, 186, 235, 242, 249, 259, 269, 295 Reighard, 144 Ricco, 52, 117, 118, 120 Richarz, 287 Richter, 108 Ritter, 42, 45, 172, 175, 190 Rivers, 45, 95, 146. 147, 148, 149, 150, 151, 168, 252, 253. 275, 298 Rollet, 127, 128, 279 Ronne. 187, 189, 192 Rood, 45, 94 Rose, 191, 267 Rosier, 191 Rouse, 140 Runge, 38 Ruppert, 204 Sachs, 40, 55. 128, 140, 169, 205, 264 Samojloff, 86, 90, 132 Schafhautl 45 Schatcrnikoff, 53, 54, 55. 191, 208 Schaum, 287 308 INDEX OF AUTHORS Schenck, 95, 179, 181, 218, 285, 288, 289, 297 Scholer, 191, 288 Schon, 109 Schopenhauer. 191 Schoute, 119,' 120 Schultze, Max, 9, 10, 11, 19(3, 197, 198, 203 Scina, Ragona, 12(3 Seebeck, 163, 170 Seguin, 110 Sewall, 205 Sherman, 84 Sherrington, 93, 94, 95, 218, 259, 275 Shinn, 152 Siebeck, 72, 73 SiemerUng, 288 Simmance, 94 Simon, 50, 51 Siven, 211 Smith, Miss E. M., 136, 138, 139 SneUen, 89 Spicer, Holmes, 192 Stargardt, 210 Stegmann, 60, 206 Steindler, 30, 31, 32, 230, 233, 234, 235, 295 Steiner, 16 Stilling, 180, 181 Stobbe, 259 Stort, van Genderen, 12, 15 Stout, 24 Swan, 90 Tait, 12, 13 Talbot, 16, 45, 91, 92, 93, 95 Thiirmel, 100 Tigerstedt, 4, 6 Titchener, 103 Tonn, 175, 207 Torelle, 142 Treitel, 51 Trendelenburg, 13, 14, 35, 36, 53, 55, 206 Trotter, 45 Tschermak, 49, 51, 53, 58, 61, 74, 81, 83, 84. 103. 106, 108, 125, 126, 128, 129, 174. 257. 263, 267, 268, 289 Tufts, 94. 99, 223 Turberville, 158 Tyndall, 60 Uhthoff, 30, 31. 32, 187, 188, 189, 191, 192. 209, 288 Valentine, 152, 155, 156, 157 Vaughan, 50 Verworn, 252 Vierordt, 42 Vinci, Leonardo da, 126, 251 Vintschgau, 181 Virchow, 146 Voeste, 107 Volkmann, 118 Wadsworth, 9 Waller, 16 Wanach, 116 Ward, 251, 276 Washburn, 111, 131, 134, 143, 144 Watson, J. B., 133 Watson. Prof. W., 4, 39, 40, 48, 94, 106, 173. 182, 183, 184, 190, 199, 222, 228, 236, 239, 240, 241, 242, 243, 244, 245, 247 Waugh, 134 Weber, E. H.. 19, 20, 42, 91 Weber, H. F., 60 Wehrli, 192 Weiss, 197 Welcker, 9 Welponer. 14 Wendt, 211 Weyde, van der, 163, 175, 179 Whisson. 158 Whiting, 94 Whitman, 94, 99 Wild, 94 Williams, 60 Wii-th, 107 Woinow, 202 Wolliflin, 84, 210 Wundt, 84, 107, 126, 127, 128, 181, 289 Yerkes, 133, 134, 142 Young, C. A., 85 Young, Thos., 158, 198, 199, 200, 202, 219, 220, 232, 271. 274, 276 Young- Helmholtz, 163, 199, 200, 213, 215, 216, 217, 220, 224. 230, 258, 263, 270, 273, 274, 285, 288 Zahn, 74 Zehender, 183 Zolotnitzki, 143, 144 Cambridge: printed by john clay, m.a. at the university press /I iTT -'Ui-(iii!'ii fiHliiiiiiii'iii^iliilllililpiiiliiiiiill^^ :l''.:ii;',rilll