Relaxed circadian rhythm in zooplankton along a latitudinal gradient

Oikos 11: 585 591, 7 doi: 1.1111/j.7.3-199.1575.x, Copyright # Oikos 7, ISSN 3-199 Subject Editor: Dag Hessen, Accepted 1 November Relaxed circadian rhythm in zooplankton along a latitudinal gradient Lars-Anders Hansson, Eloy Becares, Margarita Fernández-Aláez, Camino Fernández-Aláez, Timo Kairesalo, Maria Rosa Miracle, Susana Romo, Deborah Stephen, Kirsi Vakkilainen, Wouter van de Bund, Ellen van Donk, David Balayla and Brian Moss L.-A. Hansson (lars-anders.hansson@limnol.lu.se), Inst. of Ecology/Limnology, Ecology Building, SE-3 Lund, Sweden. E. Becares, M. Fernández-Aláez and C. Fernández-Aláez, Dept of Ecology, Faculty of Biology, Univ. of Léon, ES-71 Leon, Spain. T. Kairesalo and K. Vakkilainen, Dept of Ecological and Environmental Sciences, Univ. of Helsinki, FIN-151 Lahti, Finland. M. Rosa Miracle and S. Romo, Dept of Microbiology and Ecology, Univ. of Valencia, ES-1-Burjassot, Valencia, Spain. D. Stephen and B. Moss, School of Biological Sciences, Univ. of Liverpool, Liverpool, UK, L9 3BX. W. van de Bund and E. van Donk, Centre for Limnology, Netherlands Inst. of Ecology, NL-331 AC Nieuwersluis, the Netherlands. To test whether aquatic invertebrates are able to adjust their diel migratory cycle to different day length and presence of predators, we performed standardized enclosure experiments in shallow lakes at four different latitudes from southern Spain, with strong nightday cycles, to Finland where daylight is almost continuous during summer. We show here that nearly continuous daylight at high latitudes causes a relaxation in diel migratory behaviour in zooplankton irrespective of predation risk. At lower latitudes, however, similar conditions lead to pronounced diel rhythms in migration. Hence, zooplankton may show local behavioural adaptations in their circadian rhythm. They are also able to make risk assessments as to whether diel migration is beneficial or not, manifested in a lack of diel migration at near constant daylight, irrespective of predator presence. Our results provide an additional explanation to previous knowledge regarding diel migrations among aquatic invertebrates by showing that both physical (light) and biological (predation) factors may affect the migratory behaviour. Circadian rhythms, i.e. activities controlled, or assumed to be controlled, by an internal biological clock, are extremely common among animals in both terrestrial and aquatic systems. Generally the day night cycle synchronizes such internal rhythms and in the absence of such cycles, e.g. at high latitudes with continuous daylight during summer and almost continuous darkness during winter, the circadian rhythms among animals may become less pronounced (Hut et al., van Oort et al. 5). This has also been noticed in aquatic systems since sunlight determines the conditions for visual predation (Stich and Lampert 1981), suggesting that diel migration by prey organisms should be strongest when and where day night cycles are pronounced (Gliwicz 1999, 3). Diel rhythms in aquatic systems may be exemplified by the extremely widespread phenomenon of diel vertical migration (DVM) in zooplankton (Lampert 1993, Ringelberg 1999). This phenomenon has even been categorized as the biggest animal migration, in terms of biomass, on our planet (Hays 3). DVMs are generally characterized by downward migrations into deep, dark waters during day and a return to surface waters during night (Dawidowicz et al. 199, Ringelberg 1999, Hays 3). In addition to DVM, zooplankton may also perform diel horizontal migration (DHM) between a refuge, e.g. a macrophyte bed, and open water (Lauridsen et al. 199). Most studies, both in marine and freshwater systems, on diel migration in zooplankton are focused on vertical migration in deep waters, where DVM ranges over tens, or even hundreds, of meters (Cohen and Forward 5). The ultimate reason for such migrations is suggested to be predator avoidance since 585

visually feeding fish predators have a reduced feeding rate in the dark, which then functions as a refuge (Bergman 1998, Hays 3). In shallow systems, where light reaches the bottom, the refuge from visually hunting predators may instead be the sediment, macrophytes or other structures. (Timms and Moss 198, Lauridsen et al. 199, Romare and Hansson 3), whereas the open water provides algal food, but also exposes the herbivore more to predators. Hence, small animals like zooplankton have to decide whether they should migrate or not, which leads to a plastic behaviour depending on the risk of predation (Zaret and Suffern 197, Ringelberg 1999). For example, marine zooplankton may adjust their migratory behavior to day-length, i.e. they spend less time close to the surface at long, than at short day-lengths (Hays et al. 1995). However, studies on terrestrial vertebrates, for example reindeer and ground squirrel (van Oort et al. 5, Hut et al. ), suggest that if the daynight cycle is removed, the animals lose their circadian rhythm. In our study we experimentally test the hypothesis that diel migration patterns in invertebrate zooplankton are less pronounced at high latitudes, i.e. at nearly constant daylight, than at lower latitudes with more pronounced diurnal cycles. Our study is not focused on distinguishing between DVM and DHM, but rather to assess whether diel migration between a refuge and the open water differs with latitude, i.e. night-length. The basis for this hypothesis is that since migration is costly in terms of lower fecundity (Lampert 1989, Winder et al. 3), it is disadvantageous to migrate at constant, or near constant, daylight. In this way, our study may provide an additional explanation to diel migration among zooplankton. In order to test this hypothesis, we performed standardized enclosure experiments with and without fish predation. The experiment was performed simultaneously in four different places in Europe, ranging in latitude from 398N (Valencia, Spain) to 18N (Lahti, Finland). This latitudinal gradient offers a rare opportunity to compare diel utilization of refuges at nearly constant day light (Finland) with situations with more than 9 hours of complete darkness (Spain). Material and methods The experiments were performed along a latitudinal gradient comprising Lake Xeresa in Valencia, Spain (398?N, 8?W), Lake Naardermeer, the Netherlands (583?N, 581?E), Little Mere, UK (538?N, 8?W), and Lake Vesijärvi, Finland (18?N, 583?E). Experiments were carried out in enclosures filled with adjacent water containing natural levels of zooplankton, where twelve had no fish and the other twelve had fish predators and treatments were randomly assigned to enclosures. The experimental design, as well as methods were standardized among countries. Enclosures were made of polyethylene (15 mm thick) cylinders mounted top and bottom on plastic hoops, providing a enclosure volume of.79 m 3 at a depth of about 1 m. The bottom hoop was firmly buried in the sediment so the water in the enclosures was in contact with the sediment and the top of the enclosure was open to the atmosphere. The enclosures were attached to a wooden or plastic framework. The experiment started with fish additions on 8 June 1998 in all countries and in order to let zooplankton adjust to the level of predation risk in the different treatments, diel sampling was performed three weeks later (3 June to July). The fish used were small (5 mm) individuals with a total biomass (wet weight) of about g m, or a minimum of two fish individuals, per enclosure of locally appropriate zooplanktivorous species. The fish concentration was kept as low as possible in order to reduce the predation pressure, but provide the zooplankton with a continuous supply of fish kairomones, as well as a risk of predation. Inevitably, different planktivorous fish species had to be used in different locations, but previous studies (Lazzaro et al. 199) suggest that biomass and size of fish, rather than taxa, are the key features of predation on zooplankton. In Finland the fish species added was roch (Rutilus rutilus), whereas three spined sticklebacks (Gasterosteus aculeatus) was added in U.K and rudd (Scardinius erythrophthalamus) in the Netherlands, and mosquitofish (Gambusia holbrooki) was used in Spain. Enclosures were checked several times a week and occasional dead fish replaced. Details about fish species and management of experimental enclosures are given in Stephen et al. () and Moss et al. (). Although occasional houses, which may provide some dim light during night, are situated around most of the lakes investigated, they are unlikely to have affected the diel migration of zooplankton within enclosures (Longcore and Rich ). The algal biomasses (chlorophyll) were relatively similar among lakes (Table 1), suggesting that the light penetration through the water is also unlikely to have affected the results. The night lengths ranged from 9.15 h in Spain to.91 h in Finland, with the Netherlands and UK in between with about 7 h of night (Table 1). It should, however, be noted that at northern latitudes, e.g. in Finland, summer nights are never completely dark despite the sun being below the horizon ( the midnight sun ), whereas at low latitudes, e.g. in Spain, darkness is almost complete once the sun sets. This means that effects from night length are very conservative with respect to the Finnish data since there is still some light available even during night. 58

Table 1. Latitudes, night lengths (h), water temperature and chlorophyll a concentrations (mgl 1 ) at the four European experimental study sites. Dominant genera of cladocerans and copepods in each country are also shown. Spain the Netherlands UK Finland Latitude N398 N58 N538 N18 Night length (h) 9.15 7.3 7.17.91 Temp. (8C) 9. 1.9 17.. Chlorophyll a 7 5 15 11 Cladocerans Chydoridae Chydoridae Chydoridae Diaphanosoma Diaphanosoma Polyphemus Polyphemus Scapholeberis Scapholeberis Scapholeberis Sida Sida Simocephalus Simocephalus Bosmina Bosmina Chydoridae Chydoridae Ceriodaphnia Copepods Cyclops Cyclops Cyclops Cyclops Eudiaptomus Diaptomus Eudiaptomus This experiment on diel zooplankton migration between the open water and potential refuges among macrophytes, along enclosure walls or close to the sediment, constituted a separate study performed within the framework of a larger project focusing on climate effects on trophic interactions (Hansson et al., Moss et al., Stephen et al. ). Zooplankton were sampled with a plexiglass tube (length 1 m; diameter 7 mm) in the open water of each enclosure at midday (11 13 local time) and midnight (3 1, local time). Hence, at each sampling occasion about.5% of the zooplankton in each enclosure were removed, suggesting that the sampling procedure did not affect their abundance. Zooplankton was preserved and counted with a stereo microscope (Vakkilainen et al. ). Day length is here defined as the total time the sun was above the horizon and night time is, accordingly, minus day length. In all countries sampling was performed the first day/nights in July and therefore the night length was calculated for 1 July 1998. Data for calculating the night length at the different sites were obtained from http://www. qpais.co.uk/modb-iec/dayleng.htm. For calculating the overall differences in zooplankton distribution during day and night among fish treatments, we used a twoway ANOVA with predation (fish) and day/night as independent, and zooplankton abundances as dependent variables. We also performed a one way ANOVA on each fish treatment separately with day/night as independent variable. The obtained value of the F statistic from this ANOVA provides a test for the statistical robustness of the observed differences in abundances between day and night samples. Hence, we have here used the Fstatistic to describe how pronounced the difference was between zooplankton distribution during day and night. This way of using Fstatistics has previously been used in similar ways (Olson ). Results and discussion Zooplankton migration has thrilled ecologists for more than a century, although the progress has been surprisingly slow and the relatively modest consensus is that a multiplicity of factors interact to produce the behaviour (Dini and Carpenter 199). The most widespread explanations include food limitation (Geller 198), metabolic advantages (McLaren 193), light (Ringelberg 1991), and predator avoidance (Gliwicz 198, Lampert 1989, Dini and Carpenter 199). That diel vertical migration can be induced by the presence of fish or fish exudates has been shown repeatedly (Lampert 1989, Levy 199, Bollens and Frost 1991, Ringelberg 1991, Dini and Carpenter 199, von Elert and Loose 199, Nesbitt et al. 199). The zooplankton communities in our study were very similar among countries, generally dominated by cladoceran genera as Chydorus, Scapholeberis, Bosmina, Simpcephalus and cyclopoid copepods (Table 1). In order to minimize the effects of direct predation on zooplankton, we performed the diel study with low fish abundances (about g m ) and we recorded no negative effects on zooplankton abundances from fish predation (F 1,7 B/., p/.5), except for a tendency of lower cladoceran abundances in Spain and the Netherlands, suggesting that fish effects were mainly affecting behaviour of the zooplankton, and to a lesser extent abundances, which was the aim with the study. In Spain both cladocerans and copepods showed a strong diel migration with higher abundances in the open water during night than during day (Fig. 1; ANOVAs based on all fish treatments F 1,7 /7.13; 587

35 3 Copepods Spain 7 Cladocerans Spain 5 5 15 3 1 5 1 1 The Netherlands 1 The Netherlands 8 8 Abundance (l -1 ) 5 UK 3 5 UK 3 15 1 1 5 3 5 Finland 7 Finland 5 15 1 3 5 1 a b a b Fig. 1. Mean number (9/SE) of copepods (left panel) and caldocerans (right panel) during day (open bars) and night (filled bars) in the open water of enclosures without (a; n/1) and with fish (b; n/1). Experiments were performed along a latitudinal gradient including Spain, the Netherlands, UK and Finland. 588

pb/.11 and F 1,7 /8.9; pb/.5 for cladocerans and copepods, respectively). At a higher latitude (the Netherlands) the copepods still showed strong diel differences (F 1,7 /9., pb/.), whereas cladocerans showed no significant diel differences in utilisation of the open water (F 1,7 /.5; p/.8). Moving somewhat further northwards revealed a slight tendency for higher abundances of cladocerans in the open water during night (Fig. 1; F 1,7 /.5; p/.1), but less so for copepods (Fig. 1; F 1,7 /.; p /.3). At the highest latitude (Finland) the diel pattern was completely erased for both cladocerans (F 1,7 /.5; pb/.5) and copepods (F 1,7 /.5; P /.1), and especially the copepods even showed a tendency towards higher abundances in the open water during day than during night, i.e. an opposite diel migration (Fig. 1). Using the F-statistic as a proxy for how robust differences between zooplankton distributions were during day and night indicates that the most robust differences (highest F-values) are to be found where the night is longest (southernmost latitudes; Fig. ). Plotting the obtained F-statistic against the night length gave positive correlations for both copepods and cladocerans (Fig. ; r/.8 and.8, respectively). In Finland, with less than 5 h night length, the diel differences in zooplankton distribution was completely absent, or even showing a tendency to be reversed (indicated by negative y-axis values in Fig. ), for both copepods and cladocerans. Similar results, i.e. a reduced diel migration, was found by Loose (1993) when experimentally exposing Daphnia to continuous darkness, i.e. removing diel light differences. In Spain, however, where the night length is more than 9 h, differences in distribution between night and day were strong, i.e. both copepods and cladocerans were performing diel vertical migration (Fig. ). The night lengths in the Netherlands and the UK in the beginning of July are almost identical (7.3 and 7.17 h), which, accordingly, lead to a similar diel distribution for cladocerans (Fig. ), whereas copepods in the Netherlands showed much stronger diel migrations than in the UK (Fig. ). It may also be noted that the response to increased night length was stronger (steeper slope of the correlation) for copeepods than for cladocerans (Fig. ). Potential zooplankton refuges from fish predation in our study include the sediment surface, enclosure walls and the macrophytes. The size and shape of the first two were identical among countries and are therefore unlikely to have caused the differences in zooplankton behaviour. The size of the macrophyte refuge, on the other hand, differed considerably among enclosure set-ups, ranging from.9/. g m in UK to 1189/ g m in the Netherlands (means9/1sd). Finland and Spain had macrophyte biomasses of 5.59/.3 and 9.9/.5 g m. However, the macrophyte 8 Copepods 8 Cladocerans F-statistic - - - 5 7 8 9 1 Night length (h) - 5 7 8 9 1 Night length (h) Fig.. The relation between night length at different latitudes (N398 Valencia, Spain, N58, Naardmeer, the Netherlands, N538 Little Mere, UK and N18 Vesijärvi, Finland) and the F-statistic obtained from the ANOVA on differences in copepod and cladoceran distribution between day and night. The figure shows that the F-statistic, i.e. the robustness of differences in zooplankton distribution between day and night, increases with night length. Filled and open symbols denote enclosures with and without fish, respectively. Correlations for copepods are: r/.8; t /.9; pb/.5, and for cladocerans: r/.8; t / 3.85; pb/.5. 589

refuge size was not correlated with strength of zooplankton diel migrations, here expressed as the F-statistic of differences between day and night (r//., t 7 /.577; p/.1; (cladocerans) and r/.3; t 7 /1.37: p/.1 (copepods)). Hence, shape and size of the refuge were not likely to have affected the zooplankton migratory behaviour in our study. In addition to fish predation and macrophytes, also the ontogeny of zooplankton may affect their migratory behavior (Hays 1995). Despite these potentially confounding effects, the impact of latitudinal changes in light levels on zooplankton diel migations remained. Our main hypothesis was that utilisation of a diel refuge is useless if the night is too short and this hypothesis was borne out by the data showing considerable latitudinal differences in how pronounced the diel migrations among zooplankton were, that is, the differences between number of zooplankton between night and day were less pronounced in northern countries (short night) than at lower latitudes (long nights). In accordance with our results, marine copepods in the Arctic have been shown to not perform diel vertical migrations during periods of midnight sun (Blachowiak-Samolyk et al. ). The likely explanation for this pattern is that when night is short it may be less beneficial, in terms of reduced predation risk, to migrate. In a large scale experimental study where fish kairomones were added, Daphnia performed strong diel migrations (Loose 1993). However, when the animals were exposed to complete darkness, the diel migrations immediately ceased, and the Daphnia were evenly distributed throughout the water column, despite fish kairomones were still present. One explanation to these results is that the animals were protected by the dark throughout the water column and had no reason to migrate in order to hide from a potential predator. In our study, the zooplankton instead showed a similar response at nearly constant daylight, despite the presence of fish. Together, these studies may indicate that presence of predation risk is not sufficient, but a light:dark cycle is also necessary to initiate diel migration in zooplankton. Although there is nowadays a general consensus that predation is an ultimate cue in inducing diel migrations in zooplankton (Zaret and Suffern 197, Stich and Lampert 1981, Lampert 1989, Levy 199, Bollens and Frost 1991, Ringelberg 1991, Dini and Carpenter 199, Nesbitt et al. 199), our study shows that in addition to predation risk, the strength of the day night cycle is a cue for zooplankton migration. Hence, we show here that just as for some vertebrates (van Oort et al. 5), aquatic invertebrates are locally plastic in their circadian rhythm and are also able to make risk assessments regarding if diel migration is beneficial or not (Ringelberg 1999), manifested in a lack of diel migration at near constant daylight, irrespective of predator presence. References Bergman, E. 1998. Foraging abilities and niche breadths of two percids, Perca fluviatilis and Gymnocephalus cernua under different environmental conditions. J. Anim. Ecol. 57: 3 53. BlachowiakSamolyk, K. et al.. Arctic zooplankton do not perform diel vertical migration (DVM) during periods of midnight sun. Mar. Ecol. Prog. Ser. 38: 11 11. Cohen, J. H. and Forward, R. B. J. 5. Diel vertical migration of the marine copepod Calanopia americana. Twilight DVM and its relationship to the diel light cycle. Mar. Biol. 17: 387398. Bollens, S. M. and Frost, B. W. 1991. Diel vertical migration in zooplankton: rapid individual responce to predators. J. Plankton Res. 13: 1359135. Dawidowicz, P. et al. 199. Vertical migration of Chaoborus larvae is induced by the presence of fish. Limnol. Oceanogr. 35: 131137. Dini, M. L. and Carpenter, S. R. 199. Fish predators, food availability and diel vertical migration in Daphnia. J. Plankton Res. 1: 359 377. Geller, W. 198. Diurnal vertical migration of zooplankton in a temperate great lake (L. Constance): a starvation avoidance mechanism? Arch. Hydrobiol. Suppl. 7: 1. Gliwicz, Z. M. 198. Predation and the evolution of vertical migration in zooplankton. Nature 3: 778. Gliwicz, Z. M. 1999. Predictability of seasonal and diel events in tropical and temperate lakes and reservoirs. In: Tundisi, J. G. and Straskraba, M. (eds), Theoretical reservoir ecology and its applications. Braz. Acad. Sci. and Backhuys Publishers, Sao Paulo, pp. 991. Gliwicz, Z. M. 3. Between hazards of starvation and risk of predation: the ecology of offshore animals. Int. Ecol. Inst., Excellence in Ecology 1, Oldendorf/Luhe. Hansson, L.-A. et al.. Responses to fish predation and nutrients by plankton at different levels of taxonomic resolution. Freshwater Biol. 9: 1538 155. Hays, G. 1995. Ontogenetic and seasonal variation in diel vertical migration of the copepods Metridia lucens and Metridia longa. Limnol. Oceanogr. : 1115. Hays, G. et al. 1995. Spatio-temporal patterns in the diel vertical migration of the copepod Metridia lucens in the northeast Atlantic derived from the continuous plankton recorder. Limnol. Oceanogr. : 975. Hays, G. 3. A review of the adaptive significance and ecosystem consequences of zooplankton diel vertical migrations. Hydrobiol. 53: 1317. Hut, R. A. et al.. Gradual reappearance of post hibernation circadian rhythmicity correlates with numbers of vasopressin containing neurons in the suprachiasmatic 59

nuclei of European ground squirrels. J. Comp. Physiol. B. 17: 597. Lampert, W. 1989. The adaptive significance of diel vertical migration of zooplankton. Funct. Ecol. 3: 17. Lampert, W. 1993. Ultimate causes of diel vertical migration in zooplankton: new evidence for the predator-avoidance hypothesis. Arch. Hydrobiol. Beih. Ergebn. Limnol. 39: 79 88. Lauridsen, T. et al. 199. The importance of macrophyte bed size for cladoceran composition and horizontal migration in a shallow lake. J. Plankton Res. 18: 839. Lazzaro, X. et al. 199. Planktivores and plankton dynamicseffects of fish biomass and planktivore type. Can. J. Fish. Aquat. Sci. 9: 1 173. Levy, D. A. 199. Reciprocal diel vertical migration behaviour in planktivores and zooplankton in British Columbia lakes. Can. J. Fish. Aquat. Sci. 7: 175517. Loose, C. J. 1993. Lack of endogenous rhythmicity in Daphnia diel vertical migration. Limnol. Oceanogr. 38: 1837 181. Longcore, T. and Rich, C.. Ecological light pollution. Front. Ecol. Environ. : 191198. McLaren, I. A. 193. Effects of temperature on growth of zooplankton, and the adaptive value of vertical migration. J. Fish. Res. Bd. Can. : 857. Moss, B. et al.. Continental scale patterns of nutrient and fish effects on shallow wetland lakes: synthesis of a pan-european mesocosm experiment. Freshwater Biol. 9: 133 15. Nesbitt, L. M. et al. 199. Opposing predation pressures and induced vertical migration responses in Daphnia. Limnol. Oceanogr. 1: 13 1311. Olson, J. V.. Infrasound signal detection using Fisher Fstatistic. Inframatics : 1 7. Ringelberg, J. 1991. A mechanism of predatormediated induction of diel vertical migration in Daphnia hyalina. J. Plankton Res. 13: 8389. Ringelberg, J. 1999. The photobehaviour of Daphnia spp. as a model to explain diel vertical migration in zooplankton. Biol. Rev. 7: 3973. Romare, P. and Hansson, L.-A. 3. A behavioral cascade: top-predator induced behavioral shifts in planktivorous fish and zooplankton. Limnol. Oceanogr. 8: 195 19. Stephen, D. et al.. Continental-scale patterns of nutrient and fish effects on shallow lakes: introduction to a pan European mesocosm experiment. Freshwater Biol. 9: 1517 15. Stich, H. B. and Lampert, W. 1981. Predator evasion as an explanation of diurnal migration by zooplankton. Nature 93: 39398. Timms, R. M. and Moss, B. 198. Prevention of growth of potentially dense phytoplankton populations by zooplankton grazing, in the presence of zooplanktivorous fish, in a shallow wetland ecosystem. Limnol. Oceanogr. 9: 7 8. Vakkilainen, K. et al.. Responses of zooplankton to nutrient enrichment and fish in shallow lakes: a pan European mesocosm experiment. Freshwater Biol. 9: 119 13. van Oort, B. et al. 5. Circadian organization in reindeer. Nature 38: 195 19. Winder, M. et al. 3. On the cost of vertical migration: are feeding conditions really worse at greater depths? Freshwater Biol. 8: 383 393. von Elert, E. and Loose, C. J. 199. Predator-induced diel vertical migration in Daphnia: enrichment and preliminary chemical characterization of a kairomone exuded by fish. J. Chem. Ecol. : 885895. Zaret, T. M. and Suffern, J. S. 197. Vertical migration in zooplankton as a predator avoidance mechanism. Limnol. Oceanogr. 1: 8 813. 591