Recent Developments in Vision Research (1966)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

zo Pupillary Reactions to Light The kind and approximate number of publications dealing with the different parameters of light stimuli are shown in Tables 2-8. An exten- sive, thoughtful review of some of this literature has been given by Schweitzer (1955,1956). The work can be divided roughly into experi- ments on the pupillary effects of (1) the intensity of the stimuli, (2) the time-characteristics of the stimuli (duration, wave form, frequency), (3) the area of the stimuli (spatial summation), (4) the retinal location of the stimuli (fovea, parafovea, or periphery), and (5) the color of the stimuli. Stimulus Intensity 1. The literature By far the largest group of publications contains data on stimulus intensity (Table 2, A-C). The early authors used daylight, or various flames (candle, gas, petroleum, amyl acetate or alcohol), as light sources. After about 1900, electric lamps became generally available, and, more recently, electronic flash sources. Infrared, ultraviolet, and even X-ray and radium sources were used occasionally. Light intensity was varied by means of different apertures or iris diaphragms, by adjusting the distance between the source and an aperture, by rheostats (for electric lamps), by filters (paper, parchment, opal glass, milkglass or frosted glass, and later neutral gray gelatine filters or neutral gray glass filters or wedges), by rotating Nichol prisms or by rotating crossed polaroid filters. The light was measured by recording the current across the lamp and/or the color temperature of the lamp, by photo-electric devices, by thermopile galvanometers, comparison photometers, or spectrophotometers. In the clinical work contained in Table 2A, light intensity usually was varied by relatively simple means, such as apertures, filters, or different settings of a rheostat, without particular control of area, retinal location, or color. In a few papers included in this group, the authors had merely used an instrument previously developed by someone else, without adding original contributions ^Silberkuhl, 1896; Tange, 1901; Groethuysen, 1921; Engelking, 1922; Mehrtens &Barkan, 1923; Gifford &Mayer, 1931; Bomer, 1933; Gasteiger, 1934; Lodato, 1934) or having made only minor modifications (Sander, 1929; Mazzuchoni, 1925; Borsotti, 1939; Loewenfeld, 1956).

21 TABLE 2 Stimulus Intensity A. Methods with Clinical Emphasis in Which Intensity Was Varied by Simple Means Year Author Year Author Year Author 1882 Schadow 1920 Landolt 1933 Bomer 1894 Schirmer 1921 Groethuysen 1933 Feldman 1896 Silberkuhl 1921 Kleefeld 1934 Bujadoux & Kofman 1899 Schafer 1922 Barkan 1934 Gasteiger 1899 Somme r 1922 Enge Iking 1934 Lodato 1901 Tange 1922-59 Lowenstein 1937 Bujadoux & Gourevitch 1902 Schirmer 1922 Kofman & Bujadoux 1903 Fuchs 1923 Mehrtens & Barkan 1937 Nayrac & Franchomme 1904 Bumke 1925 Mazzucconi 1939 Borsotti 1904 Piltz 1928 Lehrfeld 1940 Frydrychovicz & Harms 1905,06 Weiler 1929 Barbieri 1907 Hubner 1929 Modonesi 1949-54 Harms 1907,08 Krusius 1929 Nicolai 1956 Loewenfeld 1910 Sachs 1929 Sander 1957-61 Shakhnovitch 1910 Weiler 1930 Engel 1959-61 Samojloff, Sokolova, & Shakhnovitch 1911 Hembold 1931 Gifford & Mayer B. Experimental Work in Which Stimulus Intensity Was Varied in Connection with Other Problems Year Author Chief problem Year Author Chief problem 1892-93 Bordier visual acuity 1939-40 Hecht ft Pirenne color (owl) 1892,93 Sachs color 1942 Bart ley frequency, ocular discomfort 1900 Abelsdorff color 1903 Friberger speed 1942 Wagman 4 Gulberg color 1903 Schafer color 1948 Barany & Hallden retinal rivalry 1904 Abelsdorff & Feilchenfeld 1948 Morone dazzling, fatigue color 1952 Wirth retinal rivalry 190S Easier color 1956.59 Van der Tweel various 1907 Hess retinal position 1957 Bleichert servo-analysis 1907 Polimanti color 1957 Bleichert & Wagner se rvo - analy gis (frequency) 1908a Hess color (animals) 1909 Hesse retinal position 1959 Shakhnovitch color (cat) 1910 Hess color (animals) 1959-63 Stark el al servo -analysis 1923 Laurens color 1960,61 Clynes servo-analysis 1931 Zeldenruat chronaxia 1962 Alpert & Campbell color 1933 Machemer speed 1962 Bouma color 1933 Stiles & Crawford directional sensitivity 1963 Feinberg & Podolack latent period Note: a in Year column refers to order in References.

22 TABLE 2 (cont'd) C. Experimental Work Especially Concerned with Effects of Stimulus Intensity Difference Threshold Year Author Static Flash Threshold Darkness 1760 Lambert * 1881,82 v. Vmtschgau n * 1884 Gorham * 1 888 Chaveau r * 1888 Cohn * 1893 Du Bois-Reymond & Greeff * 1897 Garten * 1898,1905 Ovio * 1899,1900 Lans * 1900 Vervoort * 1907,13,14 Schlesinger n * * 1908 Hess r* * 1914,15,16 Hess * 1918 Blanchard * 1918 Reeves r * 1919 Engelking n 1920 Reeves * r * 1926 Holladay * 1927,32 Ferree & Rand * 1929,30 Stiles * 1932 Cradle & Ackerman * 1933 Ferree, Rand, & Harris * 1934 Biffis r * * 1934b Luckiesch & Moss * * 1936-37 Crawford * * 1937 Hartinger * 1938 Elsberg & Spotnitz * 1938,43 Kappauf * 1938 Talbot r* * 1939 Brown & Page * 1942 Wagman 8. Nathanson * 1943 Hartley * 1944 Moon & Spencer * 1947 Corrado * 1947 Venco & Marucci n 1948 Flamant * 1948 Spring & Stiles * * 1949 DeLaunay * 1952 De Groot & Gebhard * 1953 Alpern & Benson * 1953 Fry & Allen n * * * 1954 Fugate n * 1955,56 Schweitzer n * 1956 Fugate & Fry * n * * 1956 Hopkinson * 1956 Schweitzer & Bouman n * 1957 Kadlecova & Peleska * 1959 Alpern, Kitai, & Isaacson n * * 1959-60 Kawabata n * 1959,61 Lowenstein 4 Loewenfeld n * * 1960-63 Hakerem n * 1961 Kadlecova * * 1962a Alpern & Campbell * 1963 Burke n * 1963 Lowenstein, Kawabata, & Loewenfeld n * Note: Symbols: static = eye adapted to light stimuli; flash n - light pulse; flash |~* = sudden increment; threshold = threshold reactions of dark -adapted eye; difference threshold = increment or decrement of previously steady illumination; darkness = sudden or gradual withdrawal of light. a,b in Year column refers to order in References.

23 In Table 2B, a number of experiments are listed in which light in- tensity was considered secondarily, the chief subject of the investigation being concerned with other problems. Table 2C summarizes experimental work in which stimulus lumi- nance was of major interest. In these investigations, different kinds of light stimuli were used: (1) the subject's eye was successively adapted to different levels of brightness of the source, and the pupillary diameter was noted for each intensity step ("static" in Table 2C); (2) the pupillary reaction to a sudden light-flash was studied. In some cases, this flash was transient, of definite and controlled duration (symbol|~~| in Table 2C); in others the light was left on, and pupillary be- havior after the light stimulus was not considered (symbol ["""" in Table 2C). In most of these studies the subject's eye was said to be dark - adapted, but it is clear that the intended "darkness" was true darkness only when infrared-sensitive recording or viewing devices were available. Whenever visible light was used for observation, this form of stimulation actually represented a sudden increment of light above the adapted state; (3) some investigators tried to establish the minimal light increase needed to obtain a just-discernible response ("threshold" in Table 2C). The reservation concerning the quality of "darkness" holds for this kind of experiments also, and, in tests done under observation with visible light, it was actually the incremental threshold that was determined; (4) in the experiments marked "difference threshold" the authors intentionally looked for the smallest increment or decrement of light that would provoke a just-discernible reaction; (5) a final group of authors studied the time course of the pupillary dilation that occurs when light to which the eye has been adapted is sud- denly turned off. In the case of Flamant (1948), the withdrawal of light was gradual, since she took intermittent measurements while the eyes were exposed to the darkening sky at dusk; these measurements were, then, actually taken during gradually shifting, slow adaptation to dark- ness, and they could therefore be included with equal justification among the "static" group. Fry & Allen (1953) studied the recovery time of the pupil after exposures to light shorter than needed to adapt the eye. Because of these different methods of stimulation, a variety of results were found, some of them contradictory. A detailed considera- tion of these divergencies would lead to endless discussions. In the present paper, therefore, a brief description will be given of the general features of pupillary responses, recorded under various controlled con- ditions; and the discrepancies in the literature will be discussed in connec- tion with the pupillary phenomena which they concern. Since the differences between the light stimuli listed in Table 2C under "static, " "flash, " and "darkness" are essentially differences in timing, they are considered in the section devoted to that subject. The descriptions in this section are limited to reactions elicited by 1-second (sec) light flashes of different intensities.

24 2 . Features of the reactions to light of different intensities When the dark-adapted eye is exposed to light flashes of 1 sec duration, the pupillary threshold is found to be very low. Using appro- priate recording techniques (Lowenstein & Loewenfeld, 1958), small but distinct pupillary reactions usually can be obtained well within the first log unit of stimulus luminance above the subject's scotopic visual thresh- old. As the intensity of the light is increased over a range of approxi- mately 3 log units, the pupillary contractions become more constant and extensive. Throughout this low-intensity range of luminance the responses are, however, typically shallow: the contraction is preceded by a long latent period, and it is slow, inextensive, and of short duration (See Figs. 1, 2, 9, 10A, 11). When the light intensity is further increased the re- flexes begin to grow markedly in amplitude, speed and duration, and, depending on the autonomic nervous balance of the subject (cf. below), maximal values are reached about 7-9 log units above the scotopic visual threshold: the reactions outlast the 1-sec light flashes (approximate du- ration = 1.5-1 .8 sec); the latent period has decreased to a minimum (about 0.2-0.3 sec), while extent and peak speed of contraction have reached maximal values (about 4 millimeters (mm) and 7-10 mm/sec, respec- tively). It will be noted that the increments in pupillary reactivity for equal increments in stimulus luminance become greatly enhanced im- mediately above the low-intensity range. This sudden increase in effec- tiveness of the light flashes is due to the fact that the cone-threshold has been exceeded (See Figs. 18, 19). This scotopic-photopic break is typical for the pupillary increment curve when white light is used (See also Flamant, 1948; Kadlecova, 1960; Lowenstein & Loewenfeld, 1959a); it is absent when red light is used, because the pupillary threshold rises above the photopic visual threshold, and the lower part of the increment curve is missing (Fig. 2A, line of squares). It is also absent when white light is used in the pigeon's eye, with a predominantly cone-retina (Fig. 2, lines of triangles). Very powerful light flashes fail to add further to the amplitude and speed, or to reduce the latent period of the reactions, but they cause marked prolongation of the contraction: after such stimuli, the pupil may remain in spastic miosis for several seconds, and the following redilation is slow. (See section on "Time Characteristics of the Stimuli; Very strong light. ") 3. Modifying effects of fatigue and emotional excitement Fatigue and emotional excitement are so much a part of everyday life, and their modifying influence upon pupillary diameter and reactions so profound, that their effects must be well understood and constantly borne in mind in experiments that, typically, last considerable time. The light reflex, in spite of its autonomic nature, does not take place independently of the subject's level of consciousness. While the

25 6 IB 26 1646566676 86 96 L H Fig. 1 . Extent, duration and latent periods of pupillary reactions to one- second light flashes. Symbols represent average values of 30 individual reactions per intensity step. Subject's eyes dark- adapted. Extent of pupillary contractions (in mm, left-hand scale) and of durations of latent period and of total contraction time (in sec, right-hand scale) plotted against stimulus lumi- nance, with arrow T marking subject's scotopic visual thresh- old. Log luminance = .6 - 9.6 above visual threshold for 1-sec light flashes emitted by Sylvania glow modulator tube (with 31 mm condensing lens = slightly divergent beam measuring approxi- mately 20 mm at subject's left cornea). L_ = lowest and H = highest intensity of Grass photic stimulator, seen by both eyes, approximately 9 inches distant. Circles: Note the characteristic scotopic-photopic break in the rising curve of extensiveness of pupillary contractions (see also Fig. 2, 18, 19). Maximal reflex amplitude was reached at 8-9 log units above visual threshold. Squares: Duration of the pupillary reflexes remained short in the low-intensity range (see also Figs. 10, 11); at 3-4 log units above visual threshold, duration increased to a new plateau (about 1.6-1.8 sec). In response to very bright light, the contractions were much prolonged, because the pupillary re- dilation became slow and delayed. Triangles: Latent period for contraction was inversely related to stimulus luminance, falling from a maximum of more than 500 milliseconds (ms) near threshold to a minimum of about 240 ms at 8-9 log units above threshold. (Absolute time values here indicated are probably somewhat longer than true latent periods. For explanation, see section in text concerning latent period on p. 48).

26 4.0 "*- ,-©- T .<*J A f / 3.5 3. 2.» A (y ' / H / ,** ii.s I / / CONTRACT V 9 / / / S^/ / B E 0 O' / A D A 100 90. 80 70 60 50 £20 t- O 10 o »* 0. B x / * 7 -o' "10 '9 -8 "7 -6 LOG LUMINANCE —> Fig. 2. Extensiveness of pupillary contractions to light in normal human subject and in pigeon (Fig. 20, Lowenstein et al.. Amer. J. Ophthal. 57:569-596(1964). In A, actual extenfof pupillary con- tractions plotted as ordinate (in mm); in B, data re-plotted as per- centage contraction, taking maximal contractions for human sub- ject and for pigeon as 100 per cent, to correct for smaller size of pigeon's iris; each symbol represents average of at least 20 indivi- dual reflexes. Abscissa shows intensity of stimuli in terms of neu- tral grey filter transmittance, with 0 the maximal output of Sylvania glow modulator tube, used as in Fig. 1 . Stimulus duration = 1 sec; stimulus color = white (circles, triangles), or with Wratten #29 red filter (squares). Human scotopic visual thresh- old marked by white arrows; human visual threshold when using red filter marked by black arrows. Note the low threshold and distinct double slope of the human pupillary increment curve for white light (circles), the high threshold and single increment curve for the human reactions to red light (squares), and in the pigeon for white light (triangles). subject is alert, the central synapse of the pupillomotor reflex arc in the Westphal-Edinger nucleus is subject to supra-nuclear inhibitory influences. Simultaneously, hypothalamic discharges, brought into play by sensory or emotional stimuli provided by the environment, or, in man at least, by spontaneous thoughts or emotions, travel via the lower brain stem, cervical cord, and peripheral sympathetic chain to the dilator muscle of the iris (Figs. 3, 4). Under the influence of these mechanisms the pupil in healthy, alert subjects is relatively large and quiet in darkness (7-9 mm, Fig. 5, line Bl; cf. also Lowenstein, Feinberg, & Loewenfeld, 1963). In healthy, well-rested subjects this condition may be maintained for long periods of time. But when the subject becomes tired, the pupils

27 Fig. 3. Schematic representation of pupillary light reflex pathway (Fig. 1, Lowenstein, 1959) Note: a£ = aqueduct of Sylvius; cg = ciliary ganglion; cis = short ciliary nerves; 1g = lateral geniculate body; oc = optic chiasm; on = optic nerve; o^ = optic tract; p£ = posterior commissure; pr = pretectal area; we = (Westphal-Edinger) oculomotor nucleus; III = third nerve. gradually become smaller and begin to oscillate. In ever deepening waves of sudden, spontaneous arousal and gradual slipping into a doze, the pupils dilate rapidly, then re-contract gradually in an unsteady, wavering decline. The more the subject is tired, and the less he tries to suppress his sleepiness, the shorter the time of initial mydriasis, and the deeper and more frequent the following pupillary oscillations. Eventually, the spontaneous intervals of re-awakening cease altogether, and the subject actually falls asleep. At the moment immediately pre- ceding sleep, the pupils are quite small (Fig. 5, lines B2-B3). At this time a psycho-sensory stimulus such as a sudden sound, conversation, etc., restores the waking condition and, depending on the type and in- tensity of the stimulus, may maintain it for some time. It has been shown that the pupillary oscillations that appear in the tired subject originate in the central nervous system. As the subject drifts toward

28 1C Fig. 4. Schematic representation of pupillary pathways in sagittal view of brain (Fig. 2, Lowenstein, 1959). Solid lines: Efferent sympathetic path- from hypothalamus via cervical cord and peripheral sympathetic chain to eye. This efferent system is under control of cortico-thalamic-hypothal- amic mechanisms. Dash-dot line: Efferent parasympathetic path from Westphal- Edinger nucleus via third nerve and ciliary ganglion to the iris sphincter. Dotted lines: Inhibitory paths to the Westphal-Edinger nucleus: (1) direct afferent connections in brain stem reticular formation, and (2) descending connections from cortex, thalamus and hypothalamus. Note: a£ = anterior commissure; as_ = aqueduct of Sylvius; av = subclavian ansa of Vieussens; £ = cortex; cb_ = cliliospinal center of Budge; ££ = corpus callosum; cg = ciliary ganglion; cis = short ciliary nerves; cil = long ciliary nerves; f_= fornix; gg_ = Gasser- ian ganglion; gs = ganglion stellatum; ha = habenular nucleus; m = mammillary body; mcg = middle cervical ganglion; mi = massa intermedia; nc = naso-ciliary branch of the ophthalmic 5th nerve; oc = optic chiasm; on = optic nerve; p = pons; p_£ = posterior commissure; p_i = pineal body; scg = superior cervical ganglion; III = (Westphal-Edinger) oculomotor nucleus.

29 Bl B2 6 *- _a—, j^ Fig. 5. Spontaneous pupillary movements in tired subject (Fig. 7, Lowenstem_e_t aL., 1963). Diameter of right pupil recorded as ordinate (in mm) against time as abscissa (in 0.1 sec units in A, and in seconds in B). In lines B, ordinates reduced approxi- mately to 1/2 size shown in A. In line A, 1-sec light flashes were presented at times framed (intensity 8-9 log units above scotopic visual threshold). Lines B were recorded while the subject sat in complete darkness, fixating on a small, red light spot at about 6-foot (ft) distance. Line A: At beginning of the test the pupil showed normal dia- meter, and light flashes elicited normal reflexes. Line B: The pupil was large and remained relatively quiet during the first minute in darkness (Bl). Irregularities appeared soon, with small, fast oscillations prominent (_b,b in line B2, showing pupillary movements during the 8th minute in darkness); these sudden, small contractions and redilations, resembling mini- ature light reflexes, probably are associated with imperfect fixation of the tired subject (cf. Lowenstein e_t al., 1963, p. 142). Irregular, extensive waves of pupillary dilation and con- traction appeared shortly thereafter (a-<—«-a in line B3, recorded during 16th minute of observation); they accompanied waves of drowsiness and spontaneous arousal of the subject, and became more and more frequent and extensive until the subject finally fell asleep. Spontaneous lid closures, which did not affect the pupil in darkness, marked by £. sleep, supranuclear inhibition of the Westphal-Edinger nucleus decreases, and sympathetic activity is gradually lost. The consequent relative pre- ponderance of the parasympathetic outflow is revealed by the smallness of the pupil at the time immediately preceding sleep. At the moments of

30 spontaneous or reactive awakening, sympathetic activity and supranuclear inhibition of the third nerve nucleus cooperate in dilating the pupil. Chronically increased fatigue is not an uncommon finding, and, unless recognized, will cause difficulties in experiments in which the pupil is used as an indicator. Many persons habitually fail to sleep sufficiently, and the after-effects of various illnesses may be far more long-lasting than is recognized. Many subjects, though neurologically in good condition, will thus be found to be more subject to fatigue than is usual for persons in their age group, or is explained by energy spent during an experiment. When sitting quietly in darkness and not occupied by some activity, it is difficult for such subjects to stay awake. After a brief period of wake fulness with large, steady pupils, the oscillations described above will appear in an exaggerated manner: the pupils will fluctuate wildly over a large range as the subject fights an increasingly overwhelming sleepiness. Pupillary reflexes are superimposed upon this constantly shifting equilibrium of autonomic innervation of the iris, which is further modi- fied by humoral adrenergic mechanisms (Loewenfeld 1958, pp. 327-344) and by the mechanical limitations of the iris muscles (Fig. 6). A light stimulus of a given intensity and duration will therefore not necessarily elicit a pupillary reflex of predictable speed and amplitude. As the sub- ject or the experimental animal becomes sleepy, the pupil becomes smaller and the reactions less and less extensive; immediately preceding sleep, the miotic pupil hardly reacts to light (Fig. 7, e,f, g). Sensory or emotional stimulation have an opposite effect upon the pupillary di- ameter: with increasing excitement the pupil becomes larger and larger. The light reflex, however, does not benefit by the large pupillary diam- eter in darkness; it is supressed when the degree of supranuclear in- hibition and sympathetic excitation exceeds an optimal level (Fig. 7, b, £•!>• Figure 7 shows clearly that these changes in the pupillary reflexes are not merely a matter of amplitude. The shape of the reflexes varies with the degree of excitement or drowsiness. The characteristic square, w-shaped, v-shaped, and flattened, inextensive reflexes can be seen. The very same reaction patterns are observed in patients with lesions or irritation at various locations within the nervous network of pupillary control (compare the reactions shown in Fig. 7 with Fig. 8A, D1-D3,F, Gl, and G2), and they can be produced at will in experimental animals by electrical stimulation or surgical destruction, at the same sites. The only difference between the effects of physiological fatigue and excite- ment and those of pathological conditions is that physiological reflex changes are transitory and are changed to the normal pattern as soon as the subject or animal calms down, or awakens, as the case may be; in contrast, the pathological reaction shapes are permanent; they can deteriorate further but they cannot be restored to normal form.

The striking similarity between the reflex forms of fatigue and excitement on the one hand, and those of pathological conditions on the other, shows that the effects of emotional and sensory stimulation as well as those of fatigue are not diffuse and disorganized. While the sub- ject gradually falls asleep, specific nervous centers cease to function in orderly sequence, and they are called back into action by increasing psychosensory stimulation. Fig. 6. MECHANICAL LIMITATIONS 0 SYMPATHETIC EXCITATION PARAS YMPATHET 1C STIMULATION SUPRANUCLEAR INHIBITION t t t t t MECHANICAL LIMITATIONS Mechanisms affecting pupillary size and reactions (Fig. 5 of Lowenstein & Loewenfeld, 1964b). Parasympathetic innervation causes active pupillary constric- tion (black arrow), its central nervous inhibition incomplete, passive dilation of the pupil (cross-hatched arrow). Sym- pathetic excitation dilates the pupil rapidly and completely (white arrow). Adrenergic substances, entering the blood under the influence of central nervous mechanisms elicited by strong sensory or emotional stimulation, may reinforce and prolong pupillary dilation (dotted arrow). Finally, the limits of mechanical capacity of the iris muscles may modify the movements when extremes of mydriasis or of miosis are approached (small black arrows). These mechanisms are active to a remarkably similar degree in all mammals, and pupillary reactions elicited under similar experimental conditions show relatively minor variations among species (see Figures 7,8); in birds, with striated pupillary sphincter, the reactions are a great deal faster and show different time-amplitude patterns (see Figure 13).

32 Fig. 7. Effects of emotional excitement and of sleepiness upon pupil- lary light reflex in normal rats (Fig. 1, Lowenstein e_t al., 1963). Pupillary diameter recorded (in mm) against time (in 0.1 sec units). Because of smallness of rat's eye, ordinate enlarged by factor of 4, as compared to human pupillograms. Reactions of left eye shown. Animals in darkness except for 1-sec periods framed at_a - g when light flashes were presented. a = Normal light reflex in alert but not excited rat; b-d = inhibited reactions, elicited after sensory stimulation~with increasing emotional excitement, the pupil enlarged and the light reflexes became less and less extensive, showing characteristic w-and v-shapes found in all mammals under similar conditions; £-_£-_£_= light reflexes elicited while the animal was sleepy. Note the square, w-and flattened v- shapes of the responses. In any experiment concerning the pupillomotor effectiveness of a given kind of light stimulus, or light adaptation, these mechanisms may alter the reactions. Variations due to momentary changes in the physio- logical state of the subject are especially marked near the pupillary threshold, because the small pupillary reflexes elicited by weak light are easily suppressed by antagonistic influences (Fig. 9). Moreover, in a number of normal subjects the degree of tiredness or of emotional tense- ness may vary considerably, and the same experimental situation may elicit different results (see also Section below, "Other Features of Pupillary Reactions to Light, " and Fig. 25). It is thus necessary to be aware of such possible alterations in every experiment, and to safeguard the accuracy of the results by using a sufficiently large number of sub- jects, and by repeating the reactions a sufficient number of times.

33 TABLE 3 Pathology of the light reflex Condition Pupillary manifestations D3 Maximal central (diencepha- lic) irritation D2 Dl Submaximal central (diencepha- lic) irritation Parasympathetic (non- irritative) nuclear or post- nuclear lesion of the third nerve Lesion in the afferent path- ways of the light reflex 8 Very large pupils and absence of light reflex, or sluggish, in1 tensive reflex of short dura- tion; bilateral W- or V- shaped light reflex; pupils larger than normal; bilateral Prolonged latent period; pupils slightly larger than normal; sluggish, inextensive light reflex; unilateral or bilateral Prolonged latent period; slower than normal, W- or V- shaped light reflex; pupils remain equal, even in uni- lateral lesions A normal optimal reactions to light E Peripheral sympathetic lesion Central (diencephalic) condition Gl Central (diencephalic and G2 mesencephalic) condition H Parasympathetic (mesence- phalic) irritative condition Pupil smaller than normal; contraction speed slightly increased; second redilation phase absent; usually unilateral Shortened latent period; pupils smaller than normal; fast, abrupt, though less than normally extensive light re- flexes; bilateral Pupils small; W- or V-- shaped light reflex; bilateral Prolonged latent period; in- extensive, sluggish light re- flexes; unilateral or bilateral Note: Cf. Fig. 8.

34 Fig. 8. Dynamic structure of normal and pathological light reflexes (Fig. 4, Lowenstein, 1959). Eyes in darkness = a; during 1-sec intervals b, exposed to light stimuli about 8-9 log units above scotopic visual threshold. Heavy horizontal dash-dot line indicates diameter of normal pupil in darkness. In curves A-H, pupillary diameter recorded as ordinate (in mm) against time as abscissa (in 0.1 sec units). Curves A' - H' show speed of pupillary contraction and dilation occurring within each reflex. Curves obtained by plotting against time (in 0.1 sec units, abscissa), extent of contraction or of redilation occurring within each successive 1/10 sec (expressed in mm/sec, ordinate). Normal pupillary light reflex under experimental conditions described = A and A'; dotted lines B, C, Dl, D2, and D3 (and corresponding speed curves B' - D3') = abnormal reflexes in conditions causing larger than normal pupils; solid lines E, F, Gl, G2, and H (and corresponding speed curves E - H ) = ab- normal reflexes in conditions causing.smaller than normal pupils. A summary of pupillary reactions shown in this graph, and of causitive conditions, are contained in Table 3.

35 Fig. 9. Enhancement and suppression of low intensity light reflexes by central nervous inhibitory mechanisms (Fig. 3,Lowenstein & Loewenfeld, 1959). Normal subject (36-year-old man), tired on day of examination; eyes dark-adapted. Pupillary diameter recorded as ordinate (in mm) against time as abscissa (in 0.1 sec units), with solid line representing direct reflexes of right pupil, the broken line consensual reac- tions of left pupil. One-second white light stimuli, approximately 2 log units above subject's scotopic visual threshold, were pre- sented at times framed at L^- • First line: Reactions elicited while the subject gradually fell asleep showed enhanced contractions with no redilation in darkness; at c c_c_the subject closed his eyes. Second line: Shortly before 4th light reflex, the subject was awakened by a verbal stimulus ("wake up ! "). During awaken- ing, the pupil dilated and the 4th light reflex was suppressed. The 5th light flash elicited a normal threshold reaction.

36 TABLE 4 TIME CHARACTERISTICS OF STIMULI A. Simple control of duration by means of mechanical shutters, electric contacts, and electronic flash Year Author Means R Year Author Means K 1869 von Arlt lever, tuning fork * 1947 Thomson photo-flash 1881 v. Vintschgau lever * 1947 Thschirren photo- shutter * 1882 v. Vintschgau electric spark 1947 Venco & Marucci photo-shutter * 1903 Fuchs telegraph key * 1948 Barany & Hallden shutter 1904 Bumke shutter 1949-54 Harms photo-shutter 1904 Piltz air shutter * 1949-63 Lowenstein rotary disk shutter * 1905 Weiler lever-shutter * 1950-54 Cuppers shutter * 1907 Schlesinger falling shutter 1952 Wirth photo-shutter 1910 Weiler lever-shutter * 1954 Fugate shutter * 1913 Schlesinger shutter 1956 Fugate & Fry shutter * 1919 Enge Iking metronome * 1956-63 Loewenfeld electric flash tube * 1921, 39 Kleefeld photo-shutter 1956 Schweitzer & Boum an rotating disk shutter 1922-57 Lowenstein photo -shutter * 1923 Cradle & Eisendraht * metronome 1956 Petersen electronic flash * 1957 Drischel photo-shutter * 1929 Barbie ri electric contact 1957-63 Lowenstein electronic tube * 1932 Poursines falling cam shutter 1957-63 Shakhnovitch mechanical shutter * 1933 Walter falling cam shutter 1959 Alpern, Kitai, & Isaacson * 1939 Borsotti photo-shutter * falling shutter 1940 Fryd rychovitz & Harms 1959 Samojloff mechanical shutter * photo-shutter 1963 Burke electronic tube * 1943-63 Morone et al photo-shutter * 1959-63 Hakerem electronic tube * 1947-63 Lowenstein electric magnetic shutter * R = duration of stimulus was registered B. Study of effects of duration, wave form, or frequency of stimuli Year Author n R | O Year Author I) R | O 1760 Lambert * 1955,56 Schweitzer * 1801 Himly * 1956 Fugate & Fry * * 1845 Listing * 1956 Hopkinson * 1868 Hensen & Volckers * 1956,59 Van der Tweel * * 1882 Schadow * 1957 Becker * 1888 Chaveau * 1957 Bleichert * 1903 Friberger * 1957 Bleichert & Wagner * 1904 Piltz * 1957 Stark & Sherman • 1913,14 Schlesinger * 1957 Stegemann * 1918,20 Reeves * 1957 Drischel * 1922-65 Lowenstein * | * 1958 Stark & Campbell * 1923 Laurens * 1958 Stark & Cornsweet * 1926 Santamaria * 1959 Kawabata 1 * 1929 Nicolai * 1959a,b Lowenstein & Loewenfeld * * * 1932 Lythgoe * 1959 Samojloff * 1935 Machemer * * 1959 Stark * 1938 Talbot * 1959 Stark & Baker » 1939 Borsotti + * 1960,61 Clynes * * *d 1942 Bartley * 1961 Stark, Redhead, & Payne *d 1942 Wagman & Gulberg * I962b Alpern & Campbell * 1944 Stem * 1962 Hakerem * * 1947 Venco & Marucci t *d 1962a Stark * * * *d 1948 Morone + 1962b Stark * 1950 Campbell & Whiteside # 1963 Feinberg & Podolak * * 1950-54 Cuppers * * 1963 Lowenstein, Kawabata, & * * 1952 Wybar * Loewenfeld 1953 Fry & Allen * | * 1963 Loewenfeld * 1954 Young & Biersdorf * 1963 Lowenstein & Loewenfeld * 1954-55 Du Bois-Poulsen & Loisillier * 1963a,b Stark * 1963 Troelstra * Note: Symbols: D - stimulus duration (temporal summation); R = repeated stimuli (trains of intermittent flashes); S - sinusoidal stimulation; O = other time characteristics; d = darkness reflex (cf. text). a & b in Year column refer to order in References.

37 Time Characteristics of the Stimuli 1. The literature Only relatively few authors have done special experiments on the influence of duration, wave form, or frequency of the light stimuli upon pupillary reflexes. In earlier experiments, and in most clinical work in which stimulus duration and frequency were considered at all, they were controlled by simple devices (photographic shutters, or shutters that used a swinging pendulum, swinging cam, rotating disk, or other me- chanical device). More recently, electronic flash tubes allow more variable and more accurate control of stimulus duration and frequency (Table 4A). Experiments specifically concerned with the time characteristics of light stimuli are summarized in Table 4B. They were designed to study the effects of (1) stimulus duration (temporal summation), (2) stimulus frequency (intermittent light flashes), (3) stimulus wave-form (sinusoidal stimulation), and (4) other time characteristics of pupillary reactions related to light. Under this heading are grouped experiments on (a) pupillary movements during steady illumination, and (b) pupillary reactions elicited by short interruptions of steady light ("darkness reflex"). 2 . Influence of stimulus duration (a) Moderate intensity range: It has been mentioned above that the duration of a pupillary contraction to a timed light flash depends on the intensity of the stimulus. The effect of changes in stimulus duration, likewise, is different for dim and for bright light. In the low-intensity range, the short-lasting pupillary contraction usually is followed by redilation at about the same time when either short or longer stimuli are used (Fig. 10A), and even when the light is left on continuously (Fig. 11). With brighter light, the contraction movement is sustained by continued stimulation, so that the reaction to a long, bright stimulus is much more extensive than the one elicited by a short, bright stimulus (Fig. 10B). The general statement frequently seen in the literature (usually quoted after Reeves) that "the pupil reaches full contraction after 5 seconds" is thus incorrect, and a measurement of the pupil 5 sec or more after the onset of stimulation will not necessarily reveal the effectiveness of the stimulus. If the light is weak, the transient contrac- tion will have disappeared long before. For this reason, authors who investigated the effectiveness of light by the "static" method generally tended to report much higher thresholds for the pupillary reflex than those who recorded the dynamic responses to shorter light flashes.

38 t o./sec. Fig. 10. Reflexes to short and to longer light stimuli at low and at high levels of stimulus intensity (Fig. 4, Lowenstein & Loewenfeld, 1959). Pupillary diameter recorded as ordinate (in mm) against time as abscissa (in 0.1 sec units). Broken lines show the pupillary responses to 0.1-sec light flashes (double-arrows _a); solid lines show reactions to 1.0-sec stimuli (double-arrows b). A (first line): Dim light was used, about 3 log units above sub- ject's scotopic visual threshold. Reactions to short (a) and to long (b) light flashes were alike; both showed long latent period, low peak speed, small extent, and short duration typical for low- intensity reactions. B (second line): Light intensity was increased to about 9 log units above visual threshold of the subject. The long, bright light stimulus (b) caused the pupillary contraction to continue for a longer time, and thus to become more extensive than the reflex elicited by the short, bright light flash (a), even though latent period and peak speed of the two reactions were alike: compared to the low-intensity reflexes of line A, the latent period was shortened, 'and the contraction speed increased.

e 7 s 4 3 ( 0we-* 'I i——i i i |Q '- >|^Q- Fig. 11 . Pupillary reactions to continuous light stimulation. Pupillary diameter recorded as ordinate (in mm) against time as abscissa (in seconds). Beginning of continuous light stimu- lation of subject's left eye (left pupil dilated by cyclogyl) marked by double arrow. Stimulus intensities: 2.6 log units above scotopic visual threshold for solid line, 4.6 log units for broken line, 6.6 log units for dash-dot line, and 8.6 log units for dotted line. Note the inconstancy of the pupillary contraction to dim light, and development of pupillary oscillations under influence of steady illumination. These oscillations were large, irregular, and inconstant when light of medium intensity was used; they became faster, less extensive, more uniform, and sustained with increasing stimulus luminance. (b) Threshold reactions: Among the experiments listed in Table 4B, only the ones by Fry & Allen (1953), Fugate & Fry (1956), Schweitzer (1955,1956), Kawabata, 2 and Hakerem^ can be said to have considered the question of temporal summation systematically. As a general result, it appears that close to the pupillary threshold the duration of the light flash becomes important: for durations up to approxi- mately 1 sec, the pupil responds to the total energy of the flash, so that intensity and duration are interchangeable in this near-threshold brightness range. (c) Very strong light: It has been shown in Fig. 1 that the pupillary contractions to intense light stimuli are prolonged. Even a very short, exceedingly bright flash of light will cause the pupil to constrict maxi- mally, and to remain in spastic contraction for a number of seconds. The most probable explanation of this behavior is the assumption that the after-image elicited by such powerful light flashes is sufficiently intense 2 Personal communications, (H. Kawabata) 1959-1961. 3 Personal communications, (G. Hakerem) 1959-1963.

40 to cause a continued pupillary response. There can be no doubt that after- images are capable of affecting the pupil, as shown by an experiment by Alpern & Campbell (1962a). When the eye has been adapted to very bright light, and this light is suddenly turned off, the pupil, after a short delay, begins to enlarge; soon thereafter, however, the dilation is interrupted by a new contraction which may last many seconds before giving way to resumed dilation as the dark-adaptation process continues. In a similar experiment done in the Lowenstein Laboratory of Pupillography (Columbia University), it seemed as though the contraction period coincided with the appearance of a strong positive after-image * . The chapter of pos- sibly related pupillary and after-image phenomena is an interesting one, and should be worked out under better conditions of control than those used by Alpern & Campbell and by the author and colleagues (control of stimulus area and color, simultaneous recording of pupillary move- ments and of the subject's statements). (d) Stimulus duration in clinical examinations of the pupil: Numerous authors have realized the importance of standardizing the stimulus duration in clinical examinations of the pupil, and have used various timing devices to control it (Table 4A). Among the investigators who recorded the pupillary reflexes, the question of the most suitable stimulus duration has been discussed (Cttppers, 1954; Drischel, 1957a; 1957b; 1957c; Petersen, 1956). Lowenstein, during the years 1922 to the present, has experimented extensively on the effects of different stimulus durations in clinical pupillographic examinations, and has chosen 1-sec light flashes as the kind of stimuli most likely to allow a clear-cut distinction to be made between the normal and different patho- logical reflex patterns. Reactions to shorter light flashes show these distinctive features to a lesser degree. In the intensity range usually employed for such tests, they result in a single, short contraction which is followed immediately by redilation. One-second light flashes, on the other hand, cause a continuation of the contraction; and it is during these * It should be -.stressed that such a dip in the pupillary redilation curve upon withdrawal of light is never found in alert subjects after adaptation to moderate stimulus luminances. When the eye has been adapted to light up to about 8 log units above the visual threshold, and this light is suddenly turned off, the pupil re-dilates quickly and in a smooth curve, reaching nearly full dilation within less than one minute. Crawford (1936-37) has reported a dip in pupillary size some 5 minutes after the beginning of dark adaptation, and has assumed a shifting from cone to rod function as a possible cause. It is, however, quite certain that, under the experimental conditions described by Crawford, there can be no retinal discharge powerful enough to affect the pupil at so late a time in the dark adaptation process. As has been described above, many subjects become quite drowsy after sitting quietly in darkness for a few minutes, and the pupillary contraction which accompanies fatigue is marked (cf. above. Chapter BI, 3).

41 later periods of the reaction that the characteristic features of escape from stimulation (W- and V-form), or the flat, square reaction shape (holding action) are prominently displayed (Figs. 7,8). It has been objected that prolongation of the stimulus beyond the latent period of pupillary contraction causes the light entering the eye to become progressively reduced, thus setting up a complex situation which was held responsible for the reflex changes described by Lowenstein. It should be realized, however, that in the intensity range used by Lowen- stein for such examinations (8-9 log units above scotopic visual thresh- old) such a reduction is insignificant, because the stimulus luminance may be decreased by a factor of 100-1000 without significant changes in the reflex shape. In any event, these reflex shapes were found unaltered when the patient's stimulated eye was atropinized (with the consensual pupillary reaction recorded), and when the light intensity was thus held constant throughout the 1-sec stimulation period. Some rather queer notions are sometimes encountered in the discussions about the best possible stimulus duration. Thus, Talbot (1938a) was of the opinion that only the light step, that is, the change to a higher level of brightness acts as stimulus, while continued contraction due to continued illumination constitutes an adapted state during which no energy is expended, while Drischel (1957a, 1957b, &1957c) is of the opinion that only short light stimuli elicit the "natural" features of the reflex, which would prompt one to conclude that the only "natural" illumination is provided by lightning at midnight. It is the author's opinion that in clinical examinations of the pupil standardization of stimulus duration, and accurate recording of the reflexes, are of greater importance than the particular duration used. 3. Influence of stimulus frequency The majority of investigations in which repeated light flashes were presented at a fairly slow rate were concerned with alterations of the reflex form and amplitude which develop under the influence of trains of stimuli. These changes are primarily dependent on central nervous re- flex mechanisms and are, therefore, omitted from this review. To the author's knowledge, Hartley (1943) was first to record pupillary oscillations to short, repeated light flashes presented in rapid succession. With the more recent availability of electronic flash sources and of convenient methods for continuous recording of pupillary move- ments, such studies have become more numerous (Du Bois-Poulsen & Loisillier, 1954-1955; Bleichert & Wagner, 1957; Stark & co-workers, 1957-1963; Kawabata,4 Lowenstein &Loewenfeld, 1957 to present; Clynes, 1960-61; Feinberg &Podolak5). 4 Personal communications, 1959-1961. 5 Personal communication, (R. Feinberg &G. Podolak) 1963.

42 When one or both eyes are exposed to short light flashes in rapid succession, the individual contractions elicited by individual stimuli summate- With increasing stimulus frequency, the summation becomes more pronounced: the mean pupillary diameter decreases, and the indi- vidual pupillary oscillations become smaller (Fig. 12). In mammals, the oscillations become very small at the relatively low frequency of 3/sec. Different peak frequencies have been reported by different authors (3-9/sec), and since the fast oscillations are exceedingly small, the top frequency will depend to some extent on the sensitivity of the recording instrument. The author and co-workers did not observe oscillations much above 4/sec. Hakerem , using an averaging computer, found slightly faster activity (5-6/sec). The conditions of experiments in which even higher rates were found were, in the author's opinion, not sufficiently controlled to rule out artefacts. The limiting factor for the fastest rate at which the pupil is able to follow such short light flashes is clearly imposed by the slowness of the smooth muscle effector. In birds, which have a striated pupillary sphincter, much higher frequencies can be recorded. In the experiment on a pigeon shown in Fig. 13, for example, 10 cycles per sec were fol- lowed easily, and with greater amplification higher frequencies undoubt- edly could be recorded. Stark and his co-workers have found similar high oscillation rates in the owl. As a clinical tool, the high-frequency reactions have been found by Lowenstein to be most useful. In cases with spastic miosis, just as in normal subjects after conjunctival instillation of physostigmine (Loewen- feld, 1963), the pupil is unable to oscillate normally, and the iris is driven into a tetanic contraction at lower than normal stimulation rates. In contrast, miosis due to sympathetic lesions or to decreased supra- nuclear inhibition (see p. 24) does not interfere markedly with the fast pupillary oscillations. 4. Sinusoidal stimuli Some authors used, instead of intermittent light flashes, sinusoidal alternations of brighter and dimmer light, or of light of two colors. Such stimulus wave-forms were produced by changing the current to the lamp, or by inserting a rotating Nichol prism, rotating disk shutter, or rotating crossed polaroid filters into the light path. The majority of these studies were chiefly concerned with problems of servo-analysis that fall outside the scope of this review. Findings of peak speed of the responses agree fairly well with those obtained by intermittent light flashes. 5. Other time characteristics of pupillary reactions to light (a) Pupillary movements which occur during steady illumination: When one or both eyes are exposed to a constant light stimulus, the pupils ^Personal communication, 1959-1963.

43 o. i see. Fig. 12. Pupillary oscillations elicited by repeated short light flashes in man (Fig. 6, Loewenfeld, 1963). Pupillary diameter recorded (in mm) against time as abscissa (in 0.1 sec units), with solid line representing right pupil, broken line, left pupil. Moments of presentation of 5 ms white light flashes at rate of 1 per sec (A), 2 per sec (B), and 3 per sec (C) marked by small arrows. Light source, Sylvania glow modulator tube; light intensity = about 9 log units above scotopic visual threshold; area = about 5°, centrally fixated, with rest of retina illuminated by intraocular stray light. In mammals, the slow smooth-muscle pupillary sphincter is driven into tetanic contraction at low stimulus frequencies. Mean pupillary diameter becomes smaller with increasing flash rates, and individual pupillary oscillations less and less extensive.

44 7 f Imm 1 s \ / \ / \/ / •\ , •\, ^ ^ / f^ 3^ ^v/ vi^\ ^ i ^ / ^ x-t-\' " "> ^f * \/ V •* 3 V VI I/ s •V j (, a/ I . — = $S f" r r" • - .* r •* r" P i? -k " " - --- •- -w — • - ^ I.. — H -> . t. - - - >- '-j * - - . - _^- -^-- \^v • *r~- -N_^- '-^^- *v ^*J "._r- --^-' **-^v s*^f <-*• ^= =c -4 — ' r--^ *• 1 L^^, fc 7-. f « V °t' S /^ •V/ i — — f ' ' ' - r r '1 1 "" "~ I . ~ U u I : M Li j L J il ., Fig. 13. Pupillary oscillations in pigeon elicited by repeated short light flashes Moments of presentation of 5 ms white light flashes (same as in Fig. 12) marked by up-strokes of event marker pen; the event marker was run via an electronic relay, actuated by the current to the Sylvania glow modulator tube. The relay circuit had a uniform holding action, independent of the duration of the pulse to the Sylvania tube, i.e., the light flashes were shorter than indicated by the event marker. Because of the striated iris sphincter, the pigeon's pupil moves much faster than that of .mammals, with stimulation rates of 10 per sec followed easily by discrete oscillations. contract, then redilate partially and begin to oscillate. These movements were observed by many of the early authors, and various interesting attempts were made to measure their rate and amplitude (Table 4B). At a later time, these oscillators played a considerable role in the clinical literature under the names of "pupillary unrest" or, if more marked, "hippus. " Their presence or absence was supposed to indicate various pathological conditions.

45 Several theories were proposed to explain the mechanism of these movements. Most frequently they were thought to accompany the pulse and/or the respiratory cycle, or to be elicited by accommodation, or by psychological influences. Simultaneous recording of pulse, respiration, and pupillary move- ments failed, however, to show any phase-coincidence among these functions and the pupillary oscillations. Accommodation can be ruled out by adequate fixation upon a far target, and psychological influences, just as pulse and respiration, should operate in darkness as well as in light, while the pupil, in the absence of tiredness, is large and almost immobile in darkness. A final explanation holds the shading effect of the iris responsible. Lambert (1760) first noted that the image of a candle flame, slowly moved toward the pupil from the corneo-scleral limbus, at first did not affect the pupil markedly; but as soon as the tip of the image entered the pupil, a sudden contraction occurred and was followed by oscillations. In 1944, Stern advocated this experiment as a clinical test, whereby the edge of the pupil was grazed by a small point of light, and the pupillary oscilla- tions were counted over a period of time. Later workers reported differ- ent oscillation frequencies for normal and for pathological cases. (Campbell & Whiteside, 1950; Wybar, 1952; Stark & co-workers, 1957- 1963). In Fig. 11 such oscillations are shown. They were elicited by light from a Sylvania glow modulator tube, used with a 31 mm condensing lens to form a slightly divergent beam measuring about 20 mm at the level of the subject's cornea (9 inches distant). Intensity was varied by neutral grey filters. It is obvious that extent and rate of the pupillary oscillations differ with different stimulus intensities With weak light the pupillary contraction was followed almost immediately by redilation. The pupil regained its dark-adapted diameter within seconds, and showed no more oscillations than it did in darkness (solid line). In response to brighter light, the first contraction became more extensive and was followed by large, irregular oscillations. After some time, these oscillations be- came quite shallow while the mean diameter of the pupil enlarged (broken line). When the stimulus was further increased in intensity, the small oscillations became faster and more regular. They were maintained throughout the 2-minute exposure time, even though the mean pupillary diameter enlarged (dash-dot line). Upon stimulation by bright light, the pupil contracted extensively; it remained spastically contracted for a number of seconds, and the small oscillations appeared only gradually, as the eye adapted to light. The initial, maximal constriction was main- tained throughout the 2-minute adaptation period (dotted line). In Fig- 14, the results of a similar experiment are summarized. Pupillary oscillations were recorded for 2-minute adaptation periods over a 9-log unit intensity range. As in the experiment of Fig. 11, only

46 I20 t .6 I.6 26 3.6 46 5.6 66 76 8.6 96 .25 .20 f CO F C 0 .I5 o o o r— .I0 ^ 0 .051 C 0) •7C C 1C QC 0 x 0) Fig. 14. Extent and frequency of pupillary oscillations in light Pupillary oscillations recorded from normal subject's right, consensually reacting pupil, with pupil of left, stimulated eye immobilized by cyclogyl. Abscissa showing light intensity in log units above subject's scotopic visual threshold (marked by arrow); left-hand ordinate and solid columns representing number of pupillary oscillations within 2-minute periods of steady illumination (averaged num- bers of 3 experiments per intensity step). Right-hand ordinate and shaded columns showing average extent of same oscillations (in mm). In dim light the small oscillations were no more frequent or extensive than in equivalent periods in darkness (not shown); they became larger and more numerous with increasing stimulus luminance, until the light intensity exceeded about 8-9 log units above threshold, when they became'less extensive and the average frequency within the first 2-minute periods of light dropped somewhat, because very bright light prevents pupillary oscillations in the early period of stimulation (see Fig. 11, dotted line).

47 a few, inextensive pupillary oscillations were found when the stimulus light was weak; in the medium intensity range the movements became more numerous and extensive; finally, during exposure to bright light they became fast, fairly regular, and smaller. The decrease in frequency shown in Fig. 14 for the highest intensity step is only apparent; it is caused by the fact that the pupil does not oscillate at all during the early, spastic phase of the response (see Fig. 11, dotted line). It is, in the opinion of the author, unlikely that the shading effect of the iris plays a major role in the mechanism of these oscillations, at least in experiments like those just described, in which a fairly large beam of light covered the entire iris. In the experiment of Fig. 14, the stimulated eye had been immobilized by cyclogyl, and the movements were recorded from the consensually reacting opposite pupil. This ex- periment has been repeated in the same subjects by stimulating either the eye with the normally reacting pupil or the eye with the immobilized pupil, and no difference has been found in the rate and amplitude of the oscillations under these two experimental conditions. These facts do not rule out the possibility that, with a smaller stimulating spot and more elaborate frequency analysis, using a computer and longer stimulation periods, such differences might be found. It should be mentioned, however, that changes in stimulus luminance due to shading by the iris, giving rise to such marked pupillary movements, should be expected to be visible; but in the experience of the author it was precisely in the medium-intensity range, in which the light looked entirely unchanging to the subject, that the pupillary oscillations were most pronounced, while with weak light, which appeared to come and go throughout the 2-minute period, the pupil was stable. It must be concluded that the mechanism of these movements is yet unexplained. It is possible that the analysis of pupil oscillations under steady illumination may become a useful clinical tool. It is certain, how- ever, that records of frequency and amplitude of such movements without careful control of stimulus intensity are meaningless, both in normal and in pathological cases. (b) The darkness reflex (symbol d in column O of Table 4B): When one or both eyes have been adapted to light, and when this light is inter- rupted by a short period of darkness, an interesting tri-phasic pupillary response occurs: during the dark period, the pupil dilates; after the light is re-admitted, it contracts beyond the previous light-adapted diameter, then re-dilates to the pre-stimulatory baseline. This reaction, first ob- served by Lowenstein (1939; Lowenstein &Givner, 1943), has been named by him the "darkness reflex." It is not identical to the pupillary redila- tion that follows a short light flash. The conditions differ insofar as, in the case of re-dilation, the eye is adapted to darkness, and the stimulus constitutes a short change to a higher level of illumination, while in the case of the darkness reflex, the eye is adapted to light, and the stimulus is provided by a short change to a lower level of energy.

48 Clinically, Lowenstein found that the dilation in darkness, the contraction upon re-admittance of light and the final re-dilation to base- line could be altered separately or in certain combinations, depending on the location of the pathological condition within the nervous reflex path- ways. These changes are not considered here. When the adapting light is dim, or when the dark interval is shorter than about 0.3 sec, the first dilation in darkness is missing, while con- traction and re-dilation are relatively well preserved. Kawabata, ? recording the pupillogram and electro-retinogram simultaneously, found that the dark stimulus elicits a retinal "off"- response, re-admittance of light, an "on"-response. The "on"-wave resembles the usual retinal reactions to stimulation by light, and is much larger than the "off-wave. If it is assumed that these electrical action potentials parallel the pupillomotor events, it would be easy to understand why the dilation period of the reflex is missing when the adapting light is dim, or when the dark interval is short. In the first case, no effective "off"-discharge would be generated; in the second, the pupillary dilation elicited by the smaller "off "-re act ion would be overcome by the contrac- tion elicited by the more powerful "on"-discharge. The "on"-phase of the darkness reflex is of particular interest because it is undoubtedly due to the fact that the retina, during the short dark interval, gains sufficient sensitivity to respond to the re-appearing, previously adapted light level as to an increment in lumi- nance. Stark, Redhead, and Payne (1961) have determined the interrela- tion between the duration of the dark interval and the amount by which the re-appearing light may be reduced in order to cause only a small, stand- ardized pupillary contraction, and have found a gain of about 1 -5 log units in the sensitivity of the pupillary receptor system within the first 2 sec in darkness. This experiment points to a most fascinating field of appli- cation, because it appears possible in this manner to test objectively the adaptive effects of very short dark periods. (c) The latent period: The question of the latent period of the pupillary contraction to light is one of the most promising unfinished chapters in pupillary physiology. Much effort and ingeniousness were spent by some of the early authors in attempts to measure the latent period (see Listing, 1845; von Arlt, 1869; Chaveau, 1888); but without recording apparatus, the values obtained could not be accurate. Later data, based on experiments with various recording techniques, ranged from 0.17 to more than 0.5 sec. Part of these discrepancies were, of course, based on measuring errors; it should be remembered that not long ago even reasonably accurate recording of the movements of this delicate diaphragm were exceedingly difficult to do. An additional source of discrepancies consisted in the fact, realized only later, that the pupil may, at the moment of light stimulation, happen to be in the 7 Personal communication, 1959-61.

49 descending part of a spontaneous contraction. Under such circumstances, the latent period may seem very short. Thus Lowenstein, in early papers, stated that the latent period may drop to as low as 0.06 sec. He has since corrected this error many times in various publications, but, as often happens in such cases, it remained one of his most faithfully quoted statements. A final, more important cause of the disagreements is the fact that the latent period for contraction to light varies with stimulus intensity. From a minimum of about 0.2 sec for very bright light, it may grow to about 0.5 sec near the pupillary threshold. The exact duration of the longest possible latent period in response to dim light has not been established adequately. Data obtained by the author and co-workers were based on averages of reactions which were recorded with a strip chart ink writer. The threshold reflexes of the pupil are so shallow that it is difficult to mark the exact onset of contraction and it is therefore almost certain that the figures thus obtained are slightly higher than the true durations. Hakerem is at present conducting experiments on this problem, using an averaging computer. As to the shortest possible latent period, the limiting factor is undoubtedly the slowness of the smooth muscle pupillary sphincter. It is not likely that the minimal latent period in mammals will be found to be much shorter than 0.2 sec, because the iris responds after a delay of approximately 0.15-0.2 sec to strong electrical stimulation of the third nerve or the ciliary ganglion (cat, monkey). In birds, with a striated sphincter, the minimal latent period is only about 0.06 sec, proving that, for bright light stimuli, the time delay consumed by the nervous reflex mechanism must be very short indeed. The most interesting and potentially useful feature of the latent period is its variability. It is affected by changes in the sensory as well as in the motor parts of the reflex arc. On the one hand, according to Lowenstein it is prolonged in cases with lesions in the midbrain or effer- ent parasympathetic nerve path; it can be altered by the use of drugs, and current work by Feinberg &Podolak indicates a possible age-trend. On the other hand, the inverse relation between the latent period and stimulus intensity makes it appear possible to use the latent period as a measure of the effectiveness of light stimuli. As shown in Fig. 15, individual differences in the reactions to a given stimulus situation are much less marked for the latent period than for the amplitude of the same reactions. The prolongation of the latent period with dim light does not appear to be related primarily to the properties of the effector muscle, since it is almost as pronounced in the pigeon as it is in man (Fig. 16). Stimulus Area (Spatial Summation) In clinical examinations of the pupil, the extent of the stimulus area usually has not been considered, and most of the experiments of the

50 C..W - / ' A / 600. t B 1.5- / 500- A 0/P t A 1.0" O D ' o 1 / /; 0 Z 0 H 0 2 400 SQ S 0.5. / ° / Q. L D' / Z B— — 43 ~-<ft z JL P ^D,'''-*/ 111 o «300- • 0 I I E L0G LUMINANCE Fig. 15. Extensiveness and latent periods of pupillary reactions to foveal red light flashes in three normal subjects (Fig. 18, Lowenstein, _et^£l_., 1964a). In A, extents of pupillary contractions plotted as ordinate (in mm), and in B, latent periods of same reflexes (in ms), with each symbol representing average of from 8 to 20 reflexes. Stimulus luminance in log units shown by abscissa, where O indicates subject's foveal visual threshold for 1° white light flashes of 1 sec duration; visual threshold for red stimuli (Wratten filter #29) of same area and location marked by arrow. Three normal subjects were used: (1) a twenty-year old woman with unusually reactive pupils (circles); (2) a twenty-year old man with average pupillary reactivity (squares); (3) a twenty- four year old slightly tense man with somewhat more than average supranuclear inhibition (triangles). The visual thresholds of the three subjects differed less than 0.1 log unit. Pupillary contractions of subject (1) were up to five times more extensive than those of subject (2), and up to 10 times more than those of subject (3). In contrast, the latent periods of the three subjects agreed well. interaction of stimulus area and intensity as regards their pupillomotor effectiveness were done by authors primarily interested in problems of physiological optics (Table 5). The following are the most important results of these experiments:

51 250- 200- T |50 o o 100- 50 OX -5 -4 -S -2 -I 0 LOG FILTER T RAN SM ITTAN CE Fig. 16. Latent period of pupillary contractions to light of different intensities in pigeon (Fig. 21, Lowenstein, etal., 1964a) Average latent periods of pigeon's light reflexes (extents shown in Fig. 2) plotted as ordinate (in ms) against stimulus intensity as abscissa (in terms of transmittance of neutral grey filters used, with O, the maximal available intensity); each symbol represents average of at least 20 reflexes. Note the short latent period of the pigeon's striated pupillary sphincter muscle when bright light was used, and the length- ening of the latent period with decreasing light intensity (see also Fig. 15). (1) when a small stimulus field is enlarged, the pupillary thresh- old is lowered, and (2) when an already large field is further increased in extent, a further improvement of the responses is found (Abelsdorff &Feilchenfeld, 1904a; Burke, 1963; Ferree, Rand, & Harris, 1933; Hakerem, 8 Schweitzer, 1955-1956); (3) spatial summation appears to be more pronounced for the pupil than for visual perception: while the pupillary threshold is distinctly higher than the visual threshold for small test patches, the two thresh- olds become nearly alike when very large stimulus areas are used (Schweitzer, 1955, 1956; Hakerem, 1959-1963; Burke, 1963); and it is common experience that in average room illumination the pupil contracts 8 Personal communication, G. Hakerem, 1957-1963.

52 TABLE 5 Area and Retinal Position of Stimuli - Experimental Location Clinical 1 Clinical 1 \ rj :. Year Author -' : Year Author -- ~ 1760 Lambert • * 1834 Luckiesh & Moss * * 1900 Vervoort • • : - WJ i? Crawford * • 1904 Abelsdorff & - » IIM Eluberg & Spotnitz • Feilchenfeld \" n Talbot » • 1905 Easier = « 1938 Brown & Pajp • 1907 Hess * 1940 Frydrychovicz & Harms * * 1908 Hess * 1943 Bart ley " 1909 Hesse * 1946 Berens i Zuckerman • • 1909,13 Behr 1947 Corrado * 1910 Sachs 1947 Venco & Marucci * 1913 Srhlesmger * 1948 Flam ant * UH Jess 1949 De Launay • 1914 Schleainger • 1949-54 Harms * * 1914 Utbrich 1950 Sautter * - • 1914 Walker 1953 Fry & Allen • • 1918 Weve 1954 Fugate • 1919 Weve • 1954-55 Du Boia-Poulaen & • • 1922 Vogt * Loisillier !''Jn Holladay • • 1955,56 Sk-hweitzer * • ;''l'T Ferree & Rand I • 1956 Fugatf & Fry * 1929 Barbierl 1956 Schweitzer 4 Bourn an * 1931.32 Braun 1956,59 Van der Tweel * |" U Schmeltzer 1962 Alpern & Campbell . * 1932 Ferree & Hand • * I0CI Burke * • 1933 Ferree, Rand, & Harris • • ]''I.,! Hakerem " 1934 Blffis • * • ' 1963 Lowenstein, Kawabata. & Loewenfeld * * • when one eye of a subject, previously shaded from the light, is suddenly uncovered so that the light can be seen with both eyes. Under these conditions, there is no sensation of increased brightness. Retinal Location of Stimuli 1. The literature In contrast to the lack of interest among clinicians in the extent of the retinal area stimulated, retinal location has been the center of much attention, and a number of instruments were designed to demonstrate "Wernicke's hemianopic pupillary immobility" (Wernicke, 1873; Wilbrand, 1881). Some of these instruments were less complex than the titles of their descriptive articles might lead one to believe, consisting essentially of a light source and simple means for directing a small beam of light into the eye from different directions (lens,pinhole, shade, etc.; Heddaeus, 1893; Fragstein &Kempner', 1899; Kempner, 1899; Wolff, 1900; Stoewer, 1903; Bach, 1904; Bartels, 1904; Friedlander &Kempner, 1904; Veraguth, 1905; Walker, 1914). More elaborate clinical instru- ments and experiments are listed in Table 5.

53 The question of the pupillomotor effectiveness of light striking different areas of the retina often circled around the additional question of the identity of the pupillary receptors. The fact that brighter light was needed to elicit a vigorous response from the retinal periphery than from the fovea led to the conclusion that most of the pupillomotor recep- tors are confined to a relatively small, central retinal area (3-4 mm diameter, according to Hess, 1907). This statement was later interpreted as meaning that the cones alone function as receptors for the pupil, even though it is clear that Hess intended no such implication. The discussion, once opened, flourished in the usual manner, and even today, doubts or a flat denial of a pupillomotor function on the part of the rods are often expressed. It is interesting that among those who have experimented on the subject, these doubts are apparently less common (Table 7). Con- sideration of this problem will be postponed until the experimental facts pertaining to it have been presented. TABLE 6 Opinions Expressed about Pupillary Receptors: Rods versus Cones Year Author R C Year Author R C Year Author R C 1900 Abelsdorff t t 1934 v. Studnitz * * 1957 Bleichert 7 * 1904 Abelsdorff & Feilchenfeld * * 1936-37 Steinhardt * * 1959 Alpern, Kitai, & Isaacson * * 1939 Brown & Page - * 1905 Basler * * 1939-40 Hecht & Pirenne (owl) * 1959 Lowenstein & * * 1907 Hess * * Loewenfeld l908b Hess * * 1942 Wagman & Gulberg *• * 1962 Hake re m * * 1919 Enge Iking * * 1962 Campbell & * * 1923 Laurens * * 1948 Flamant * * Alpern 1927 Schlesinger 9 *' 1949-54 Harms - * 1963 Burke * * 1929 Barbieri * * 1949 DeLaunay - * 1963 Lowenstein, Kawabata, & Loewenfeld * * 1933 Ferree, Rand, & Harris * * 1953 Fry & Allen * * 1955,56 Schweitzer * * 1934 Crawford * * 1956 Fugate & Fry * * Note: R = rods; C = cones. * = pupillomotor function stated; - = pupillomotor function denied; ? = author was undecided. b in Year column refers to order in References. 2. Pupillary movements elicited from the fovea and from the retinal periphery When the eye has been dark-adapted, it is not difficult to obtain pupillary reactions from the retinal periphery. The threshold of these responses is low, even when relatively small stimulus patches are used, and it parallels the visual threshold in the different retinal areas. When light of a color other than red is used, the central scotoma of the dark- adapted eye can be demonstrated, as shown in Fig. 17. This figure also shows that the pupillary threshold reactions cannot be the result of stray light, because the energy needed to obtain contractions by stimuli placed on the blind spot is very much greater than that needed to stimulate adjacent areas of the retina.

54 TABLE 7 Experiments on Color Sensitivity Year Author Method of Stimulation and Chief Experimental Result 1892 Sachs color-color or color-grey substitution; light reflected from colored papers; moved new color over adapted color; pupillomotor effectiveness directly related to apparent brightness of color 1893 Sachs color- color substitution (glass filters in movable frame; intensity varied by grey filters); result same as 1892 1900a Abelsdorff color-color substitution (colored filters in movable frame); studied effect of color adaptation in man and animals 1900b Abelsdorff color- color or color-^rty substitution (prism in modified Helmholtz-Knnig color mixing apparatus; intensity varied by Nichol prism); apparent brightness parallel to pupillomotor effectiveness; found Purkinje shift after light-adaptation 1903 Schafer color- grey substitution with spectral colors and color mixtures (Helmholtz color mixing apparatus); 1905 Basler color-color or color-grey substitution (gelatine-glass-liquid filters in movable frame); studied effect of color adaptation; found red more effective in fovea, blue in periphery 1907 Hess adaptation to spectral colors (Nerst lamp and prism); studied adaptation and spectral sensitivity in day and night birds 1907 Polimanti color-grey substitution (mirror and spectroscope in double light path); pupillomotcr effectiveness of spectral colors parallel to apparent brightness 1908a,b Hess spectral colors (Nerst lamp & prism); also colored masks of various shapes; day birds more sensi- similar to night birds and to normal, dark -adapted man for dim Light; fovea in normal man relatively insensitive to blue light 1913,14 Schlesinger intermittent stimulation in different pans of visual field (color filters); selective extinction of reac- tions to one color need not affect those to a different color 1914,15 Hess as 1908-10; also difference thresholds with color and grey filters and wedges; extensive work on normal and color blind man, many animals; results similar to 1908-10; night birds have no Purkinje shift 1922 Enge Iking color-color substitution (color wedges in Hess' pupilloscope); determined color sensitivity in a color-blind patient 1923 Laurens adaptation to equal energy spectrum (Hilger spectrometer, brightness calibrated by distance from tungsten source) , determined pupillary effectiveness of colored light in dark-adapted and light- adapted man, pigeon, alligator; established Purkinje shift 1929 Barbie ri color filters in pupil-perimeter; determination of clinical color fields 1939-40 Hecht & Pirenne step stimuli, large field (tungsten with Wratten filters); determined minimal light of different colors visibility curve 1942 Wagman & Gulberg step stimuli, large field (tungsten with Wratten filters); determined minimal light of different wave lengths needed to obtain 1/2 mm contraction in dark-adapted human eye; curve matched dim light visibility curve 1955,56 Schweitzer intermittent, small teat flashes at various retinal sites (tungsten, red or green interference filters); also larger fields, ring-shaped fields, etc.; extensive work on pupillary threshold in dark-adapted eye; found central scotoma with green but not with red light; color sensitivity parallel to scotopic visibility 1959a,b Lowenstein & Loewenfeld intermittent stimuli (glow modulator tube, red, green, blue Wratten filters); determined pupillary thresholds in dark-adapted and light-adapted eye 1959 Shakhnovilch red-blue substitution (fitters); found reaction to color change in cat, said to be independent of luminosity 1962b Alpern tt Campbell color- grey sinusoidal alternation; threshold pupil change as criterion (Hilger-Watts monochromator, rotating polaroid filter in double light path); found mixed cone -rod spectral sensitivity curve for 2° foveal target, which was shifted to purely photopic curve when rods were suppressed by blue background field 1962 Bouma adaptation to equal energy monochromatic spectrum in photopic brightness range; 4-5 mm pupil con- traction as criterion; pupillary spectral sensitivity curve followed scotopic visibility curve 1963 Lowenstein, Kawabata, & Loewenfeld intermittent stimulation with small test patches in different retinal areas (red, green, blue Wratten filters); compared pupillary threshold reactions and increment curves to visual thresholds and flicker fusion curves Note: a,b in Year column refers to order in References.

55 /relB. •\ —© pup sensitivity ' vis. sensitivity X 550 nyi Exp. tl'LXI. Distance from fovea 20' 15* Moiolly IS* 20* Temporally Fig. 17. Pupillomotor and visual sensitivity to green light in fovea and retinal periphery for dark-adapted eye. (Schweitzer, 1956) Log relative effectiveness of 0.25-sec test flashes of 2° area (tungsten light with rotating shutter and interference filter) represented by ordinate; retinal positions indicated on abscissa. Visual thresholds (solid line) and pupillary thresholds (broken line) are parallel, with pupil about 1.5 log units less sensitive in this experiment (pupillary reactions detected by observation with infrared-sensitive converter-detector; the values for the pupil tend to become somewhat more sensitive when the reflexes are recorded). Note the central depression for green light, shown in the pupillary reactions as well as for the visual threshold, and the lack of sensitivity at the blind spot. The question arises as to why so many careful investigators have failed to observe these responses. There are two reasons for the diver- gent results, namely (a) the differences in the experimental procedure used, and (b) the differences in criteria as to what constitutes pupillary responsiveness.

56 (a) Influence of experimental procedure: As already described, the pupillary reflexes elicited by dim light are inextensive; they are easily suppressed by antagonistic influences, and are short-lasting even in the presence of continued illumination. When the stimuli are confined to the retinal periphery, the reactions grow only slightly in extent and speed as the light is increased above threshold values (Fig. 18, line of dots). Equal increments in stimulus luminance above threshold cause much greater increments in pupillary activity when the light strikes the fovea (Fig. 18, line of crosses). Finally, the small peripheral reflexes are suppressed more profoundly by adaptation to light than are reflexes elicited by brighter light at the fovea. LUMINANCE Fig. 18. Increments of pupillary reactivity with increasing light intensity at fovea and in retinal periphery (green light). Average extent of pupillary contractions plotted (in mm, 12 re- actions per intensity step) against stimulus luminance (in arbi- trary log units, with 100 per cent maximal intensity available in experiment). Subject's eyes dark-adapted; stimulus area = 1°; stimulus dura- tion = 1 sec; stimulus position = fovea (0°, crosses), or 15° in temporal retina, horizontally from fovea (dots). Visual thresh- olds marked by arrows, with shaded arrow indicating foveal, solid arrow peripheral threshold. With increasing stimulus intensity the pupillary contraction extents rose in a steep slope when the fovea was stimulated (solid line). The reactions elicited 15° peripherally showed a lower threshold than in the fovea, but only small increments to a low plateau, and a second rise when the stimulus luminance exceeded the cone threshold (broken line). The maximal reactions elicited peripherally were less extensive than those obtained at the fovea.

57 A number of the experiments that resulted in a denial of pupillomotor activity in the retinal periphery were done by the "static" method, that is, the pupillary diameter was measured after the eye was adapted to light of different luminances, and it is clear why no reactions could be found: they had already run their course, and the pupil had returned to its dark- adapted diameter at the time the observation was made. It should be added that the majority of these experiments were done with imperfect dark adaptation. Even "dim" illumination of a subject's eye must obviously be above cone threshold if it is used to render the examined pupil clearly visible to the experimenter (plus the intensity needed to overcome the loss of energy proportional to the square root of the distance between the sub- ject's and the observer's eye). It is, therefore, reasonable to assume that the small, fleeting, peripheral reflexes were merely suppressed, or escaped observation be- cause they occurred earlier than was expected. Once a doubt is expressed, however, it is surprising how quickly a "small" response may become "negligible, " and thence "absent"; and once it is considered absent, there is a tendency to explain it away, should it yet be observed. In the case of the peripheral retinal responses, the contractions, when seen, were thought to be due to stray light scattered by the ocular media, whereby it was not explained why stray light, which must by necessity be less in- tense than the focused beam, should have been capable of eliciting pupil- lary reflexes when the stimulus itself was not. (b) Criteria of pupillary responsiveness: Since the retinal periphery is less efficient than the fovea for the production of extensive pupillary contractions, it appeared as though the pupillomotor representation of the fovea far outweighed that of the periphery, whenever the reflex ampli- tude was used as a measure of reactivity. Despite their low amplitude, the peripheral reactions are, however, more sensitive than the foveal ones, that is, their threshold is lower. The sensitivity of a retinal element, expressed by the threshold, is thus a quality quite separate from its motor effectiveness, as expressed by the increments in reflex ampli- tude for light intensities above threshold. Pupillary reactions are not unique in this respect. For example, when the visual flicker fusion curves are compared with pupillary incre- ment curves, obtained with the same kind and intensity of stimuli, the low threshold and shallow increment curves for the retinal periphery, and the high threshold and steep increment curves for the fovea are very similar for the two functions (Fig. 18, 19). (c) Pupil pe rime try: A number of investigators have made efforts to develop pupil perimeters that would allow the exploration of the visual field by small, well-outlined light patches. There are certain technical difficulties to overcome, but it is hoped that such instrumentation will become practicable, and may allow the objective scanning of the visual field in man and animals, thus opening a most interesting field of future investigation, both clinical and experimental.

58 30 •o25- o S 20 &I5 0) W •510 e? u: 5 •!• '3 "l I00% Fig. 19. Increments of critical flicker fusion frequency with increasing light intensity at fovea and in retinal periphery (green light). Same subject and same apparatus as in Fig. 18, with subject's eye dark-adapted. Flash duration held constant at 10 ms. Values plotted indicate the lowest rate at which, for each intensity step, 1-sec trains of light flashes were consistently seen as "steady" (not "flickering"). Crosses and solid line = foveal stimulation: dots and broken line = stimuli 15° in temporal retina. Note the parallelism between the flicker fusion increment curves and the pupillary reflex increment curves of Fig. 18. Stimulus Color 1. The literature A fascinating group of experiments were concerned with the in- fluence of the wave length of stimulus light upon the pupil (Table 7). It was the purpose of most of these investigations to develop an objective method of testing spectral sensitivity in normal man, color-blind man, and various animals (Table 8). In the earliest experiments by Sachs (1892), the color of the light was changed by reflecting it from colored papers; soon these were re- placed by transmission filters (gelatine, glass, liquid solutions of dyes, colored wedges), or by spectroscopic prisms or color gratings. It is unfortunate that some of the most interesting earlier work (see especially Hess) was done under conditions that cannot be repeated today, because methods of accurate measurement of light intensity and wave length had

59 not yet been developed. Especially in view of the limited technical possibilities of the time, the ingeniousness and fruitfulness of this work is admirable. TABLE 8 Experiments on Animals Cephalopodes: sepia, eledone Hess (1910, 14, 15)*0 Teleosts: eels, other fish v. Studnitz (1934)=0 Amphibia: frog v. Studnitz (1934)=0; Lowenstein & Loewenfeld (1950- 60) +=0 Reptiles: alligator Laurens (1923)* 0 turtle v. Studnitz (1934)=0 lizard v. Studnitz (1934)*=0 Birds: pigeon Abelsdorff (1900a)*0; Hess (1907, 08a, 10, 14b, 15)*0; Laurens (1923)*0; Gundlach (1934)+; Lowenstein & Loewenfeld (1945, 1963)+* 0= chicken Hess (1907,08a)0*; Lowenstein & Loewenfeld (1950)=+0 magpie, falcon, buzzard Hess (1908a)0* owl Abelsdorff (1900a)*0; Hess (1907, 08a, 10, 14b, 15)* 0; van der Plank (1934)*0; Hecht & Pirenne (1939)*0; Stark e_t al (1960)= Mammals: mouse Keeler (1927)* 0 rat Lowenstein & Loewenfeld (1950-55)0= + guinea pig Abelsdorff (1900a)0*; v. Studnitz (1934)0*; Lowenstein & Loewenfeld (1950-55)0+= rabbit Abelsdorff (1900a)0*; Hess (1915)0*; Wagman & Nathanson (1942)0; Lowenstein & Loewenfeld (1950- 63)0+= cat Hess (1915)0*; Gundlach (1934)+; Kappauf (1938, 43)0; Shakhnovitch (1959)*; Lowenstein & Loewenfeld (1943- 60)0+= dog Hess (1915)0*; Lowenstein & Loewenfeld (1942-50)0+= monkey Hess (1915)0*; Lowenstein & Loewenfeld (1942-60)0+= Note: Symbols represent functions tested: * = color sensitivity; 0 = light sensitivity; + = speed of movement; = = other.

60 A method widely used was that of color substitution: after the eye had been adapted to light of one color, the color was changed suddenly, and the luminance of the second color was varied until the sudden change evoked no response, or a small contraction of standardized amplitude. In this manner, the relative physiological effectiveness of the two colors was revealed by the energy needed for each of them to become "pupil- lomotor equivalent." One color alone could be balanced similarly against neutral grey filters. The method of switching filters consisted most often in mounting them in a movable frame, their edges adjoining, and by flip- ping the frame back and forth across the stimulus light path. Sudden ro- tation of a prism was also used for this purpose. Recently, Alpern & Campbell (1962b) used a rotating polaroid filter for slow substitution of lights from a double path. In other experiments, colored light flashes were presented, and the amount of energy needed to cause a pupillary response of predetermined extent (threshold, or higher) was measured, or the amplitude of pupillary reactions that resulted from stimulation by light of various colors but equal energy content was recorded. 2. Pupillary spectral sensitivity A summary of the results of these experiments follows: The pupillomotor effectiveness of a colored light stimulus is related to its apparent brightness; for each color, the threshold for pupillary re- actions lies slightly above the corresponding visual threshold. This is true for all areas of the retina, and in the dark-adapted as well as the light-adapted eye (Fig. 20). In other words, the Purkinje shift exists for the pupil as well as for visual sensation (Fig. 21). In the dark-adapted eye, the pupillary spectral sensitivity curve for large stimulus areas agrees well with the human dim-light visibility curve, and with the spectral sensitivity curve for night-birds, with predominantly rod retinae (see especially Hess, 1908a,1910, 1914b, 1915 ; Hecht & Pirenne, 1939-1940; Wagman & Gulberg, 1942; Fig. 22). In the dark- adapted eye, Alpern & Campbell (1962b), using the color substitution technique with fairly bright stimuli, found a mixed cone-rod spectral sensitivity curve, even upon foveal stimulation with small test patches. The authors concluded that the rod-contribution must have been produced by stray light to extra-foveal regions, and, indeed, when the experiment was repeated in the presence of a blue background field, the pupillary spectral sensitivity curve for the foveal stimuli agreed completely with the CIE photopic visibility curve, and with the flicker fusion spectral sensitivity curve obtained from the same subjects under the same experi- mental conditions (Fig- 23). In a recent experiment by Boum a (1962), stimuli in the photopic brightness range caused pupillary reactions with a scotopic spectral sensitivity curve. Since Bouma used fairly extensive pupillary contrac- tions as criteria of reactivity, the explanation proposed by Alpern &

61 0 I 2 LOG LUMINANCE -> Fig. 20. Visual and pupillary thresholds in different areas of retina Experiments done after complete dark-adaptation. Subject's left eye stimulated by light flashes of 1° area and 1 sec dura- tion. Stimuli presented at fovea (first line), and 15° or 40° from fovea in nasal field (temporal retina, horizontal meridian, second and third lines, respectively). In each line pupillary thresholds indicated by P, visual thresholds by V; light intensity indicated on abscissa in arbitrary log units, with 4 the maximal intensity available in the experiment. Source = Syivania glow modulator tube, with flashes elicited by Grass stimulator via constant voltage power supply. Light reflected by circular white cardboard targets placed 98 cm from subject's eye. Wratten filters as follows: red = # 29 (solid arrows); green = # 99 (cross-hatched arrows); blue = # 45 + 47 (white arrows); intensity regulated by neutral grey filters. The #29 Wratten filter is not a pure red; it was chosen nevertheless because it does not overlap with the green filter. Note the close agreement between pupillary and visual thresh- olds for all colors and for all retinal locations examined. Campbell (1962b), namely, that the results were likely to be due to the influence of stray light, appears probable. In Schweitzer's work (1955, 1956), and in the experiments in the author's laboratory with threshold stimuli, stray light probably played no significant role, since the light

62 l00 90 80 t70 S60 £50 $30 • 20 o l0 DIM UGHT, dark-adopted.-* BRIGHT UGHT light-adapted 400 WAVELENGTH 500 in mj 600 700 Fig. 21. Purkinje shift in pupillary spectral sensitivity (Laurens, 1923) Curves constructed from photographic records of pupil size, obtained after 15 minutes of adaptation to darkness or to light of different wave lengths (equal energy stimuli, using tungsten source and Hilger wave length spectrometer). "2 o •£-3- o O O PUPIL: OWL (Hecht and Pirenne) MAN (Wagman) \ ihuman scotopic visibility curve 400 500 600 WAVELENGTH in mp -»• 700 Fig. 22. Pupillary spectral sensitivity curves (re -plotted from Hecht Pirenne, 1939-40; and Wagman & Gulberg, 1942) In both experiments curves constructed by measuring energy of light of different wave lengths needed to obtain pupillary threshold reactions (0.3-0.5 mm), with sudden exposure of eye to light, and with relatively large field illuminated (tungsten source and Wratten color filters).

63 20 1.6 1.2 S o w o. «> 0.8 e o -" 0.4 15* blue background field • = Pupil, with 2* foveal target —— m CIE photopic visibility curve — » Flicker fution photometry 450 500 550 WAVELENGTH in m|j -> 600 650 700 Fig. 23. Foveal pupillary and psychophysical spectral sensitivity curves for two subjects (After Fig. 8, Alpern & Campbell, 1962) Dots: Mean differential threshold measurements for pupillary reactions to 2-sec test stimuli (2° test patches, centrally fixated, seen against continuous blue background field approxi- mately 15 200 td). in diameter —retinal illuminance between 100 and Interrupted line: C-I.E. photopic visibility curve. Solid line: Mean results of psychophysical measurements of photopic luminosity (flicker fusion photometry) on the same two subjects with the same apparatus. Rods surrounding fovea suppressed by blue adapting field, preventing them from reacting to stray illumination from foveal test lights . Under these con- ditions foveal pupillary spectral sensitivity curve agreed com- pletely with photopic visibility curve, and flicker fusion spectral sensitivity curve. flashes used were too weak to elicit pupillary reflexes from the blind spot (Fig. 17). Figure 24 again shows the close agreement between pupillary and visual thresholds, even for blue light at the fovea. When the stimulus luminance was increased, however, the effect of stray light was easily recognized: for equal increments in light intensity, the increases in pupillomotor activity were much steeper for blue, green, or white than for red light. Under the same experimental conditions, the visual flicker fusion curves, which are influenced far less by stray light than is the pupil, were precisely parallel for the four colors (Lowenstein, Kawabata, & Loewenfeld, 1964).

64 30 25 2.0 1.5- o <r O o E £ 0 / /v / / / / / / o' / P .' / J I / / A / / / ' /_/ / / ™ / '// / lA f i LOG LUMINANCE Fig. 24. Extent of pupillary contractions to white, red, green, and blue light stimuli at fovea (Fig. 13, Lowenstein et al., [in press] 1964. Extent of pupillary contractions plotted as ordi- nate (in mm) against stimulus luminance as abscissa (in arbitrary log units, with O marking subject's foveal visual threshold to 1-sec white light flashes of 1° area (white arrow)). Foveal visual threshold for red (Wratten filter # 29) shown by black arrow, green (Wratten # 99) by cross-hatched arrow; blue (Wratten #45 + 47) light flashes of same duration and area by shaded arrow. Pupillary reactions to white light (no filter) indi- cated by circles, those to red light by dots, those to green light by triangles, and those to blue light by squares. Note the close agreement between pupillary and visual thresholds, and the markedly less steep increment slope for red light than for other lights (cf. also Figs. 18,19). Day birds such as the pigeon, with predominantly cone retinae, have pupillary spectral sensitivity curves similar to the fovea of the light-adapted human eye, and their pupillary increment curves fail to show the rod-cone break which is typical for the dark-adapted human eye (Tables 7, & 8, & Fig. 2).

65 In color-blind patients, also, there appears to be close parallelism between the visual and pupillary color responses. Such clinical data are, however, still fragmentary. It appears, then, that the pupil follows faithfully the visual color sensitivity under all conditions so far examined. This fact opens a large field of most interesting possible applications. Other Features of Pupillary Reactions to Light 1 . Adaptation (a) Adaptation to darkness: It has already been mentioned that the "dark adaptation used in pupillary experiments was often far from per- fect. In the older work, some kind of illumination was needed to observe the subject's pupil, and even the "infrared" light used since about 1935 for photographic or cinematographic recording of the pupillary responses usually contained some long-wave visible light. Though it was often stated, as it had been at an earlier time about dim white light, that this illumina- tion had only little pupillomotor effect, the results may have differed from those obtained in true darkness. Strictly speaking, then, dark adaptation existed only in those experiments in which the pupillary reactions were recorded by flash photography without other illumination, or were viewed or recorded by infrared-sensitive devices that operated with light beyond the visible spectrum. (b) Light adaptation: Just as different authors used stimuli of various durations, wave forms, and brightness ranges, of different sizes, shapes, or retinal locations, and of different color, a wide variety of adapting lights were used. The combinations of adapting and stimulating lights are too numerous to be mentioned individually. It should not be forgotten in this connection that all methods in which the "difference threshold" was determined, as well as those in which "static" light stimuli were used, were, in effect, dealing with light-adapted states. In some experiments, pre-adaptation to light, or adaptation to a dim background field, were used to minimize the effects of stray light scattered by the ocular media when a light flash of higher intensity entered the eye. This method was termed "masking technique" (Hesse, 1909; Vogt, 1922; Braun, 1931; Schmelzer, 1931; Biffis, 1934; Fry & Allen, 1953; Fugate, 1954; Alpern & Campbell, 1962b). The pupillomotor effects of light adaptation were considered generally only in relation to the reduced sensitivity of the retina, or the altered color sensitivity brought about by the adapting light. It should be stressed, however, that light adaptation influences pupillary reactivity by a second, central nervous mechanism (Lowenstein & Loewenfeld, 1961). In the presence of an adapting light, the supra-nuclear impulses inhibiting the Westphal-Edinger nucleus are reduced, and the functional

66 state of the parasympathetic motor nucleus is thus altered (cf. above section on Pupillary Reactions to Light—Modifying effects of fatigue and emotional excitement). Depending on the autonomic balance of the indi- vidual, this lessening of supra-nuclear inhibition may depress the light reflex, enhance it, or leave it unchanged (Fig- 25). II III Illllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll Fig. 25. Effect of dim illumination upon pupillary reaction to light in three normal subjects with different degrees of supra-nuclear inhibition (Fig. 8, Lowenstein & Loewenfeld, 1961). Pupillary diameter recorded as ordinate (in mm) against time as abscissa (in 0.1 sec units). At a, eyes dark-adapted (solid lines), or adapted to diffuse, blue-green illumination, about 4.5 log units above subject's scotopic visual threshold (broken lines); during periods framed at b, bright, white light flashes presented, approximately 9 log units above absolute visual threshold. A: The subject was a tense, excitable but otherwise normal 22 year-old woman. The pupil was only slightly smaller after adaptation to dim light than in darkness, and the reflex elicited by the bright light flash was slightly enhanced. B: The subject was a normal, calm, well-rested 24 year-old woman. The decrease in pupillary diameter after adaptation to dim light was somewhat greater than in subject A, and the light reflex elicited by the bright flash was slightly reduced. C: The subject was a 36 year-old man, hyper-fatigable because of habitual lack of adequate rest. In darkness his pupillary re- flexes remained within normal limits, but adaptation to dim il- lumination had a profoundly depressing effect: The pupil became quite small, the light reflexes inextensive and pathologic in shape.

67 When the subject is calm and well rested, adaptation to dim back- ground illumination causes the pupil to become slightly smaller than in darkness, while the light reflexes, elicited by standardized light flashes, are unchanged or slightly diminished (Fig. 25B). When the subject is tired, the same background light has a profoundly depressing effect: the pupil becomes quite small, and reflexes to standard light stimuli inexten- sive and pathological in shape (Fig. 25C). When the subject is tense, or emotionally excited, the pupils are large in darkness, and remain large in dim light. In such subjects the background field may enhance the contractions evoked by the standard light flash (Fig. 25A). The central nervous nature of these changes is proven by the fact that they can be obtained also by exposing the opposite, non-stimulated eye to dim light. In experiments in which the eye is adapted to light, it is necessary to consider these central nervous effects. Even weak background light may alter the reflexes considerably, whereby opposite changes may occur in different subjects. 2. Effect of the entrance pupil Since the luminosity of a beam of light entering the eye is affected by the size of the entrance pupil, investigations were done to test the effect of different pupil sizes upon visual functions. These papers, in which the pupil was considered exclusively in its role as a component in the optical system of the eye, have been excluded from this review. In a number of experiments efforts were made to eliminate the effects of variations of the entrance pupil. The means used were: (1) immobilization of the pupil of the stimulated eye, most often by cycloplegic drugs, and occasionally by miotics; (2) small artificial pupils close to the stimulated eye; and (3) optical means, that is, placing the focus of the stimulating light beam in the plane of the subject's pupil. (For the latter two methods, careful immoblization of the subject's head is, of course, essential). 3. Binocular interaction Several kinds of studies fall into this category: (a) Galen, in the second century A.D., used a clinical test that has enjoyed wide popularity throughout the centuries. With the patient facing a source of light, first one eye was covered, and then the other, while the first eye was exposed to the light. Whenever vision was impaired in one eye, the pupil would be larger when it was opened, and smaller when its healthy fellow-eye was exposed. Galen believed the difference to be due to some defect that interfered with the normal flow of 'Vital spirits" from the brain through

68 the hollow optic nerve to the eye; vision was impaired because less of the vital energy was available to illuminate the outside world. Some centuries later it was recognized that the change in pupillary size is due to the difference in afferent conduction from the two eyes. Kestenbaum (1928-1929) has attached the name of "pseudo-anisocoria" to this phenomenon, but since the pupils actually are not unequal at any time, it is the author's opinion that this term is unfortunate. Recently, Leavatin (1959) has advocated a modification of this test, as the "swinging flash- light test", that is, using a flashlight that is moved from one eye to the other. (b) Summation of afferent impulses from the two eyes has been found to exist, within the limits of mechanical possibility: the pupil was found to respond to the total light energy stimulating the two eyes unless the light was so intense that unilateral stimulation alone resulted in maximal constriction (Silberkuhl, 1896; Abelsdorff & Feilchenfeld, 1904; Weiler, 1905; Blanchard, 1918; Reeves, 1918-1920; Luckiesh &Moss, 1934b; Bartley, 1943; Thompson, 1947; Venco &Marucci, 1947; Flamant, 1948; Kawabata9). (c) "Retinal rivalry" has been reported to affect the pupil, so that a larger percentage of near-threshold stimuli resulted in a pupillary contraction when the stimulated eye was in the dominant phase that when it was in the inhibited phase (Nicolai, 1929; Barany feHallden, 1948; Wirth, 1952). These experiments were, however, done without records of the pupillary movements; in view of the great difficulty in judging accurately the small, variable pupillary threshold responses, .especially in dim light, the question cannot be considered as settled. (d) Similar difficulties attend pupillary investigations on amblyopic patients. Some authors described good pupillary reactions in amblyopes: Harms (1949) and Dole'nek, Kristek, Nemec, &Komenda (1962) reported reduction in pupillomotor activity). Harms stated that in cases of unilat- eral amblyopia both pupils enlarged whenever the patient fixed with the amblyopic eye, and became smaller when he fixed with the normal eye. Since, however, the fovea is much more efficient for pupillary contrac- tion than the extrafoveal retina, some means would have to be provided in such experiments to assure exact positioning of the stimulating light spot on the retina, in order to rule out the possibility that the change might be caused by a difference between the effects of foveal and of eccentric fixation. 9 Personal communication, H. Kawabata, 1959-1960.