J. Physiol. (I954) I23,

357 J. Physiol. (I954) I23, 357-366 THE MINIMUM QUANTITY OF LIGHT REQUIRED TO ELICIT THE ACCOMMODATION REFLEX IN MAN BY F. W. CAMPBELL* From the Nuffield Laboratory of Ophthalmology, University of Oxford (Received 13 August 1953) If an emmetrope views a near object the visual axes of the eyes converge and the refractive power of the crystalline lenses increases, resulting in sharp vision without diplopia. While there is close harmony between the degree of convergence and the extent of accommodation, accurate accommodation can occur in the absence of convergence. Normally the conscious awareness of the distance of an object also assists in the production of the correct degree of accommodation. But accommodation can be elicited even without knowledge of the distance of the object from the eye, by suddenly placing a negative spherical lens of moderate power before the eye. The adjustment to this change in vergence of the light occurs rapidly without conscious effort on the part of the subject. It may justifiably be classed as a reflex. The purpose of this paper is to determine the minimum amount of light required to elicit the accommodation reflex. Some observations on the behaviour of the accommodation mechanism in darkness are also reported. METHODS Accommodation changes were determined objectively by photographing the third Purkinje- Sanson image formed by reflexion from the anterior surface of the lens. The dimensions of the image were then measured directly on the negative with a measuring microscope. The size of such an image depends on the radius of curvature of the reflecting surface. When the eye accommodates for a near point the radius of curvature of the anterior surface of the lens becomes less and the size of the third catoptric image decreases. To facilitate measurement a double light source was employed and the appearance of the double image is shown in Fig. 1. A 2J electronic flash tube (Mullard, L.S.D. 7) was used as the source of light for photography. This source was doubled by placing mirrors above and below the tube with a screen before the tube. A 1W compact filament lamp was placed in front of the screen to permit initial adjustment of the camera. A 35 mm 'Wrayflex' camera fitted with a 5 mm f/2 lens and a 5 mm extension tube was used to photograph the images. Photographs were taken at an effective aperture of f/16. Forty-five exposures could be taken at intervals of 2 see without re-loading the camera. Careful focusing and aligning are essential. A single-lens reflex camera of this type is ideal for the purpose. * Present address: Physiological Laboratory, University of Cambridge.

358 F. W. CAMPBELL The test object was a circular disk of light provided by an adaptometer placed 5 m from the subject. The light source of the adaptometer was a small wattage lamp fed from a constant voltage transformer (Colour temperature, 3 K). The luminance of the test object was altered by means of calibrated neutral density wedges. An iris diaphragm placed in front of the stimulus controlled its size. The subject's head was supported by a chin and forehead rest. One eye was occluded with a screen. A special lens holder was constructed to hold lenses of 2 cm diameter before the subject's other eye. These lenses could be placed close to the eye without interfering with photography. All experiments were carried out in a dark room. Fig. 1. Untouched photograph of the Purkinje-Sanson images of the left eye. The bright image to the left of the photograph is formed by the anterior surface of the cornea. The larger and dimmer image near the centre of the pupil is the 3rd image formed by the anterior surface of the lens. The smaller image to the right is formed by the posterior lens surface. To determine the light minimum for the reflex a moderate degree of accommodation was elicited by placing a - 3D lens before the eye and instructing the subject to view the test object. Seven photographs were then taken. The luminance of the stimulus was reduced in steps of -25 log1 units between each photograph. This sequence was carried out with the test object subtending angles of 5', 9', 1-25', 3', 1', 25', 3', 1 and 1-5. The initial luiminance of the test object was adjusted to a convenient level for each size of test object. The subjects closed their eyes between each photograph and re-focused the test object after each adjustment of luminance. Approximately 3 sec was allowed for the subject to find and to accommodate on the test object. To determine the effect of the luminance of the test object on the maximum amount of accommodation which could be exerted, a - 6D lens was placed before the subject's eye. The subject was instructed to accommodate until the test object, which subtended 1, appeared in sharp focus. Photographs were then taken with the test object at known luminances. In addition, a series of

THE ACCOMMODATION REFLEX 359 calibration photographs was made at the high luminance level of 12 tl (microlamberts) with negative lenses of known power before the eye. In this way it was possible to express approximately the amplitude of accommodation in dioptres by referring to the calibration data. The subject re-focused the test object after each adjustment of luminance. While these experiments were in progress the subjects were requested to report any changes observed in the colour, luminance or size of the test object. In order to examine the behaviour of accommodation in the absence of any light sensation, use was made of the foveal scotoma found at low levels of imuimination in the dark-adapted eye. The subject was presented with a disk subtending 3' at 5 m distance and the lulminance of the object was gradually diminished until it was no longer visible foveally, although it was still visible parafoveally. Before each photograph, the subject was instructed to look 'directly at the light' so that its image fell on the fovea. The photograph was taken when the subject reported that the light was no longer visible. In this way the visual alignment necessary for the accurate estimate of the curvature of the centre of the anterior surface of the lens was ensured in spite of the absence of visual stimulation. RESULTS The light minimumfor the accommodation reflex Fig. 2 shows the result obtained with subject H.M.C. while viewing test objects subtending -5', 3' and 25' through a -3 D lens. Photographs were taken after each reduction in the luminance of the test object (-25 log. steps). Far E 1EiO -5' 3'1 25" -1-9 Near 5 4 3 2 1 Log. luminance in microlamberts Fig. 2. The effect on accommodation of diminishing the luminance of test objects subtending -5', 3' and 25'. Accommodation for near was induced by placing a - 3 D lens before the eye. The subject re-focused the test object after each photograph. It can be seen that, when the luminance of the -5' object was 13,15,L or brighter, accommodation for near was induced by the -3D lens. However, when the luminance of the test object was further reduced accommodation diminished suddenly, as shown by the increase in the size of the third Purkinje-Sanson

36 F. W. CAMPBELL image. Similarly, with the 3' and 25' test objects accommodation diminished below a luminance level of 12.5 and 11.25 pl respectively. In Fig. 3 the critical luminance in log1 ul below which accommodation diminished is plotted against the visual angle of the test object. If the relation between the size of the stimulus and its luminance obeys Ricco's law (Area x Luminance = Constant) then the gradient of the experiment curve should be parallel to the interrupted line drawn in Fig. 3. It is clear that for 1 _ IV \1 I I I. I 1 2 3 4 5 Log. luminance in microlamberts Fig. 3. The relationship between the threshold luminance for the accommodation reflex and the size of the test object. If the relationship obeys Ricco's Law the experimental curve should have the same slope as the interrupted line. objects subtending from.5' to 1' this law is substantially true. That is, within this range of visual angle the factor determining the light minimum for the accommodation reflex is the total light energy of the stimulus. The results obtained with larger test objects subtending from 1' to 1.5 deviate slightly from this law. The minimum light energy required to elicit the accommodation reflex is greater at these larger angles. This experiment was repeated on two other subjects over a smaller range of object size with the same results. Effect of lens power on the light minimum In the previous experiments a -3 D lens was used to elicit accommodation. As the power of the lens used to induce the accommodation may influence the determination of the light minimum, a further experiment was designed to estimate the light minimum for the reflex with lenses of different powers before

THE ACCOMMODATION REFLEX 361 the eye. The size of the test object was held at 1 visual angle throughout the experiment. The results obtained with subject H. M. C. are shown in Fig. 4. A linear relationship was found between the dioptric power of the lens used to elicit accommodation and the log1o of the threshold luminance at which accommodation occurred. By extrapolation it may be calculated that with a lens of zero power before the eye the light minimum would be 1 pl with a 1 stimulus. Thus the results presented in Fig. 3 may be approximately 1,L higher than the true threshold for the accommodation reflex due to the use of a - 3D lens to stimulate the reflex. -5 -o - -3 vla- 3-2 -1 Fig. 4. 1 2 Log.luminance (microlamberts) Effect of the power of the negative lens used to induce accommodation on the threshold luminance for the reflex for a 1 test object. Effect of luminance on the amplitude of accommodation The effect of varying the luminance of the test object on the amplitude of accommodation is shown in Fig. 5. A -6 D lens was placed before the eye of the subject (H. M. C.) to stimulate the necessary degree of accommodation. It can be seen that as the luminance of the test object decreased the degree of accommodation exerted to overcome the blurring of the -6 D lens diminished, finally reaching a very low value at 1-25 ul. It should be stressed that in this experiment the subject was instructed to exert voluntarily enough accommodation to overcome the blurring due to the -6 D lens and that the accom-

362 F. W. CAMPBELL modation response was not necessarily involuntary as in the previous experiments. 6- X4/ v 2 1-1I 1 2 Log. luminance (microlamberts) Fig. 5. The relationship between the lulminance of a 1 test object and the amplitude of accommodation. Sulbective observations During these experiments the subjects were instructed to report if they failed to see the test object when viewing it foveally. When their reports were compared with the results obtained photographically it was found that the test object could always be seen when its luminance was -25 log1 unit lower than the light minimum for the accommodation reflex but could only rarely be seen when it was O5 log1 unit lower. If the subjects viewed the test object parafoveally it could be detected 1-2 log1 units lower still, the exact value depending upon their state of dark-adaptation. Frequently, subjects would report a sudden apparent increase in the size of the object as its luminance was lowered. This change always occurred at the point of sudden diminution in accommodative power. Blurring of the test object was never observed when its luminance was below the minimum required to elicit the accommodation reflex. Changes in the curvature of the lens below the light minimum necessary to elicit the accommodation reflex The behaviour of the accommodation mechanism in the absence of a visual stimulus was investigated in thirteen subjects and the result obtained with subject F. W. C. is shown in detail in Fig. 6. For calibration purposes twelve photographs were taken with the stimulus at a high photopic brightness with a, -5 or ID negative lens before the eye. This calibration sequence was

THE ACCOMMODATION REFLEX 363 carried out at the beginning and at the end of the scotopic experiment giving a total of four measurements ( e) at each level of accommodation. The means of these measurements were found ( x ) and the best straight line (AB) drawn through them. Six photographs were then taken at 2 min intervals after the subject had been in absolute darkness for 15 min, and the luminance of the stimulus had been reduced below the photopic threshold and was not perceived by the subject when viewed with the fovea. The mean of these measurements () was then determined and the degree of accommodation assessed from the calibration line AB. In this experiment the mean degree of accommodation present under scotopic conditions was 62D and the range was found to be from 38D to 87D. 1-7 E~~~~ E1*6 8- -8 v -5 154 o -2-4 -6-8 1. Accommodation (dioptres) Fig. 6. Determination of the state of accommodation in darkness. * calibration measurements at, -5 and 1 D of accommodation. x, mean of readings. Line AB drawn through mean of calibration measurements., readings obtained in darkness. Similar experiments were undertaken on a further twelve subjects, and the results are shown in Table 1. It can be seen that in these subjects the refractive power of the lens increased by an average amount of 64D in the absence of a visual stimulus. There is, however, a fairly wide scatter in the individual readings obtained from each subject. This variation is not due solely to the experimental error of the method, for subject no. 2 shows very little variation between readings. It seems likely that fluctuations of accommodation occur in most subjects in the absence of light perception. The frequency of these fluctuations could not be determined as photographs could only be taken once every 2 sec.

364 F. W. CAMPBELL TABLE 1. The increase in the refractive power of the lens found in darkness. (At least 6 measurements in darkness were obtained from each of thirteen subjects) Increase in refractive power in dioptres Subject Minimum Maximum Mean 1-38 -87-62 2 1.1 1-3 1-2 3 1-28 4 1-67 5-25 -52 5 6 *5 1. *75 7 1. -6 8 *35-85 5 9 1-1 -75 1 *6-25 11 1. -6 12 5 95 75 13.5 1. -8 Mean -64 DISCUSSION These experiments demonstrate that the accommodation reflex is activated only when the light energy of the test object exceeds a critical value which is approximately 1,L for a test object subtending 1. This level of luminance is near to the lower limit of the foveal cone threshold for visibility (Mandelbaum, 1941). It may therefore be concluded that the receptors involved in the accommodation reflex are the foveal cones. Comparison of the foveal threshold for visibility, as reported by the subject, with the threshold for the reflex, as detected photographically, reveals that the threshold for the reflex is 25--5 log1 unit higher than the threshold for visibility. This implies that the accommodation reflex requires about twice the light energy necessary to excite sensation at the fovea. The reason for this higher threshold becomes clear if the nature of an out-offocus image on the retina is considered. When the eye is in perfect focus for a given object there is a certain amount of light distributed outside the geometric image on the retina as a result of diffraction, chromatic aberration and spherical aberration. The observer is not normally aware of this degradation of the image border. However, if the vergence of the light is altered so that the retina is not situated at the point of optimum focus there is a further decrease in the contrast between the image border and its immediate surround. When we become conscious of this decrease in contrast we speak of blurring. In the case of a white disk on a dark background the blurred edges of the outof-focus image will have a lower luminosity than the central regions of the image. If the luminance of the object is diminished until it approaches the sensory threshold the perception of the less intense blurred edge of the image will become impossible. Thus the finding of a slightly higher threshold for the accommodation reflex in these experiments suggests that appreciation of the

THE ACCOMMODATION REFLEX 365 blurred edge of the image is necessary for the initiation of accommodation. Furthermore, the finding that the response of the accommodation mechanism to a -6 D lens increases with increase in luminance of the object indicates that the eye only accommodates sufficiently to overcome the blurring which is detectable. The higher the luminance of the object the easier will be the detection of this out-of-focus blurring, and the greater will become the accommodation response. The edge of an out-of-focus image will vary not only in luminosity but also in hue owing to the chromatic aberration present in the normal eye. If the visual apparatus could detect and discriminate the colour of these fringes the accommodation mechanism might adjust itself more accurately. Fincham (1951) compared the effectiveness of white and monochromatic light stimuli on the accommodation reflex in a group of fifty-five young adults. He concluded that 6% of the subjects obtained information on the vergence of the light by means of the colour fringes produced round an image by chromatic aberration. To account for the remaining 4% he suggested that small scanning movements of the eye may permit differences in the vergence of the light to be detected by means of the retinal direction effect (Stiles & Crawford, 1933). It seems improbable, however, that two such entirely different mechanisms are involved in different groups of the population. It is also difficult to believe that the foveal cones could differentiate the unsaturated colour fringes around a 5' object at levels of luminance only -25 logl unit above the foveal threshold for visibility. Furthermore, the brightness difference produced by the Stiles-Crawford effect across a tiny out-of-focus confusion circle would be small. The very low brightness difference threshold necessary to detect such an effect is not found at Lthe luminance levels used in this paper (Hecht, 1935). At scotopic levels of luminance below the cone threshold the optical system of the eye might be expected to assume the condition of minimum refractive power. The final experiment shows, however, that the accommodation mechanism takes up a position about '6D greater than this minimum power. This increase in the power of the lens may account for part of the refractive error found in 'nocturnal myopia'. Otero (1951) undertook a similar investigation and found an average increase of 1 1 D in the refractive power of three subj ects examined in total darkness. Koomen, Scolnik & Tousey (1953), on the other hand, detected no lens change in three subjects but some accommodation in a fourth subject. It would thus appear that in many, but not all, subjects there is under scotopic conditions an increase in the refractive power of the lens which rarely exceeds 1-2D. Spherical and chromatic aberration must also contribute to the total apparent change of 1V5 to 2D found in 'nocturnal myopia' (Koomen, Scolnik & Tousey, 1951). It is interesting to note that Whiteside (1952) found an increase in the refractive power of the lens under

366 F. W. CAMPBELL bright photopic conditions when the eye was presented with a visual field containing no detail. It may be concluded that if the fovea is deprived of visual detail at any level of illumination the mechanism of accommodation takes up a position of relatively fixed focus approximately *6 D greater than the minimum refractive power of the eye. SUMMARY 1. Changes in the state of accommodation were followed by photographing the Purkinje-Sanson image formed by the anterior surface of the lens. 2. For test objects subtending from 5' to 1' the factor determining the threshold for the accommodation reflex was found to be the total light energy of the stimulus. For test objects subtending from 1' to 1.5 the minimum light energy required to elicit the reflex was found to increase with the size of test object. 3. It was found that for a given size of test object the light minimum required to elicit the reflex was only -25 log1o unit greater than the foveal threshold for visibility. 4. The mechanism of accommodation was found to take up a relatively fixed position approximately -6D greater than the minimum refractive power of the eye when the luminance of the test object was below the cone threshold for visibility. It is concluded that the receptors involved in the accommodation reflex are the foveal cones and that in the absence of a foveal stimulus the mechanism of accommodation takes up a position of relatively fixed focus greater than the minimum refractive power of the eye. I wish to thank Prof. R. C. Garry for his active interest and constructive criticism. The photographic apparatus used in this investigation was supplied by the W. H. Ross Foundation (Scotland). Other expenses were defrayed by a grant from the Rankin Medical Research Fund of the University of Glasgow. REFERENCES FrIcHAM, E. F. (1951). The accommodation reflex and its stimulus. J. Physiol. 115, 13P. HECHT, S. (1935). A theory of visual intensity discrimination. J. gen. Physiol. 18, 767-789. KOOMEN, M., SCOLNIK, R. & TOUSEY, R. (1951). A study of night myopia. J. opt. Soc. Amer. 41, 8-9. KOOMEN, M., SCOLNIE, R. & TousEy, R. (1953). Measurement of accommodation in dim light and in darkness by means of the Purkinje images. J. opt. Soc. Amer. 43, 27-31. MANDELBAUM, J. (1941). Dark adaptation. Arch. Ophthal., N.Y., 26, 23-239. OTERO, J. M. (1951). Influence of the state of accommodation on the visual performance of the human eye. J. opt. Soc. Amer. 41, 942-948. STILES, W. S. & CRAWFORD, B. H. (1933). The luminous efficiency of rays entering the eye at different points. Proc. Roy. Soc. B, 112, 428-45. WHITESIDE, T. C. D. (1952). Accommodation of the human eye in a bright and empty visual field. J. Physiol. 116, 52P.