ABSTRACT
The pupil mechanism in the acone apposition eye of the semi-aquatic common backswimmer Notonecta glauca (Hemiptera) was investigated with infrared reflectometry of the pseudopupil. This method allows non-invasive continuous measurements of pupil responses in the living animal. The dynamic range of the pupil sensitivity is about 7 log units during daytime and 6 log units at night. During the day, the sensitivity range of the pupil covers the normal daylight intensities in the animal’s habitat, just under the water surface (I50=1019.2 photons m−2 sr−1 s−1). At night, the sensitivity is 1 log unit lower (I50=1020.2 photons m−2 sr−1 s−1), ensuring that the pupil is maximally open when light intensities are low. During daytime, light adaptation is completed in slightly less than 40 min, and dark adaptation takes approximately 50 min. The pupil response is only slightly slower at night. The speed of the response as well as the pupil sensitivity are dependent on the preceding adaptation history. An endogenous circadian rhythm determines the control range of the pupil aperture. However, the rhythm is easily disturbed, especially within a 3 h period before dusk and dawn. The results are compared with corresponding results from other insects with the same type of pupil mechanism.
Introduction
Pupil mechanisms are usually controlled directly by the light entering the eye, making a feedback loop for dynamic sensitivity control. This is the case in both vertebrates (Reeves, 1918) and invertebrates (Franceschini, 1972; Land and Osorio, 1990). There are, however, several examples of invertebrate pupil mechanisms that are partly controlled by an endogenous circadian rhythm, for instance in the scorpion Androctonus australis (Fleissner, 1974), the horseshoe crab Limulus polyphemus (Barlow et al. 1989), the portunid crab Scylla serrata (Leggett and Stavenga, 1981), the tenebrionid beetle Tenebrio molitor (Wada and Schneider, 1968; Ro and Nilsson, 1993a) and tipulid flies (Williams, 1980; Ro and Nilsson, 1994). Why do these animals have a circadian rhythm regulating a mechanism that is strictly light-regulated in other animals?
Tenebrionid beetles and tipulid flies have the same type of eye as the common backswimmer Notonecta glauca (Bedau, 1928; Lüdtke, 1953, Horridge, 1968): an acone apposition eye (Grenacher, 1879) with an open rhabdom, consisting of two fused central rhabdomeres surrounded by a ring of six separated peripheral rhabdomeres. Another shared feature is the adjustable pigment aperture in front of the rhabdom. However, a circadian pupil rhythm has not yet been reported in Notonecta.
Recently, we described a non-invasive method of infrared reflectometry, which allowed direct and precise monitoring of the pupil rhythm in the eye of living Tenebrio molitor (Ro and Nilsson, 1993a). This method was subsequently used for studies of the strong circadian dependence of the pupil mechanism in tenebrionid beetles (Ro and Nilsson, 1993b) and tipulid flies (Ro and Nilsson, 1994). In this paper, we use the same method to investigate the dynamics and sensitivity of the pupil mechanism in the backswimmer Notonecta. We found that Notonecta has a method of pupil control like that of tenebrionids and tipulids, but with a surprisingly weak influence from the circadian rhythm.
Materials and methods
Adult backswimmers, Notonecta glauca L. (Hemiptera, Heteroptera), were collected locally and kept in aquaria. Normal day length was maintained by the daylight from large windows.
Under CO2 anaesthesia, the animals were mounted with wax in plastic tubes, thus immobilizing the head. A piece of tissue paper, moistened with water, was placed in the plastic tube behind the animal to maintain high humidity during the experiments. Each animal was centred in a goniometer for easy alignment of the appropriate eye region (a slightly dorso-frontal location was chosen). The goniometer with the mounted animal was then placed in the experimental apparatus (Fig. 1). A ‘dark box’ eliminated disturbance from stray light from the optical bench and ensured complete darkness during dark adaptation. The optical arrangement contained two coaxial light sources, one for blue stimulation and one for infrared reflectometry.
The intensity of the blue stimulation beam was measured with a radiometer (United Detector Technology 61). At the position of the animal’s eye in the apparatus, the intensity of the stimulation beam was 1022.7 photons m−2 sr−1 s−1 (measured through a 550 nm short-pass filter, without neutral density filters). The intensity of the same stimulation light was also measured with a photometric filter (transmission range 430−750 nm) in front of the radiometer, and this ‘photometric intensity’ was calibrated against normal outdoor intensities at different times during the day (see Land, 1981; Lythgoe, 1979). The outdoor intensities were measured with the radiometer through the photometric filter on a reference surface (a Kodak Gray Card with 18 % reflectance), representing an object with intermediate reflectance in a normal outdoor scene.
Data from the reflectometrical measurements were acquired with a MacADIOS data acquisition card and LabVIEW data acquisition software.
Results
The pseudopupil in visible and infrared light
In normal visible light, the pseudopupil in Notonecta appears as a dark spot on the reddish brown eye (Fig. 2A). The dark appearance of the principal pseudopupil (Exner, 1891; English translation by Hardie, 1989) results from the absorption of light by dark screening pigment surrounding the proximal part of the crystalline cone in the ommatidia that are aligned with the direction of observation (Stavenga, 1979). Under axial illumination, a deep pseudopupil (DPP) (see Franceschini and Kirschfeld, 1971; Franceschini, 1975) can be observed at the level of intersection of the ommatidial axes of the eye. The DPP is a superimposed, somewhat blurred, enlarged optical image of the structural elements around and immediately below the cone of the ommatida directed at the observer. In Notonecta, this ‘visible-light DPP’ (not shown here) never changed in size between light and dark adaptation, or between day and night.
Using infrared light of certain wavelengths, combined with image intensification, another pseudopupil phenomenon appears in Notonecta: the ‘infrared pseudopupil’ (see Ro and Nilsson, 1994). The infrared light is reflected, not absorbed, by the screening pigment in the two primary pigment cells, which form an aperture above the rhabdom. The reflecting primary pigment cells contrast with the dark central spot of the ‘infrared DPP’, which is the result of lack of scattering of infrared light from the structures behind the aperture. This pigment aperture closes during light adaptation and opens during dark adaptation (Lüdtke, 1953). As a consequence, the infrared DPP in Notonecta changes in size depending on the adaptation state. In Fig. 2B−E, the infrared DPP is shown in the four extreme adaptation states: dark-and light-adapted at night and during the day. Infrared reflectometric measurements over a circular area determined by the outer edge of the fully dark-adapted infrared DPP show a signal that increases with the degree of closure of the pigment aperture.
The sensitivity of the pupil
The absolute sensitivity of the pupil mechanism was investigated by measuring the response of the pupil to a series of successive light intensity increments. The resulting R/logI curves (Fig. 3, circles) show that, at 50 % response, the pupil is 1 log unit less sensitive at night (I50=1020.2 photons m−2 sr−1 s−1) than during daytime (I50=1019.2 photons m−2 sr−1 s−1). The slope of the near-linear part of the curve is less steep during the day (m=0.38) than at night (m=0.66). Extended dark adaptation clearly shifted the R/logI curve towards higher light intensities for individuals tested both during the day and at night. In Fig. 3 (triangles), this effect is shown at night (during the day, a prolonged dark adaptation also affected the circadian pupil rhythm, which is discussed below).
The dynamics of the pupil
Fig. 4 shows the changes in reflection, plotted against time, during the pupil responses between the fully dark-and light-adapted states, measured in the same animal both during the day and at night. ‘Fully dark-adapted’ and ‘fully light-adapted’ were defined as 10 min of unchanged reflection. The appropriate light intensity to adapt the pupil fully to light was determined from the sensitivity measurements (see Fig. 3). The speed of the pupil mechanism was only slightly slower at night than during the day. Light adaptation (from fully dark-to fully light-adapted) took 35−37 min during the day and 36−38 min at night, whereas dark adaptation took 48−51 min during the day and 48−53 min at night (extremes of three measurements). Prior to these measurements, the animals were moved directly from the natural outdoor light intensity to the dark-box, and the period in the fully dark-adapted state was no longer than 10 min, just long enough to be sure that the pupil was fully closed. A longer period of dark adaptation prior to the measurement resulted in a slower pupil mechanism.
The circadian pupil rhythm
There is clearly an endogenous circadian rhythm involved in the regulation of the pupil mechanism in Notonecta. The stability of the rhythm, however, is not as high as in, for instance, Tenebrio molitor, where the rhythm continues almost unaffected for at least a week under constant, free-running conditions (Ro and Nilsson, 1993a). During the experiments on the sensitivity and dynamics of the pupil in Notonecta, we found that the endogenous pupil rhythm was very easily disturbed, especially during the afternoon and in the early morning hours. Therefore, a simple experiment was designed to test the stability of the rhythm and the tendency towards deviation after exposure to abnormal light intensities. An animal, entrained to normal day length, was first exposed to total darkness in the afternoon and then to a 30 s ‘flash’ of room light in the middle of the night (see Fig. 5). The results confirm that the endogenous pupil rhythm is easily disturbed, especially within a period of 3 h before dusk and dawn.
Discussion
Recently, we reported on the sensitivity and dynamics of the pupil mechanism in the compound eyes of tenebrionid beetles (Ro and Nilsson, 1993b) and tipulid flies (Ro and Nilsson, 1994). In both cases, the sensitivity of the pupil was found to correspond to the light intensities in which the animals are normally active. Both tenebrionid beetles and tipulid flies have the same type of acone apposition eye and the same type of pupillary movements during light adaptation (retinomotor movements; see Autrum, 1981) as the backswimmer Notonecta (Lüdtke, 1953). The lifestyle of Notonecta is, however, totally different from that of tenebrionid beetles and tipulid flies. Notonecta is a diurnal visually guided predator, which usually hangs under the water surface waiting for prey, but also frequently swims about actively hunting for prey. In addition, the adult backswimmer often flies between ponds both during the day and at night when the weather is warm and clear during late summer and autumn (Walton, 1935; Schwind, 1983a).
In Fig. 6, the pupil sensitivity (R/logI curves) of Notonecta is shown together with the pupil sensitivity of the tenebrionid beetle Zophobas morio and the tipulid fly Tipula luteipennis. There is a clear difference in the sensitivity of the pupil mechanism between the diurnal Notonecta and the crepuscular Tipula. The maximum sensitivity of the daytime pupil in Notonecta is adjusted to daylight intensities in shallow waters, corresponding well with the intensities in which Notonecta hunts. At depths down to 1 m, light intensities drop by less than 0.5 log units in clear water (Clarke and Denton, 1962) and the reflection loss at the water surface is less than 10 %. Tipula usually avoids bright sunlight, and its activity period is mainly during dusk and dawn, which explains why the sensitivity of the daytime pupil is adjusted to lower intensities. Zophobas is a Central American species, usually found in caves, breeding on bat guano (Tschinkel, 1984). The extreme shift of the R/logI function towards low intensities is therefore not surprising. In all three species, the pupil sensitivity at night-time is shifted towards higher light intensities, which seems contradictory. However, this is probably a way of ensuring that the pupil stays fully open at the light intensities normally encountered, in order to exploit the advantage of an open rhabdom at night (see Nilsson and Ro, 1994). It would be unfortunate if Notonecta were blinded by light-adaptation caused by the strong illumination from a full
The dynamic range of the pupil sensitivity also differs between the three species (Fig. 6). In Zophobas, the total range (day and night) is about 5 log units of light intensity, whereas the corresponding value is about 9 log units in Tipula and about 8 log units in Notonecta. The pupil of Zophobas is specialized for the low light intensities in its natural habitat. Pupil mechanisms of other insects are known to be even more specialized for a limited range of light intensities, for instance in the nocturnal moth Hydraecia micacea (Noctuidae), with a dynamic range of about 4 log units (Nordtug, 1990), and in the blowfly Calliphora erythrocephala, with a dynamic range of 3 log units, adjusted to high daylight intensities (Roebroek and Stavenga, 1990). The pupils of Notonecta and Tipula, however, are tuned to function over a wide range of intensities. This seems to be appropriate in the crepuscular Tipula, and in Notonecta, considering that it has both diurnal and nocturnal activity. In the laboratory, we have on several occasions observed backswimmers catching prey in the evening at twilight intensities.
The speed of the pupil mechanism in Notonecta is very slow. Light-adaptation during daytime takes about 36 min from the fully dark-to the fully light-adapted state. As a comparison, the corresponding value is 10 min in Zophobas (Ro and Nilsson, 1993b), 31 min in Tipula (Ro and Nilsson, 1994) and approximately 40 min in the giant water bug Lethocerus insulanus (Walcott, 1971), which have the same type of pupil mechanism as Notonecta. Dipterans and butterflies have a much faster type of pupil mechanism (adaptation typically takes 1−10 s), involving radial migration of pigment granules inside the retinula cells (see Kirschfeld and Franceschini, 1969; Franceschini and Kirschfeld, 1976; Stavenga et al. 1977; Land and Osorio, 1990). In these fast-flying insects, the pupil responds to the quick changes of light intensity in the environment resulting from their flying movements. In Notonecta, however, the slow pupil mechanism can hardly be involved in the adaptation of the eye when the animal is swimming swiftly in and out of the shade of leaves and other objects floating on the water surface. Instead, the slow pupil is probably tuning the eye to the general intensity of the habitat.
To understand fully the significance of the pupil mechanism in Notonecta, we must consider the endogenous circadian rhythm that sets the control range of the pupil aperture. At night, the dark-adapted aperture is fully open. During daytime, however, the aperture is only partially open when dark-adapted (see Fig. 2). From ultrastructural studies on several insect species with open rhabdoms and this type of pupil mechanism [e.g. Notonecta (Lüdtke, 1953), Tipulidae (Sotavalta et al. 1962; Williams, 1980), Lethocerus (Walcott, 1971; Ioannides and Horridge, 1975), Labidura (McLean and Horridge, 1977), Tenebrio (Wada and Schneider, 1967; Eckert, 1968), Coccinella (Home, 1975)], it has been established that the central rhabdomeres are always exposed to the incident light, whereas the peripheral rhabdomeres become more or less shielded from light when the pupil closes. Schwind et al. (1984) showed that all the peripheral rhabdomeres in Notonecta contain a visual pigment with an absorption maximum that corresponds to the greenish background light of the turbid waters in which it usually lives and that the peripheral rhabdomeres are therefore optimized for perception of contrast at low light intensities (see Lythgoe, 1972). The two central rhabdomeres, in contrast, were found to contain either two visual pigments with absorption maxima in the violet and ultraviolet (in the dorsal part of the eye) or only the ultraviolet pigment, combined with the capability of detecting polarized light (in the ventral part of the eye) (Schwind, 1983a,b, 1984, 1985). In the present study, we show that the pupil mechanism effectively contributes to the division of the Notonecta retina into a day/night duplex retina. In Fig. 6, we can see that, during the day and at normal daylight intensities, the pupil of Notonecta is more or less closed and the central rhabdomeres provide a high-intensity set of photopic photoreceptors. The night-time pupil, however, with its slightly lower sensitivity, is fully open at all normal light intensities at night, and the peripheral rhabdomeres are fully exposed to provide a low-intensity set of scotopic photoreceptors (see also Nilsson and Ro, 1994). This is an elegant way of providing optimal vision for an animal that is both diurnal and nocturnal.
It is still possible, and perhaps even likely, that the peripheral rhabdomeres are used as a green-sensitive subsystem in the day-adapted eye, but then with the same narrow acceptance angle and low sensitivity as the central rhabdomeres (see Nilsson and Ro, 1994).
The local nature of this type of pupil mechanism (see Ro and Nilsson, 1994) will result in dark adaptation of those ommatidia that are directed towards dark parts of the environment for a longer period. During the day, the circadian rhythm prevents the pupil from opening fully, thus avoiding the deleterious effects of long-lasting after-images caused by local and excessive opening of the slow pupil (see Nilsson et al. 1992). This is especially important to animals that hide in dark places during daytime, such as Tenebrio and Tipula. These animals, therefore, have a very strong and stable circadian pupil rhythm (Ro and Nilsson, 1993a, 1994). Notonecta, in contrast, does not hide during the day. This is probably the reason why the pupil rhythm in Notonecta is rather unstable (see Fig. 5). The backswimmer’s pupil rhythm seems to rely on the visual input provided by the change in light intensity at dusk and dawn. Nevertheless, the sensitivity of the pupil changes from day to night, as does the dynamic range and, to some extent, the speed.
ACKNOWLEDGEMENTS
We are extremely grateful to Dr Eric J. Warrant for critically reading the manuscript and for helping us with the Lipetz equation. The work was supported by the Swedish Natural Science Research Council.