Hydrobiologia 491: 301–307, 2003. E. van Donk, M. Boersma & P. Spaak (eds), Recent Developments in Fundamental and Applied Plankton Research. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

301

What goes down must come up: symmetry in light-induced migration behaviour of Daphnia Erik Van Gool∗ & Joop Ringelberg Netherlands Institute of Ecology, Centre for Limnology, Amsterdam, The Netherlands. E-mail: [email protected] ∗ Present address: Population Biology Section, Institute of Biodiversity and Ecosystem Dynamics, University of Amsterdam, Kruislaan 320, 1098 SM Amsterdam, The Netherlands Received 1 August 2001; in revised form 29 April 2002; accepted 13 August 2002

Key words: predation avoidance, diel vertical migration, behavioural mechanism, phototaxis, zooplankton

Abstract During a short period of the year, Daphnia may perform a phenotypically induced diel vertical migration. For this to happen, light-induced swimming reactions must be enhanced both at dawn and at dusk. Enhanced swimming in response to light intensity increase can be elicited by fish-associated kairomone in the laboratory, if food is sufficiently available. However, during the light change at dusk the Daphnia are still in the hypolimnion, where no fish kairomone is present and both temperature and food availability is low. Still, what goes down must come up. This raises questions about how Daphnia tunes its light-induced swimming behaviour to prevailing conditions such that a normal diel vertical migration can be performed. We investigated the symmetry in behavioural mechanism underlying these diel vertical migrations in the hybrid Daphnia galeata × hyalina (Cladocera; Crustacea), with special interest for the environmental cues that are known to affect swimming in response to light increase. That is, we tested whether fish- associated kairomone, food availability, and temperature affected both swimming in response to light intensity increase and decrease similarly. We quantified swimming behaviour during a sequentially increased rate of light change. Vertical displacement velocity was measured and proved to be linearly related to the rate of the light change. The slope (PC) of the function depends on the value of the factors kairomone concentration, food availability, and temperature. The changes of the PC with kairomone concentration and with temperature were similar both at light intensity increases and decreases. The PC also increased with food concentration, although during light increases in a different way from during light intensity decreases. Low food availability inhibited swimming in response to light intensity increase, but enhanced swimming in response to light intensity decrease. Hence, ascent from the deep water layers with low food concentration at dusk is facilitated. These causal relations are part of a proximate decision-making mechanism which may help the individual Daphnia to tune migration to predation intensity and food availability.

Introduction Many prey behave so as to reduce the risk of being predated. This behaviour may have important negative effects, because prey may not be able to forage efficiently while avoiding predators (Lima, 1998). A good example of this is diel vertical migration (DVM) of zooplankton in lakes and oceans (Lampert, 1989, 1993). Zooplankton migrate to deeper water to escape predation by visually hunting juvenile fish during the

day, and return to the food-rich surface water to eat and profit from the higher temperature at night (de Meester et al., 1999). For Daphnia, the direct energetic cost of swimming distances comparable to amplitudes of migration in the field has been shown to be negligible when measured as difference in fitness (Dawidowicz & Loose, 1992a). However, several studies have shown that the lower temperature and food density at greater depths induce reduced population growth rates (Stich & Lampert, 1984; Dawidowicz & Loose,

302 1992a,b). These ecological costs may have selectively prevented Daphnia from staying continuously at large depths. In Lake Maarsseveen (The Netherlands) for instance, vertical distribution of the hybrid Daphnia galeata × hyalina, changes with the seasons. From April to the end of May, the adults are distributed nearly homogeneously within the upper 7 m both day and night. Then, during a short period of 6–7 weeks, the mean population depth at noon is much lower than at midnight. This period of DVM coincides with the appearance of large shoals of predatory juvenile perch (Perca fluviatilis) (Ringelberg et al., 1991). At the same time the thermocline develops. The thermocline divides the lake vertically in a warm epilimnion with fish and much food, and a fish-free hypolimnion, with lower temperature and food availability. Hence, migrating Daphnia reside in a totally different habitat during day and night. Field observations on Daphnia revealed also that rapid migrations occurred when relative changes in light intensity at dawn and dusk were high and a correlation between migration velocity and the rate of the light change was demonstrated (Ringelberg et al., 1991; Ringelberg & Flik, 1994). In the laboratory, light-induced swimming responses can be elicited if the combination of the rate and the duration of a relative change in light intensity surpasses a threshold (Ringelberg, 1964; van Gool & Ringelberg, 1997). It was found that a kairomone associated with juvenile perch enhanced the swimming velocity of this light-induced reaction (van Gool & Ringelberg 1995, 1998a). It was suggested that Daphnia use kairomone level as an estimate of predation risk, and thus have a ’decision-making mechanism’. Information based on predator kairomone level, but also food concentration, determines swimming velocity in response to light changes (van Gool & Ringelberg, 1998a,b). So far, research on the phenotypic induction of DVM has been focused on the swimming in response to light intensity increases. This provided insight in the behaviour at sunrise. However, daphnids have to react to light intensity decrease at sunset too. Given that the environment a daphnid dwells in at dawn and dusk differs, questions emerged on the influence of factors such as kairomone, food and temperature on the light-induced swimming reactions elicited at dawn and dusk. In a hypolimnion without fish, fish kairomone proved to be absent as well (van Gool & Ringelberg, 2002). Because kairomone-induced sensitization is maintained for some time in kairomonefree water (Ringelberg & van Gool, 1995), the ascent migration at dusk might still be sufficient. Apart from

fish kairomones, the epilimnion and the hypolimnion differ in food concentration and temperature. Since morning descent should be comparable to evening ascent, the swimming reactions elicited by increase and decrease in light intensity must be equal. Hence, factors that affect swimming in response to light increase (van Gool & Ringelberg, 1998a), should also affect the response to light decrease. Dawn and dusk are successions of accelerating and decelerating relative increases and decreases in light intensity. For an enhancement of swimming velocity by food and kairomone these accelerations are conditionally. In this study, the symmetry in light-induced swimming responses was investigated experimentally with accelerating light intensity increase and decrease at different concentrations of fish kairomone, food concentrations, and temperatures.

Methods Light-induced swimming reactions of Daphnia were studied in three experiments. In each experiment either the level of (1) fish kairomone, (2) food concentration, or (3) temperature was varied. The experimental set-up was the same for all experiments (see van Gool & Ringelberg, 1998a). A vertically positioned Perspex cylinder (height 118 cm, diameter 9.5 cm, content 8.2 l) was placed in a Perspex jacket, filled with temperature controlled tap water. The cylinder was illuminated from above by three incandescent lamps (Philips, 25 W, 60 W, 25 W). These lamps were connected to a computer-controlled, variable resistance. Illumination intensity at the top of the cylinder was 0.082 µmol m−2 s−1 . Light intensity (PAR) was measured with a Li-Cor (LI-189) and a quantum sensor. An infrared (IR) sensitive video camera, connected to a video recorder and a computer screen was used to observe Daphnia in the cylinder. IR illumination (λmax = 950 nm; 60 W) was used to make the animals visible. This IR illumination has no effect on Daphnia behaviour. During the light changes, daphnids visible on the computer screen were followed by moving the camera from the initial position. For light increases this was from about 30 cm above the middle of the tube, down to the bottom; for light decreases the camera started at a depth of 30 cm below the center and was moved up until a depth of 10 cm was reached. Observations near the top (<10 cm deep) and near

303 the bottom (>110 cm deep) were not used because boundary effects on behaviour may be expected. We used animals from one clone of the hybrid Daphnia galeata × hyalina (stock culture M2, van Gool & Ringelberg, 1995). The daphnids were always kept in water from Lake Maarsseveen. Prior to use, this lake water had been circulated over a sand filter to stimulate the bacterial breakdown of any organic components, and was filtered through a 0.45-µm acetate filter ( 14 cm) to remove debris. This handling ensured absence of an effective fish kairomone level, since no effect could not be shown in this water using a bioassay (van Gool, unpubl. results). Cultures were maintained in 1-l bottles. Every 2 weeks, half of the content was refreshed with sand-filtered lake water. Every other day, Scenedesmus acutus was added to get a food concentration of about 1 mg C/l. The Daphnia were cultured at a day–night cycle of L:D = 16:8, with light on at 06.00 h (local time). The same light regime was applied in the experimental situation. One day before an experiment started, 60 Daphnia with eggs (first or second brood) from a culture were placed in the experimental cylinder at treatment conditions. Experimental light changes were applied between 11:00 and 14:00 h to reduce an influence of a circadian rhythm on the reactivity of the daphnids (Ringelberg & Servaas, 1971). On the day of an experiment, at 09.00 h, the experimental food concentration was reestablished and the water in the cylinder was aerated. At least 1 h (1–1.5 h) before an experiment started, the daphnids in the cylinder were exposed to the initial light intensity. Each day, only one light stimulus was given. For each treatment, this procedure was repeated three to four times, using new daphnids. We expect the light-induced swimming reaction of a daphnid to be independent of the behaviour of the other animals in the cylinder. On successive video playbacks, all daphnids (four to nine) visible on the screen, were followed. Individual daphnids were traced on the computer screen and position and time were recorded simultaneously at clicks with the cursor at about 5-s intervals. The tracks of the daphnids were stored as two-dimensional coordinates (x, y). At the start of the video play-back the daphnid followed was marked on the computer screen to ensure that each individual was observed only once. This procedure resulted in 20–28 individual behavioural responses per treatment. Fish-associated kairomones were obtained from one juvenile perch (Perca fluviatilis, approx. 4 cm length), which was kept in a 10-l aquarium with sandfiltered water from Lake Maarsseveen. Each day, half

the content of this aquarium was refreshed and faeces removed. The perch was daily fed with a small amount of Chironomus larvae. Fish kairomone level present in the experiments represented 0, 7, and 15% of the concentration in the aquarium. Food concentration (S. acutus) in the experimental cylinder was made 0.5 mg C/l. This concentration exceeds the incipient limiting level above which ingestion rates of Daphnia are constant (ILL = 0.26 mg C/l; Muck & Lampert, 1984). In the experiments on the effect of food availability S. acutus was used at four concentrations: 0.05, 0.2, 0.5, and 1.0 mg C/l. No fish kairomone was present in these experiments. Kairomone and food experiments were carried out at 17±0.5 ◦ C. The effect of temperature was studied at four different temperatures: 12, 17, 23 and 26 ◦ C. In these experiments kairomone was absent and the algal food concentration was 0.5 mg C/L. An experimental light change started with a relative change in light intensity; the initial light intensity (Io ) was changed for t = 0.33 min at a constant rate or stimulus value (It = Io · eRCL1∗t ). Then, in sequence, the rate was increased stepwise after 0.33 min by 10% (RCLi = 1.1 * RCLi−1 ). The first relative change had a value of RCL1 = 0.08 min−1 for increases in light intensity and RCL1 = −0.08 min−1 for decreases in light intensity. Vertical displacement velocity (DV) was defined as the vertical distance a daphnid moved, divided by the time needed to cover this distance. DV was measured over 20-s time intervals (moving average), thus separately for each stimulus value during the complete period in which the light intensity changed. The displacement velocity is a linear function of the stimulus (DV = PC * RCL; Ringelberg, 1993), even if the stimulus is increased stepwise over time (van Gool & Ringelberg, 1998a). The slope of this function was called the Phototactic Coefficient (PC) and was calculated for each individual Daphnia separately. Because PC is a measure of the rate by which the displacement velocity increases with the stimulus strength, we supposed that a functional relationship would exist between PC and the factors kairomone level, food concentration and temperature, during increases as well as decreases in light intensity. Therefore, regression analyses were performed. For the first two factors, a linear function described the relation adequately. The temperature data were better described by a quadratic function. Statistical analysis was performed with Systat (SAS Institute, Inc.) and StatWorks (Cricket Software).

304 Results All observed Daphnia responded to the light change in each of the three treatments. Displacement velocity increased with increased stimulus strength and the regression coefficient (PC) of the function between DV and RCL depended on the treatments. Fish kairomone had a significant effect on PC (Fig. 1). The regression function for light intensity increases was PC(incr) = 44.40 + 2.13 * Kairomone level (H0 , constant = 0, P < 0.001; and H0 , regression coefficient = 0, P < 0.001; N = 67; R2 = 0.29) and for light intensity decreases PC(decr) = 50.26 + 1.68 * Kairomone level (H0 , constant = 0, P < 0.001; and H0 , regression coefficient = 0, P < 0.001; N = 64; R2 = 0.29). Regression analysis was performed on all data and a considerable variation was present. In Figure 1, this variation is presented as 95% confidence limits of the mean. Apparently, the direction of the light change, i.e., increasing or decreasing intensity, had no effect on the PC-value at any given kairomone level (two-way ANOVA, F1,127 = 0,23, P = 0.64). Therefore, the regression function was calculated for all data together; PC(all) = 48.94 + 1.92 * Kairomone level (H0 , constant = 0, P < 0.001; and H0 , regression coefficient = 0, P < 0.001; N = 133; R2 = 0.22). This regression line has been inserted in Figure 1. Under the artificial light change, kairomone level affected the PC, but the light-induced swimming reaction was symmetrical (at light increase and decrease) with regard to the effect of fish kairomones. Food concentration influenced PC also significantly (Fig. 2). The regression functions were PC(incr) = 30.45 + 29.26 * Food concentration (H0 , constant = 0, P < 0.001; and H0 : regression coefficient = 0, P < 0.001; N = 88; R2 = 0.14), and PC(decr) = 44.01 + 18.12 * Food concentration (H0 , constant = 0, P < 0.001; and H0 , regression coefficient = 0, P < 0.001; N = 98; R2 = 0.13). For the factor food, however, the regression coefficients seem to be different for increases and decreases in light intensity (H0 , b(incr) = b(decr), P < 0.025). Again, a considerable part of the total variation was due to differences in PC within groups (food concentration). At the lowest food concentrations the mean PC was significantly smaller at light intensity increases than at light decreases (Student t-test, P = 0.007). Also at the next food concentration (0.2 mg C/l) the difference was significant (P = 0.035). Both regression lines intersected at a food concentration of 1.22 mg C/l.

Figure 1. The average value (± 95% confidence limits) of the Phototactic Coefficient (PC) at different kairomone levels (K). The open dots represent the observations at light intensity increases and the closed dots represent the observations at light intensity decreases. See text for regression functions based on all data. The open squares represent the values of similar experiments done in 1997 (van Gool & Ringelberg, 1998a; PC = 52.88 + 2.65 * K; R2 = 0.96, N = 4) and are included in the graph for comparison.

Figure 2. The average value (± 95% confidence limits) of the Phototactic Coefficient at different food concentrations. The open dots represent the observations at light intensity increases, the closed dots represent the observations at light intensity decreases. See text for regression functions based on all data.

Temperature (T) also influenced the swimming in response to relative changes in light intensity (Fig. 3). The best fit was obtained with a quadratic model. The regression functions were PC(incr) = −49.647 + 7.808 * T −0.119 * T2 (P < 0.001, N = 91, R2 = 0.754) and PC(decr) = −64.131 + 9.905 * T −0.190 * T2 (P < 0.001, N = 91, R2 = 0.843). No significant difference between the direction of light change and the temperature could be determined (two-way ANOVA, F3,178 = 0.63, n.s.). A perfect least-square regression was obtained when all data were averaged per tem-

305 perature. This led to the function: PC = −57.393 + 8.923 * T −0.156 * T2 (R2 = 1.000, N = 4). In the middle temperature range, the three functions were nearly identical, but at temperatures higher than 20 ◦ C, the quadratic terms of the separate functions are too small to fit the last two PC values. At a temperature of 7.8 ◦ C (and of 49.2 ◦ C, but this root value is theoretical because Daphnia cannot survive at this high temperature) the phototactic coefficient PC equals zero, which suggests that there is no relation between the RCL and the DV anymore.

Discussion Seasonal migrations in a vertically stratified lake require adaptations to cope with this spatially diverse environment. Flexible and situation dependent behaviour may be an answer. Nevertheless, the behaviour needs input from the environment, and the animals must act upon the variable and depth dependent information. DVM behaviour causes Daphnia to reside in quite different environments during day and night. We show how flexible light-induced swimming reactions, incorporating informational input from the environment, enables it to perform these daily migrations. However, the environmental factors we studied convey quite different types of information on the environment. Fish kairomone is an information conveying chemical without the nutritional or energetic value food has. In several models, food, as well as predators, are considered a key factor that determine the benefit of DVM (Gabriel & Thomas, 1988; Gabriel, 1993; Fiksen, 1997; Han & Straskraba, 1998). So, food may also convey information if it makes an assessment of the nutritional value of the environment possible. By what means this information can be obtained is unknown but an obvious suggestion is that a satiation (or starvation)-related variable, like gut fullness or blood sugar concentration, plays a role. Another suggestion is that the information is derived, as is the case with fish kairomones, from infochemicals, excreted by algae. Evidence for a perception of chemicals produced by algae, is available in Daphnia (van Gool & Ringelberg, 1996; Laurén-Määttä et al., 1997). A priori, we assumed the effect of environmental factors on the light-induced swimming reactions of Daphnia to be symmetrical. It was known that the influence of these environmental factors on the swimming reactions is conditional to accelerations in the relative increase in light intensity (van Gool & Ringel-

Figure 3. The average value (± 95% confidence limits) of the Phototactic Coefficient (PC) at different temperatures (T, ◦ C). The open dots represent the observations at light intensity increases, the closed dots represent the observations at light intensity decreases. See text for the quadratic regression functions calculated for all data. The line was fitted through the averages of PC for increases and decreases in light intensity: PC = −57.39 + 8.92 * T – 0.156 * T2 (N = 4, R2 = 1).

berg, 1997, 1998a). In this study we showed that Daphnia displayed similar phenotypic plasticity in swimming in response to light intensity decrease when stimulus accelerations were present. In the absence of these accelerations, the vertical displacement velocity was only marginally affected by these factors (van Gool & Ringelberg, 1995, 1997). The regression coefficient of the relation between DV and RCL (the Phototactic Coefficient PC) was lower than about 30, in the presence as well as absence of fish kairomone (van Gool & Ringelberg, 1995). For our clone of D. galeata × hyalina, this value of about 30 might be considered the lower limit of PC at a temperature of 17◦C. If during an experiment with accelerations in stimulus strength, PC would have this value, this would mean that Daphnia only responded to the individual relative changes in light intensity and displacement velocity would not specifically be increased. An example is given in Figure 2 for increases in light intensity in the absence of kairomone and at low food concentration. In the presence of fish kairomones, the phototactic coefficient was always larger, with enhancements (PC(i) /30) by a factor 2–3 at increasing as well as decreasing light intensities (Fig. 1). Light-induced swimming reactions in Daphnia depend on temperature (Fig. 3). The quadratic temperature function explains more of the variation than the linear models of the food and kairomone factors do. The phototactic coefficient PC increases rapidly from zero between 7.8 and 20 ◦ C, even in the absence of

306 fish kairomones. Consequently, we suppose that some diel vertical migration might develop in the course of the summer season, when water temperature increases, even in the absence of predating fish. The food concentration of 0.5 mg C/l, at which the temperature experiments were done, is high and not often encountered in non-eutrophied lakes. At concentrations below the incipient limiting level of 0.26 mg C/l, PC is smaller than 35. If we suppose that the temperature curve starts at a similar zero-value as the one presently found, then, for lake temperatures lower then 18 ◦ C, PC is much smaller than 35. These PC values realise small migration amplitudes, only. Ringelberg (1999) argued that the adaptive value of these small scale migrations is prevention of UV damage rather than predation prevention. On the other hand, for a kairomone concentration of 20%, for example, and for the same temperature range, PC ranges from 30 to 90, thus potentially leading to large amplitude vertical migrations. Of course, a definite quantification of PC and of the resulting vertical migrations is only possible if the fish kairomone can be quantified in absolute units and if experiments are extended to kairomone and food concentrations at different temperatures. We were interested in the symmetry of the phototactic mechanism, since at dawn and dusk the same distance must be covered by migrating animals. Symmetry in positive and negative phototaxis is present if kairomone and temperature are the variables. Food acts asymmetrically: at low concentrations, downwards swimming to increases in light intensity is not enhanced, but upwards swimming for light intensity decreases is. This remarkable phenomenon was also found by Forward & Hettler (1992) for brine shrimp larvae (Artemia) photoreactivity. They described, that starvation did not affect the negatively phototactic response, enhanced in the presence of fish predators, but activated the positive response to decreases in light intensity. After the better part of the day in the food-poor hypolimnion, a return to the algae-rich epilimnion in the evening is facilitated by this effect of the food–light decrease combination. A qualitative verbal model may illuminate the complexity of descent and ascent diel migration. At dawn, relative increases in light intensity accelerate rapidly, and if kairomone level is sufficiently high and food concentration is not too low, a rapid descent from the predation zone is possible, supported by the high temperature of the epilimnion. If food concentration is very low, as, for example, in 1991 in Lake Maarsseveen, diel vertical migration does not occur (Flik &

Ringelberg, 1993). Downwards displacement may be predominantly passive, as in 1990, or active as in 1992 (Ringelberg & Flik, 1994). At the time the metalimnion and the hypolimnion is reached, the acceleration in relative changes in light intensity have changed into decelerations and in combination with the lower temperature, downwards displacement slows down, or ceases altogether if the threshold for phototaxis is reached at about sunrise. At dusk, relative decreases in light intensity increase slowly and the threshold for phototaxis is reached at about sunset. Temperature is low so ascent is slow and might have the character of an upward drift. The absence of fish kairomone in the hypolimnion and metalimnion is not really a problem, because sensitization has been maintained during the day (Ringelberg & van Gool, 1995). The danger zone is slowly reached but once in the epilimnion swimming velocity increases because in the mean time higher relative changes in light intensity occur and temperature in the epilimnion is higher. Time is an hour after sunset and meanwhile light intensity has decreased considerable. Therefore, the danger of being eaten by visually hunting fish predators is reduced. Still other factors influence the pattern. For example, the threshold relative change depends on absolute light intensity (Ringelberg et al., 1967) and thus is different in the morning than in the evening. The rapid shifts in genotypes during a period of diel vertical migration in Lake Maarsseveen (Spaak & Ringelberg, 1997) suggests that not all genotypes induce migration behaviour equally effective. All factors and circumstances are variable, in seasonal time and from lake to lake. They act in concert and thus realise a variety of migration patterns. These patterns can be explored with a physiologically based mechanistic model (Ringelberg, 1995). The comparison of different habitat configurations can provide further understanding of the adaptiveness of diel vertical migration.

Acknowledgements EvG was supported by contract ENV4-CT97-0402 within the framework of the European Commission’s Environment and Climate Programme and was part of the project network WAtER (Wetland and Aquatic Ecosystem Research). The paper was written when EvG was at the University of Amsterdam. We thank Mous Sabelis, and two anonymous reviewers for comments on a previous version of this paper.

307 References Dawidowicz, P. & C. J. Loose, 1992a. Costs of swimming by Daphnia during diel vertical migration. Limnol. Oceanogr 37: 659–665. Dawidowicz, P. & C. J. Loose, 1992b. Metabolic costs during predator induced diel vertical migration of Daphnia. Limnol. Oceanogr. 37: 1589–1595. de Meester, L., P. Dawidowicz, E. van Gool & C. J. Loose, 1999, Ecology and evolution of predator-induced behavior of zooplankton: depth selection behavior and diel vertical migration in Daphnia. In: Tollrian, R. & D. Harvell (eds), The Evolution of Inducible Defences. Princeton University Press, Princeton, NJ. Fiksen, O., 1997. Allocation patterns and diel vertical migration: modeling the optimal Daphnia. Ecology 78: 1446–1456. Flik, B. J. G. & J. Ringelberg, 1993. Influence of food availability on the initiation of diel vertical migration (DVM) in lake Maarsseveen. Arch. Hydrobiol. 39: 57–65. Forward, R. B. Jr. & W. F. J. Hettler, 1992. Effects of feeding and predator exposure on photoresponses involved in diel vertical migration of brine shrimp larvae. Limnol. Oceanogr. 37: 1261–1270. Gabriel, W., 1993. Models on diel vertical migration. Arch. Hydrobiol.39: 123–136. Gabriel, W. & B. Thomas. 1988. The influence of food availability, predation risk, and metabolic costs on the evolutionary stability of diel vertical migration in zooplankton. Verh. int. Ver. Limnol. 23: 807–811. Han, B-P. & M. Straskraba, 1998. Modeling patterns of zooplankton diel vertical migration. J. Plankton Res. 20: 1463–1487. Lampert, W., 1989. The adaptive significance of diel vertical migration of zooplankton. Funct. Ecol. 3: 21–27. Lampert, W., 1993. Ultimate causes of diel vertical migration of zooplankton: new evidence for the predator-avoidance hypothesis. Arch. Hydrobiol. 39: 79–88. Laurén-Määttä, C., O. T. Kleiven & J. Kiviranta, 1997. Horizontal distribution of Daphnia pulex in response to toxic and non-toxic algal extracts. J. Plankton Res. 19: 141–148. Lima, S. L., 1998. Stress and decision making under the risk of predation: recent developments from behavioural, reproductive, and ecological perspectives. Adv. Study Behav. 27: 215–290. Loose, C. J. & P. Dawidowicz, 1994. Trade-offs in diel vertical migration by zooplankton: the costs of predator avoidance. Ecology 75: 2255–2263. Muck, P. & W. Lampert, 1984. An experimental study on the importance of food conditions for the relative abundance of calanoid copepods and cladocerans. Arch. Hydrobiol. Suppl. 66: 157–179. Ringelberg, J., 1964. The positively phototactic reaction of Daphnia magna Straus: A contribution to the understanding of diurnal vertical migration. Neth. J. Sea Res. 2: 319–406.

Ringelberg, J., 1993. Phototaxis as a behavioural component of diel vertical migration in a pelagic Daphnia. Arch. Hydrobiol. 39: 45–55. Ringelberg, J., 1995. An account of a preliminary mechanistic model of swimming behaviour in Daphnia; its use in understanding diel vertical migration. Hydrobiologia 307: 161–165. Ringelberg, J., 1999. The photobehaviour of Daphnia spp. as a model to explain diel vertical migration in zooplankton. Biol. Rev. 74: 397–423. Ringelberg, J. & B. J. G. Flik, 1994. Increased phototaxis in the field leads to enhanced diel vertical migration. Limnol. Oceanogr. 39: 1855–1864. Ringelberg, J. & E. van Gool, 1995. Migrating Daphnia have a memory for fish kairomones. Mar. Freshwat. Behav. Physiol. 26: 249–257. Ringelberg, J. & H. Servaas, 1971. A circadian rhythm in Daphnia magna. Oecologia 6: 289–292. Ringelberg, J., B. J. G. Flik, D. Lindenaar & K. Royackers, 1991. Diel vertical migration of Daphnia hyalina (sensu latiori) in Lake Maarsseveen: Part 1. Aspects of seasonal and daily timing. Arch. Hydrobiol. 121: 129–145. Ringelberg, J., J. V. Kasteel & H. Servaas, 1967. The sensitivity of Daphnia magna Straus to changes in light intensity at various adaptation levels and its implication in diurnal vertical migration. Z. Physiol. 56: 397–407. Spaak, P. & J. Ringelberg, 1997. Differential behaviour and shifts in genotype composition during the beginning of a seasonal period of diel vertical migration. Hydrobiologia 360: 177–185. Stich, H. B. & Lampert, W., 1984. Growth and reproduction of migrating and non-migrating Daphnia species under simulated food and temperature conditions of diurnal vertical migration. Oecologia 61: 192–196. van Gool, E. & J. Ringelberg, 1995. Swimming of Daphnia galeata × hyalina in response to changing light intensities: influence of food availability and predator kairomone. Mar. Freshwat. Behav. Physiol. 26: 259–265. van Gool, E. & J. Ringelberg, 1996. Daphnids respond to algaeassociated odours. J. Plankton Res. 18: 197–202. van Gool, E. & J. Ringelberg, 1997. The effect of accelerations in light increase on the phototactic downward swimming of Daphnia and the relevance to diel vertical migration. J. Plankton Res. 19: 2041–2050. van Gool, E. & J. Ringelberg, 1998a. Light-induced migration behaviour of Daphnia modified by food and predator kairomones. Anim. Behav. 56: 741–747. van Gool, E. & J. Ringelberg, 1998b. Quantitative effects of fish kairomones and successive light stimuli on downward swimming responses of Daphnia. Aquat. Ecol. 32: 291–296. van Gool, E. & J. Ringelberg, 2002. Relationship between fish kairomone concentration in a lake and phototactic swimming by Daphnia J. Plankton Res. 24: 713–721.