The Influence of Pigment Migration on Vision of Higher Crustacea

Exner (1891) divided the compound eyes of insects and crustaceans into two classes according to the distance separating the crystalline cones from the rhabdomes. In the apposition eye the proximal ends of the crystalline cones were contiguous with the distal ends of the rhabdomes, whereas in the superposition eye the crystalline cones were far removed from the rhabdomes. Exner believed that in an apposition eye a distinct image was always formed because only rays parallel to the axes of the ommatidia were received by it. On the other hand, the clarity of the image in a superposition eye depended on the position of the distal retinal pigment.

In the eye of an animal adapted to darkness, where the distal pigment was withdrawn to form a collar round the cones, the image was formed by rays refracted by the cones acting as ‘lens cylinders’, as well as by rays passing straight through. This Exner believed would give rise to a well-illuminated but blurred image because the focus of the rays would not be perfect. In the eye of an animal adapted to light, the distal retinal pigment moved proximally away from the cones. Exner believed that in this condition the oblique rays would be excluded so that, as in the apposition type of eye, only those rays parallel to the ommatidial axis would fall on the retina. The superposition eye could therefore form a sharp apposition image 01 a less distinct but brighter superposition image, depending on the position of the distal retinal pigment.

Exner considered that the function of the proximal retinal pigment of arthropods was to enhance stimulation in very weak light and to prevent overstimulation of the sensitive retina in very bright illumination.

Hassenstein (1954) attempted to compare the visual acuity of light-adapted and dark-adapted eyes in two species of prawns, Leander serratus and Lysmata seti-caudata. He found no change in visual acuity whether the eye was in the dark-or light-adapted state. He conceded, however, that his experiments were not conclusive..

It seemed desirable, therefore, to test Exner’s hypothesis on a number of Crustacea. The experiments described below were made on three species of prawns Leander serratus (Pennant), L. squilla (Linné) and Pandalus montagui Leach and on a mysid Praunus flexuosus (Müller). Species of Leander and Praunus were collected living amongst weeds in rock pools, while Pandalus was, trawled from depths ranging from 6 to 40 fathoms. Eupagurus bemhardus (Linné), a hermit crab found on the shore and below low-water mark, was used to typify the behaviour of a crustacean with apposition eyes (Brocker, 1935).

The eyes of the first three species are essentially of the same pattern as that of Palaemonetes vulgaris Say, which has been fully described by Parker (1891). They are typical superposition eyes in which the crystalline cones lie at a considerable distance from the rhabdomes. The crystalline tract is a slender strand, the broader distal end of which is closely opposed to the lower surface of the cones. In this region formalin-fixed frozen sections reveal that the uppermost part of the tract is packed with clear globules, apparently of oil, which are lost during ordinary histological treatment. There is also a little red pigment in this region. The lower end of the crystalline tract is closely applied to the rhabdomes, though Parker (1891) states that it is not continuous with them.

If balsam-mounted unstained sections through the ommatidial axes are examined by the interference microscope, the crystalline tracts and the rhabdomes stand out as strongly refracting elements. They seem to be optically continuous with each other (Pl. 6). Sections of eyes of Leander serratus, cut transversely to the ommatidial axes, show that the proximal end of the crystalline tract becomes hollow and expands into four processes which lie just above the rhabdome.

Employing the interference microscope on 8 p thick unstained sections we have observed that the crystalline cones have a slightly higher refractive index (Δn = + 0-003 to 0.0006) than the surrounding balsam and are separated from each other by material of lower refractive index. The distal extremity of the crystalline tract referred to above has a refractive index which is only a little different from that of the surrounding balsam, but the rest of the crystalline tract and the rhabdome have relatively higher refractive indices (Δn = + 0.015 to 0·030). They are surrounded by a matrix scarcely differing in refractive index from the mounting medium. The cornea is also highly refracting. Thus, apart from a small region at the base of the cone, the crystalline cones and tracts form continuous strands of higher refractive index surrounded by media of lower refractive index. Balsam-mounted sections, cut transversely to the longitudinal axes of the ommatidia, fail to show any evidence that the crystalline cones refract more strongly at their central axes than at their periphery. Exner’s hypothesis that they function as lens cylinders demands that the refractive index of the cone in the living eye should diminish from the centre outwards. These results therefore throw a little doubt on Exner’s assumptions regarding the function of the cones.

Observations on fixed eyes cannot afford conclusive evidence of their condition in life. They may, however, give a good indication if the fixation is rapid and satisfactory, that is if the material responsible for refraction in the living eye is not disturbed and if the mounting medium replaces the body fluids originally present in the interstices and tissues. Since the optical path difference is approximately determined by the total mass through which the light passes (Barrer, 1956) the refractive indices of the various tissues, if well fixed, should have a similar relationship to one another, though not the same value, as they had in life.

Unfortunately, the material comprising the crystalline elements of the eye is gelatinous and unstable, so that it has not yet been found possible to study the optical properties of untreated ommatidia.

The eyes of the Decapoda Natantia and the mysid Praunus flexuosus contain three groups of pigment cells. Two of these groups contain a black pigment, probably melanin; these are the distal and proximal retinulae. These cells are grouped regularly about each ommatidium. The two distal pigment cells extend from the crystalline cones down to the basement membrane, but the pigment is most concentrated in a narrow band which varies in position. The eight proximal retinulae are arranged radially around the rhabdome and penetrate the basement membrane of the eye. One of them differs from the rest ; its nucleus lies closer to the basement membrane than the nuclei of the other seven, and is placed excentrically in the Decapoda. None of these nuclei are covered by pigment since the proximal pigment always lies deeper down in the cells. Only the proximal pigment cells are innervated and these alone are believed to be the light-sensitive cells. Parker’s use of the term distal retinulae is thus misleading. The third group of pigment cells, the accessory or reflecting pigment cells, are not arranged in relation to the ommatidial elements and are fewer in number than the ommatidia according to Knowles (1950). They contain a white amorphous reflecting pigment and their processes extend from below the basement membrane to the level of the bottom of the rhabdome.

The eye of the mysid Praunus flexuosus has been described by Mayrat (1956). It differs from that of the Decapoda Natantia in several respects. The crystalline cone is formed from two, not four, elements (Grenacher, 1879). The cones have a short crystalline tract and the nuclei of the proximal retinulae lie close to the cones. There are eight proximal retinulae and one differs from the rest, as in Decapoda. The nucleus of this eighth retinula, however, lies centrally in a clear space formed at the junction of the short crystalline tract and a prolongation of the rhabdome termed by Mayrat the epirhabdome. Mayrat calls the eighth retinula the ‘accessory cell’ ; it should not, however, be confused with the reflecting pigment cells forming the tapetum which Parker calls the ‘accessory pigment cells ‘. These cells are present in Praunus and their pigment forms a tapetum which migrates reciprocally with the dark pigment of the proximal retinulae, just as in Decapoda Natantia.

The anomuran Eupagurus bemhardus has no white reflecting pigment layer (Bröcket, 1935), though it possesses a golden pigment which gives a strong reflexion to rays falling obliquely to the axes of the ommatidia. This pigment is usually lost during histological treatment. It does not seem to occupy the same position nor to migrate as does the tapetal pigment of Decapoda Natantia. The eyes of hermit crabs, whether kept in light or in darkness, thus have a similar appearance.

In a dark-adapted eye the distal pigment lies on either side of the crystalline cones, the proximal pigment lying entirely beneath the basement membrane while the reflecting pigment lies above it. During adaptation to light the distal pigment migrates proximally towards the rhabdomes, the proximal pigment migrates distally to envelop the rhabdomes, while most of the reflecting pigment migrates below the basement membrane. A small amount of reflecting pigment is left as a thin band just outside the basement membrane but entirely covered by dark proximal pigment.

An old and still unresolved problem is the method of control of the retinal pigments.

Welsh (1930 a) observed that when dark-adapted Palaemonetes were given a short flash of light and then returned to darkness the distal pigment continued to move towards the light-adapted position. Kleinholz (1934) suspected therefore that the distal pigment might be under a hormonal control and devised experiments to test whether this was true. He injected the eye-stalk extracts of light-adapted Palae-monetes into dark-adapted animals in the dark and found that the distal pigment migrated proximally. Kleinholz (1936) found that the reflecting pigment also responded to the injections but not the proximal pigment. Welsh (1939) stated that by using greater concentrations of eye-stalk extract the proximal pigment also could be made to respond.

Kleinholz (1949) removed the sinus gland from specimens of Cambarus and found that the proximal pigment continued to respond to changes in illumination but not the distal pigment. This confirmed that the distal pigment was controlled by secretions of the sinus gland but threw doubt on the possibility of hormonal control of the proximal pigment. Knowles (1950) working on Leander serratas obtained exactly similar results.

The distal retinal pigment is known to display a diurnal rhythm under constant conditions of light or of darkness in a number of crustaceans. Evidence for rhythmic changes in the proximal pigment, however, is less well founded. It has been demonstrated in conditions of constant darkness in a proportion of species studied (Table 1), but has been claimed under conditions of constant illumination only in a single instance. In the majority of species the state of the proximal pigment depends only on the conditions of illumination at the time.

Though the evidence indicates strongly that a hormone controls the movements of the distal retinal pigment it is difficult to reconcile with this view certain observations by Parker (1897). He found that not only did both pigments in excised eyes respond to changes in illumination, but also that if part of the eye of an animal was covered, the pigments of the shaded ommatidia became dark-adapted though the rest of the eye was fully light-adapted. The only evidence for the hormonal control of the proximal pigment, on the other hand, is the injection experiment of Welsh (1939) and the few instances of rhythmic behaviour given in Table 1.

All the other results indicate that the proximal pigment is independent of sub-stances circulating in the blood stream. Its behaviour during changes in illumination is certainly different from that of the distal pigment and appears to be a direct response to light.

Knowles (1950) studied the migration of the retinal pigments of L. serratus and L. squilla. He considered only the migrations brought about by changes from light to darkness and vice versa. Neither the diurnal movements of the pigments in constant light nor the rates of adaptation were dealt with. Moreover, no information is available on the behaviour of the retinal pigments of Pandalus montagui and Praunus flexuosus. For the purposes of this investigation it was therefore necessary first to make a thorough study of the behaviour of the retinal pigments of all these species.

The changes in the retinal pigments were studied with the aid of sections of the eyes of animals treated in various ways and rapidly killed in very hot water. Sections were prepared using the technique recommended by Knowles (1950).

In order to define the position of the pigments independently of the size of the eye or of any shrinkage during fixation, an index for each pigment was introduced. The index in each case was derived from the distance separating a defined level of the pigment layer from a fixed level in the eye and was expressed as a fraction of the distance from the basement membrane to the outer surface of the crystalline cones.

The numerator of the distal retinal pigment index is the distance measured from the middle of the band of pigment to the outer surface of the cones. This index was always positive and increased in value as the eye became light-adapted.

The basement membrane affords the best reference surface for the proximal and reflecting pigments. These pigments may be found both proximally and distally to the basement membrane according to the state of the eye but unlike the distal pigment they do not migrate as a regular band. During light-adaptation the outer boundary of the proximal pigment moves in a regular and measurable way beyond the basement membrane but the boundary of this pigment below the basement membrane is ill-defined and the pigment here gradually fades. It is therefore desirable to differentiate between the behaviour of the pigment below and above the basement membrane by introducing two separate indexes, the one for outer surface of the pigment above and the other for the level of its inner boundary below the basement membrane. The indexes for the surfaces below the basement membrane are given a negative sign.

The reflecting pigment is similarly treated, the levels above and below the membrane being recorded as separate indexes. The indexes may therefore be formulated as follows (see Text-fig. 1):

Freshly collected specimens of the four animals studied were divided into two groups. The first group was placed in a light-proof box in a dark room at a temperature ranging from 15 to 18⁰ C. The water was kept well aerated with a diffuser in order to avoid any interference with the pigment responses due to lack of oxygen (Bennitt & Merrick, 1932). The second group was kept similarly aerated at the same temperature under illumination of 7-6 foot-candles (f.c.) in a glass dish.

The average pigment indexes of individuals removed at intervals are given in Tables 2-5. All four species behaved similarly.

It will be seen that in the dark the distal pigment index has its minimum value, i.e. it occupies its nearest position to the cones. The proximal pigment is entirely below the basement membrane and the reflecting pigment above it. In continuous light the proximal pigment lies entirely above and the reflecting pigment has partially migrated below.

It will be further noted, from an inspection of the sample estimates of the standard errors, that in no case is the proximal or reflecting pigment index significantly different between animals killed at night or during the day.

On the other hand, it will be seen that the values for the distal pigment index under continuous illumination are significantly higher in animals killed during the day than in those killed at midnight. The distal pigment therefore moves towards the dark-adapted position during the night even if kept well illuminated. The degree of dark-adaptation is greatest in Praunus flexuosus and least in Pandalus montagui.

When exposed to light the distal pigment migrates inwards and the proximal pigment moves outwards covering the reflecting layer. The position of the outer boundary of the reflecting layer shows only a small change during light-adaptation, a thin layer always remaining between the basement membrane and the proximal pigment. It seemed reasonable to ignore changes which took place below the basement membrane as these could scarcely influence the response of the retinal receptors to light. We therefore concentrated attention on the migration of the proximal pigment distal to the basement membrane as it passed across the reflecting pigment layer and came to lie over it at the side of the rhabdomes. This movement is given by the proximal pigment index above the basement membrane. No attempt was made to follow the changes below the basement membrane as there was no clearly defined boundary to the pigment. The pigment gradually fades from this region as it becomes concentrated above the reflecting pigment.

Table 6 gives the values of the distal and proximal pigment indexes of three of the species studied after being exposed to illumination of 1.1 f.c. for various lengths of time. To eliminate the effects of diurnal variations these experiments were conducted between 10 a.m. and 2.30 p.m.

There is considerable variation, as might be expected in unreplicated experiments, but the table shows that in all three species the proximal pigment migrates more rapidly into the light-adapted position than does the distal pigment. The last column of the table gives the index after prolonged exposure to illumination and is taken from Tables 2-5. The approximate time for the migration of the proximal pigment into the light-adapted state in Praunus flexuosus was about 4 min., for Leander serratus about 4 min., and for Pandalus montagui between 4 and 6 min. The corresponding times for the distal pigment to migrate into the fully light-adapted position were 20, 90 and more than 40 min. respectively. Moreover, the proximal pigment has reached its limiting position before any detectable change has occurred in the distal pigment. The proximal pigment below the basement membrane continued to diminish in quantity for some time after the outer boundary had reached its limiting position, so that some flow of pigment distally must have continued. However, this process does not lead to any noticeable increase in the density of the pigment around the rhabdomes which appears completely enveloped in dense black pigment after 4-6 min.

The independent behaviour of the retinal pigments allows experiments to be carried out to test Exner’s views on the function of the distal pigment. Animals exposed for long periods to light during the day will have the distal and the proximal pigments in the light-adapted state. At night, however, the distal pigment will have migrated to some extent towards the dark-adapted condition, though the proximal pigment will remain lght-adapted. Animals which have been fully dark-adapted and kept for a few minutes in the light will have the proximal pigment in the light-adapted state and the distal pigment will still be in the fully dark-adapted position. It is not possible to obtain the distal pigment in the light-adapted state while the proximal pigment remains dark-adapted.

Specimens of Pandalus montagui, Leander serratus and Praunus flexuosus were first kept in complete darkness for at least 4 hr. A specimen was then removed from the light-proof box and placed at the centre of a cylindrical glass dish with a rotating striped cylinder outside. The visual acuity is expressed as the minimum angle between adjacent black and white stripes which causes an optomotor response. At the beginning of the experiment a cylinder with a small number of stripes was used. The light was then switched on giving an illumination of 1.1 f.c. at the bottom of the glass vessel. In this relatively weak light the dark-adapted animals did not show any distress or give an avoiding reaction. The time taken for the animal to respond to the rotatory pattern, either by swimming characteristically round the dish in the same direction as the rotation of the striped pattern or by nystactic eye-stalk movements, was recorded, and a cylinder with a larger number of stripes was quickly substituted. This resulted in the temporary cessation of the response. The time at which a response was again elicited was once more recorded and another substitution made. In this way the visual acuity was determined at intervals after initial exposure to the light. As a further check, in some experiments a cylinder with a larger number of stripes was used at the outset to confirm that the time at which the response was first obtained was independent of the order in which the patterns were offered. The results are shown in Table 7. It can be seen that there is some difference between individuals, but an improvement in visual acuity always occurs.

This takes place within 3 min. in Pandalus montagui and within 6-10 min. in Leander serratus and Praunus flexuosus. Text-figs. 2-4 illustrate how the visual acuity, the distal pigment index and the proximal pigment index vary with time after exposure to light for each of the three species studied. The figure demonstrates clearly that the visual acuity rises while the proximal pigment is migrating over the reflecting layer and surrounding the rhabdomes. The movement of the distal pigment which takes place afterwards has no influence on visual acuity.

Freshly collected specimens of Leander serratus and Praunus flexuosus were kept under constant illumination of 7 6 f.c. for at least 2 hr. The visual acuity was measured, using the apparatus previously described, between 10 a.m. and 12 noon when the distal pigment was fully light-adapted, and between 12 midnight and 2.30 a.m. when the distal pigment was partially dark-adapted. Some of the animals used in these experiments were killed and the eyes sectioned to confirm the position of the pigments.

Table 8 gives the values of the minimum visual angle for individuals measured in the middle of the day and at midnight. There is clearly no difference in the visual angle caused by the shift in position of the distal pigment. Judging from the mean values of the distal pigment index, the migration of the pigment at night averages about two-thirds of the distance towards the fully dark-adapted state in Leander serratus, while in Praunus flexuosus the distal pigment is fully-dark adapted. Even in Leander, however, some individuals become fully dark-adapted at night ; nevertheless, no change in visual acuity was ever observed.

The following experiments were performed to test whether the dark-adapted eye was more sensitive than the light-adapted eye. Fresh specimens of Leander serratus and Pandalus montagui were used and also Eupagurus bemhardus since this species shows no mechanism for dark-adaptation (Bröcket, 1935). One group of animals was adapted to darkness by keeping them in a light-proof box for at least 3 hr. Another group was illuminated at an intensity of 1·1 f.c. for at least 3 hr. to ensure that they were fully light-adapted.

Each group was then tested with a broad striped pattern in which the alternate stripes were separated by an angle of 72 degrees. This was well within the limit of visual acuity. The test was carried out under a dark room safe-light (Kodak deep red iso-filter) which cut off all wavelengths below 6000 Á. The intensity of illumination was calculated to be 7·2 × 10^-3 f.c. at the bottom of the vessel containing the animals. This light intensity was so low that the eyes of the observer required time to become sufficiently adapted before commencing the experiments. It was seen that only the dark-adapted individuals Pandalus and Leander were able to respond under these conditions. Evidently they are fairly sensitive to light of long wavelength. Eupagurus bemhardus gave no response at all, nor did light-adapted individuals of Pandalus and Leander (see Table 9).

The visual acuity of dark-adapted individuals of the two species which responded was then measured under the weak red light.

Striped cylinders of decreasing angular separation were rotated around the animals first clockwise and then anti-clockwise for 1 min. When no response was given after ten such trials the limit of visual acuity was considered to have been reached. Nystactic movements of the eye-stalks could not be seen, only body movements of the animals could be taken into account.

The visual angle was found to be higher than that given by light-adapted animals in white light, showing that the greater sensitivity of dark-adapted animals in the weaker red light is associated with loss of resolution.

When this experiment was repeated under white light of 1·1 f.c., initially dark-adapted prawns did not respond at once ; this we know is because when the pigments are dark-adapted the visual acuity is initially very low (see Text-fig. 1). After a short time, however, they became light-adapted and responded normally. Eupa-gurus bemhardus, on the other hand, responded immediately whether it had been previously kept in the dark or the light. Moreover, if the dark-adapted animals are exposed to an intense white light of about 500 f.c., these animals show a violent escape reaction, attempting to jump out of the dish. This reaction lasts 2-3 min. The light-adapted prawns and E. bemhardus, whether previously kept in the light or the dark, showed no discomfort when placed under strong white light.

We shall first consider Exner’s view that in the dark-adapted state a superposition image less distinct than the apposition image is allowed to reach the retinulae, but is prevented from doing so in the light-adapted eye by the distal pigment. This view can no longer be maintained because there is no change in visual acuity when the distal pigment migrates down from the cones. We must therefore examine how far this fact undermines the remainder of Exner’s hypothesis. Two possibilities must first be distinguished according to whether the dioptric apparatus or the receptor elements limit the acuity of vision. If the dioptric apparatus is limiting, then it follows that the apposition image and the superposition image formed by it must have approximately equal resolutions. If, on the other hand, the retinal elements limit the visual acuity, as seems likely (de Bruin & Crisp, 1958), then we can only conclude that both images are sufficiently sharp to meet this requirement. A third possibility also emerges that the cones do not act as lens cylinders forming a superposition image as Exner supposed. The rays received by the retinulae would then be only those falling on the cone connected with them. The visual mechanism would therefore be the same whether the distal pigment were in the dark-adapted or in the light-adapted state. Such a possibility is consistent with the optical properties of the crystalline elements so far as can be judged from fixed preparations. Since their refractive index is greater than that of the surrounding tissues, they will act as ‘wave guides’ retaining by internal reflexion any ray whose direction is close to that of the ommatidial axis. These rays would pass down to the rhabdomes which would probably scatter the light around the adjacent retinulae.

Whichever of these alternatives is accepted the function of the distal retinal pigment is obscure. Since it does not affect the visual acuity, but obviously must restrict the light reaching the retinulae, it is natural to assume that its migration may function as a light-regulating device, perhaps to protect the retinulae from excessive illumination. The apertures through which light can enter the retinulae certainly become reduced as the distal pigment moves proximally, and as the pigment is concentrated over a much smaller area it becomes denser. There is as yet, however, no experimental evidence to support this view.

The proximal pigment cells and the reflecting pigment are best considered together since their functions are complementary. The dark pigment surrounding the rhabdome in the light-adapted state will clearly prevent light from being reflected back from the rhabdomes to the retinulae other than those immediately surrounding it. On the other hand, in the dark-adapted state the reflecting pigment lies immediately behind the rhabdomes, which are now no longer covered by black pigment. Stray reflexions can therefore pass in all directions. Bearing in mind the very small distances separating the bases of the ommatidia, these secondary reflexions will cause surrounding ommatidia to be stimulated in addition to those which are primarily illuminated. Thus the visual image will be less distinct. The black pigment therefore improves the visual image in a manner analogous to that of the anti-halation backing of a photographic plate. In the dark-adapted state the reflecting pigment increases the illumination falling on the retinulae though at the expense of definition. Our experiments in dim light show conclusively that dark-adapted animals are more sensitive to light, though this may be due not only to the reflexion of incident light by the accessory pigment cells but also to the adaptation of the photoreceptors themselves. However, this effect is not shown by Eupagurus bemhardus which lacks a tapetum. We have also shown that visual acuity is reduced when the reflecting pigment lies next to the retinulae, both in weak and in normal light intensities, though the latter soon brings the proximal pigment into the light-adapted position. It follows that the observed loss in visual acuity of dark-adapted animals is not related to the amount of illumination falling on the eye, as is sometimes the case (Hecht & Wolf, 1929), but is due to changes in the arrangement of pigments beneath the light-sensitive cells. This conclusion is borne out by the close parallelism between changes in the visual acuity and the position of the proximal pigment when dark-adapted animals are exposed to light.