Temperature-Induced Pupil Movements in Insect Superposition Eyes

Animals active in a broad range of light intensities need a mechanism to adapt the eye to the different intensities they may encounter over a 24 h period. This adaptation to the ambient illumination usually involves two main mechanisms:

(i) changes in the transduction gain in the photoreceptor cells, and (ii) a pupil mechanism that absorbs or scatters the incoming light and regulates the amount of light reaching the retina (Warrant and McIntyre, 1990; for a review, see Autrum, 1981). In insects with superposition eyes, a number of different types of pupil mechanisms are known, all based on movements of screening pigment granules (for a review, see Warrant and McIntyre, 1996). In moths, the most common pupil type is the ‘longitudinal pupil’ (see Fig. 1A) in which pigment granules located in specialised pigment cells migrate between two extreme positions along the borders of the ommatidia, thus functioning as a pupil (Tuurala, 1954; Höglund, 1963a,b, 1966). In Agrotis spp., only pigment granules located in the secondary pigment cells contribute to the pupil mechanism (Tuurala, 1954; Warrant and McIntyre, 1996). At low light intensities at night (dark-adapted eye), the granules are retracted to a position between the crystalline cones (and out of the clear zone), allowing light from a large number of facets to reach each rhabdom. At high light intensities during the day (light-adapted eye), the pigment migrates proximally into the clear zone and isolates the ommatidia from one another. This reduces the number of facets superimposing light on each rhabdom, which in turn protects the retinula cells from excessive illumination. The migration of the pigment granules depends mainly on the ambient illumination; the higher the intensity, the further the pigment migrates into the clear zone and the greater the reduction in the amount of light reaching the retina (Höglund, 1963a,b, 1966).

Besides light intensity, temperature has also been shown to influence pigment migrations in a wide range of arthropods (butterflies, Srinivasan et al., 1977; crustaceans, Meyer-Rochow and Tiang, 1979, 1982; Meyer-Rochow, 1982; King and Cronin, 1994a,b; flies, Järemo Jonson, 1995; moths, Day, 1941; Weyrauther, 1988, 1989). A possible explanation for the influence of low temperatures on pigment migration is the depolymerization of microtubules, which is known to play an important role in pigment movement. Even though high temperatures are known to cause pigment migrations in some species, this phenomenon has not been thoroughly investigated and no functional explanation of the mechanism has been proposed.

Even though temperature is known to be a strong stimulus for pigment migration, no studies of the entire course of events resulting from temperature exposure have been made. We have studied the nocturnal turnip moth Agrotis segetum to investigate and describe quantitatively the role of temperature in pupil control. We have found that the pupil can be opened and closed solely by changes in temperature even in the absence of light. Closure appears to involve two processes: one that initiates pupil closure and a second that is responsible for complete closure.

The experiments were performed on male turnip moths, Agrotis segetum Schiff. (Noctuidae: Lepidoptera). Specimens were obtained from the Department of Ecology (Lund, Sweden), where they were reared from eggs to adults under a 17 h:7 h light:dark photoperiod in constant conditions of temperature (20–23 °C) and relative humidity (70 %). The adult males were used within 24 h of eclosion.

For comparison, two other nocturnal insect species, lacewings Euroleon nostras (Myrmeleonidae: Neuroptera) and codling moths Cydia pomonella (Tortricidae: Lepidoptera), were examined to determine whether temperature had a general influence on pupil response. E. nostras were collected as larvae from Öland, Sweden, and reared in the laboratory, and C. pomonella were obtained as pupae from the Department of Agriculture (Alnarp, Sweden). Both species were reared to adulthood in the laboratory at 19–23 °C under the same light cycle as A. segetum.

Many superposition eyes possess a reflecting layer called the tapetum lucidum positioned just behind the retina. The tapetum increases the sensitivity of the eye since light that has passed once through the retina is reflected for a second absorption. Light not absorbed by the retina after reflection emerges from the eye and is seen as a bright glow (eye glow) when viewed from the direction of illumination (Fig. 1B,C). The brightness of the glow is an indication of the location of the screening pigments and thus of the state of adaptation: a bright glow indicates an aggregation of pigments in the dark-adapted position, a dim glow indicates pigment dispersal to the light-adapted position (Höglund, 1963a,b, 1966). We used the intensity of the eye glow to determine the position of the screening pigments.

The experimental apparatus was an ophthalmoscope, constructed from a Zeiss photomicroscope with an epi-illumination attachment (Fig. 1D). The light source used for the eye glow measurements was a 20 W halogen lamp filtered with a far-red 770 nm Melles Griot wideband interference filter (bandwidth 40 nm). This far-red observation light is not visible to the moth and does not affect the pupil mechanism. The objective lens was a Zeiss 4×/0.10 epi-illumination objective.

Under CO₂ anaesthesia, the animals were mounted in plastic tubes and immobilised with wax before being placed in a dark box to obtain total darkness during the experiments and thus eliminate all effects of light. The short exposure to CO₂ did not seem to affect the pupil mechanism. The animal dark-adapted normally after anaesthesia, and in the experiments on the circadian rhythm the pupil did not show any signs of time shift compared with the entrained rhythm. The dark box was thermo-isolated to maintain constant temperature and was equipped with a thermocouple probe (pt-100 FC 1020, Heraeus). The relationship between temperature and voltage output from the probe was determined using a Fluke 51 thermometer with an accurate probe (Anritsu) for calibration. The temperature inside the box was adjusted by passing the air current through a heat-exchange plate sandwiched between two Peltier elements and then through the box. The apparatus allowed control of the temperature in the range between −5 and +45 °C.

The circadian rhythm of pupil activity was investigated by recording eye glow reflectance from animals kept in complete darkness and at constant temperature (17 °C) during the experiment. The reflectance was measured and recorded for 600 ms every 2 min for at least 60 h. A larger voltage response indicated a brighter eye glow and a more dark-adapted pupil. To eliminate the small possibility that the far-red observation light contributes to the closure of the pupil, a shutter was placed between the light source and the eye. This permitted illumination of the eye only during measurements of the eye glow reflectance. The illuminated area of the eye was adjusted using a diaphragm (located in the light path) to the same size as that of the pupil.

The reflected light from the eye was detected with a photomultiplier (Hamamatsu R2949HA), and the transduced voltages were collected on a Macintosh IIci computer. The eye glow results were analysed using Labview and Cricketgraph software.

The animals were mounted and then dark-adapted for at least 60 min at 20 °C prior to the experiment to obtain maximal eye glow. This adaptation was made during the last part of the light period of the entrained light cycle, and the experiments were performed during the dark period of the light cycle. After dark-adaptation and temperature-adaptation, the animals (still in darkness) were exposed to increasing temperatures in steps of 0.5 or 2 °C (see Results), and the eye glow reflectance from the eye was recorded as a voltage response for 600 ms every 10 s during the experiment. The experiment generally lasted for 5–6 h, but was always completed well before the end of the dark period. The pupil continued to respond normally even after having been exposed to temperatures exceeding 35 °C, well beyond the range of temperatures used in our experiments.

During the experiments, we measured the temperature of the air surrounding the moth inside the box and assumed this temperature to be the same as that of the eyes and head of the moth, although this is not necessarily the case. To check this, we measured the temperatures of the eyes, the head and the box in 5 °C steps from 10 to 30 °C using a Fluke 51 thermometer with accurate probes (Anritsu and Thermocoax). The statistical analyses of the temperatures were performed using Statview software. These temperatures were found not to differ significantly in the range between 10 and 25 °C (Student’s t-test; P_eye>0.1, P_head>0.7, N=5). At 30 °C, however, the eye and head temperatures were slightly lower than the ambient temperature in the box (Student’s t-test; P_eye=0.014, P_head=0.020, N=5).

The pupil mechanism in Agrotis segetum is not only controlled by light but is also under the influence of an endogenous circadian rhythm (Fig. 2). Even though the moths were tested in total darkness and at constant temperature (approximately 20 °C), the pupil opened and closed as if still experiencing a normal light cycle, opening during the previously entrained night and closing during the entrained day. The amplitude of the pupil response decreased throughout the experiment, indicating that the pigment granules did not reach the extreme positions and that the strength of the circadian influence decreases with time. To minimise the effect of this circadian rhythm, all temperature experiments were performed during the dark period of the light cycle.

After dark adaptation at low temperatures (5–10 °C), animals were exposed to a series of increasing temperatures (in steps of approximately 2 °C). Fig. 3A shows that the animals, although in complete darkness, responded to a change in temperature by altering the position of the screening pigment granules (seen as a change in eye glow intensity). The pupil responded over a wide range of temperatures, and the pigment granules migrated proximally to a more light-adapted position when exposed to either low or high temperature. The pupil does not respond to low temperature as strongly as it does to high temperature. Exposure to low temperature does not close the pupil completely, and the closure does not involve a rapid phase like that induced by high temperature (T_c). In the range between 21.0 and 22.7 °C, the pupil is effectively fully open (i.e. produces an eye glow greater than 97.5 % of maximum) and the pigments are located distally between the crystalline cones (Fig. 3B).

The temperature-induced pigment migration in A. segetum shows two distinct phases initiated at two different temperature thresholds. The animals were dark-adapted and temperature-adapted (at approximately 20 °C) until a maximal glow was obtained after approximately 80 min. The temperature was then increased by 0.5 °C every 20 min. If pigment migration occurred during this 20 min period, the time was extended until the eye glow intensity, and thus the pigment position, was stable again. The temperature was held constant during final pupil closure, and the experiment was terminated when the pigment had reached the light-adapted position and all reflectance from the eye had disappeared.

Figs 4 and 5 show two distinct closures: a partial closure at23.8±0.7 °C (the activation temperature, T_a; mean ± S.D., N=10) and a second fast and complete closure at 25.7±1.2 °C (the closure temperature, T_c; mean ± S.D., N=10). The two thresholds are significantly different (Wilcoxon signed rank test, P=0.005). When the temperature exceeds the first threshold (T_a), the onset of pigment migration is triggered, and this may ultimately result in closure of the pupil. By plotting the relative response versus temperature, it is possible to determine accurately the activation temperature T_a. Such plots reveal a distinct reduction in eye glow intensity at a specific activation temperature. T_a was found by regression-fitting straight lines to the data on each side of the reduction in eye glow and finding the temperature at which they intersect (Fig. 5B). All temperatures exceeding T_a induce some pigment migration, but a temperature only slightly higher than T_a usually causes a slow, partial and rather unpredictable pupil closure. As temperature continues to rise above T_a, the pupil closes faster and further. Exposing the animal to temperatures exceeding T_c results in a rapid closure of the pupil, and this stimulus is strong enough to keep the pupil closed, even though temperatures just fractions of a degree below T_c close the pupil only slightly. T_c thus seems to be a very sensitive trigger for pupil closure.

The pupil mechanisms of two other nocturnal insects, the lacewing Euroleon nostras and the codling moth Cydia pomonella, are also influenced by temperature. Like A. segetum, both species were strongly influenced by temperature and responded to increasing temperature with a closure of the pupil (Fig. 6A,B). The pigment migration occurred in the same range of temperatures in all three species investigated, indicating that temperature-dependent pupil closure is a general mechanism in nocturnal insects.

Although light is the general stimulus for activating insect pupils, our results show that temperature alone is a stimulus strong enough to cause pupil closure in Agrotis segetum and also in the lacewing Euroleon nostras and in the codling moth Cydia pomonella. A possible explanation for the temperature-dependence of the pupil mechanism in A. segetum is the presence of two processes: a temperature-sensitive process activated at T_a, leading to a second process at T_c that is responsible for complete pigment migration. The activation temperature, T_a, is rather constant among individuals, indicating that the eye contains a temperature-sensitive structure with a threshold at 23.8±0.7 °C. As T_a is exceeded, a chemical and/or structural process is initiated which induces pigment migration. As temperature increases towards T_c, a second process begins, and pigment migration in the eye is completed. This second process is not necessarily sensitive to temperature, but may be dependent only on the first process. There may not even be two separate mechanisms, but merely two phases of the same process, resulting in this two-step closure. Frixione and Pérez-Olvera (1991) demonstrated that light-induced pigment migration in crayfish is a two-step process that includes a first all-or-nothing phase that results in a migration approximately half-way towards the light-adapted position, followed by an adaptation phase regulating the illumination of the photoreceptors more accurately. A pupil mechanism involving a two-step process, which is sensitive to temperature, has not been noted before. However, the nature of this mechanism is unknown.

The response to increasing steps of temperature seen in A. segetum is quite different from that shown by the mealmoth Ephestia kuehniella (Weyrauther, 1988). In E. kuehniella, the pigment position seems to be directly dependent on temperature and does not show the same kind of sudden closure seen at T_c in A. segetum. Instead, the pupil in E. kuehniella closes in a stepwise manner in response to a certain plateau level for each step increase in temperature. A. segetum shows pigment dispersal that results in a closed pupil when exposed to temperatures above T_c, but a rather unpredictable pigment position when the animal is kept at temperatures between T_a and T_c. Not only is the position of the pigment influenced by temperature, but also the speed of pupil closure: lower temperatures open and close the pupil more slowly than higher temperatures (P. Nordström and E. J. Warrant, personal observations). This seems to be a rather common response in insects (Srinivasan et al., 1977; Järemo Jonson, 1995). All species examined in this paper (A. segetum, E. nostras and C. pomonella) show the same response when exposed to high temperature: a dispersal of the screening pigments to a light-adapted position. These results are consistent with work on other insects (butterflies, Srinivasan et al., 1977; flies, Järemo Jonson, 1995; moths, Weyrauther, 1988) and crustaceans (amphipods, Meyer-Rochow and Tiang, 1979; isopods, Meyer-Rochow, 1982), indicating a widespread mechanism responsible for pupil closure in arthropods exposed to high temperatures. What has not been demonstrated before is the range of temperatures within which this closure occurs and the temperatures that are most important. Interestingly, all three species investigated here respond very similarly and within a narrow temperature range. This suggests that they all share the same temperature-sensitive mechanism controlling pigment migration. Whether this response has evolutionary significance or whether it is merely a side-effect of some temperature-sensitive process breaking down at T_c, we cannot say. However, the fact that it occurs at approximately 25 °C has the beneficial side effect of closing the pupil at high temperatures and thus pre-adapting the eye for high light intensities.

Even though the function of the longitudinal pupil pigment migrations in superposition eyes is fairly well understood, not much is known about the mechanisms behind these migrations. Although the mechanisms proposed for pupil pigment migration in apposition eyes (King and Cronin, 1994b) and neural superposition eyes (Järemo Jonson, 1995) are not directly comparable with those in superposition eyes, we discuss two important factors that may also be involved in superposition eyes.

In many arthropods, a large number of microtubules are arranged in the direction of movement of the migrating pigments, thus forming a massive framework that may function as a guiding track during screening pigment transport (butterflies, Meinecke, 1981; crayfish, Frixione et al., 1979; Frixione and Tsutsumi, 1982; flies, Järemo Jonson, 1995; Limulus polyphemus, Miller, 1975). Other cytoskeletal elements, such as actin filaments, are often present in the photoreceptor cells and may also be involved in this pigment transport (Arikawa et al., 1990; Baumann, 1992; Järemo Jonson, 1995). King and Cronin (1994b) showed that mantis shrimps Gonodactylus oerstedii close the pupil when exposed to low temperatures and that this response may be due to structural changes in the eye because a lowered temperature decreases the density of microtubules. However, it is not known whether the microtubules in the eye are affected by high temperature. It is therefore possible that pupil closure at high temperature is not dependent on microtubule disruption, but that two different mechanisms are responsible for pupil closure at high and low temperatures. Microtubules have been shown to be present in the secondary pigment cells in moths (Meinecke, 1981), but whether they are involved in pigment migration is still unclear.

Ca²⁺ is very important in many phototransduction processes in invertebrates (for a review, see Dorlöchter and Stieve, 1997) and is also known to play a significant role in pigment migration in compound eyes (butterflies, Weyrauther 1989; crayfish, Frixione and Aréchiga, 1981; King and Cronin, 1996; flies, Kirschfeld and Vogt, 1980; Howard, 1984). High temperatures may increase membrane permeability in the crayfish retina, resulting in an efflux of Ca²⁺ from intracellular reservoirs that causes the pigment to migrate (Meyer-Rochow and Tiang, 1982). A high level of intracellular Ca²⁺ is generally understood to induce pigment migrations. In E. kuehniella, however, eyes treated with EGTA responded with dispersal of the pigments and closure of the pupil, indicating that the response to Ca²⁺ is the opposite of that found in flies and crayfish (Weyrauther, 1989). If the response seen in E. kuehniella is general for all superposition eyes, this would indicate that the mechanisms responsible for pigment migration in superposition eyes are quite different from those in apposition eyes and neural superposition eyes.

Agrotis segetum is a nocturnal animal active around midnight from June to September with a peak of activity from June to July (Persson, 1971; Esbjerg, 1987). During the Swedish summer, the temperature rarely exceeds 20 °C after sunset but normally ranges between 9 and 17 °C, and the moth is very unlikely to experience ambient temperatures in the vicinity of T_a during its nocturnal activity. During experiments, the pupil of A. segetum is fully open in darkness between 21.0 and 22.7 °C and closes at both higher and lower temperatures, as we have already described. However, at the lower temperatures we measured, the pupil did not close entirely.

Most flying insects produce large quantities of heat when flying (Heinrich, 1993). Even though the temperature of the head usually remains some degrees below that of the thorax, it usually remains some 5–15 °C above ambient temperature during flight (Hegel and Casey, 1982; May and Casey, 1983; May, 1995). However, since the experimental animals were constrained in plastic tubes and unable to move their wings during the experiment, they were therefore unable to alter their body temperature to the same extent as during natural flight and warm-up. This explains the mismatch between the range of temperatures in which the pupil is fully open (21.0–22.7 °C) and the range of temperatures over which A. segetum is normally active (9–17 °C). When the animals are active and flying, the head temperature should reach a level close to that inducing a maximally open pupil. The temperature measured during the experiments describes the air temperature in the dark box and not the temperature of the eye itself; in the range 10–25 °C, these temperatures are not significantly different.

In addition to external stimuli, such as light and temperature, the pupil mechanism in A. segetum is under the influence of an endogenous circadian rhythm. This is common in nocturnal animals hiding in dark refuges during the daytime (Stavenga, 1977; Ro and Nilsson, 1993a,b, 1994, 1995; Warrant and McIntyre, 1996). In terrestrial environments, high temperature is normally correlated with daytime and hence with high light intensities. A. segetum may use temperature as a backup to the regular light input received by the pupil mechanism to obtain a more reliable entrainment of the circadian rhythm. This rhythm pre-adapts the eye for different light intensities during the day and night, by either aggregating or dispersing the pigment. Light adaptation during the daytime, even in darkness, prevents the retina from being exposed to high light intensities should the animal be forced to leave its dark refuge during bright daylight (Ro and Nilsson, 1993a; Warrant and McIntyre, 1996). The regulation of the pupil mechanism in Agrotis segetum therefore seems to be influenced by a number of inputs which interact to optimise vision at different light intensities and temperatures.

We are extremely grateful to Erling Jirle at the Department of Ecology, University of Lund, for supplying us with animals. We wish to thank Lars Ebbesson for helping us with the statistical analyses. We also thank Marie Dacke, Anna Gislén, Almut Kelber and Dan-Eric Nilsson for critically reading the manuscript. This work was financially supported by the Swedish Natural Science Research Council. P.N. also wishes to thank the Kungliga Fysiografiska Sällskapet in Lund and Västgöta Nation in Lund for financial support.