Population dynamics of Daphnia spp. and implications for trophic interactions in small, monomictic lake

In Lake Oglethorpe, Georgia, USA, herbivorous crustacean zooplankton are abundant and dominate zooplankton biomass in winter, but are scarce throughout most of the summer. We used a 3.5 year study of Daphnia population dynamics to infer when food, predators or temperature constrained growth of this population. Transitions between winter and summer consumer assemblages are concurrent with seasonal changes in water temperature, thermal structure (stratification/destratification), resources (autotrophic/heterotrophic-domained production), and predator abundance and activity (e.g. Lepomis macrochirus and Chaoborus punctipennis. We sampled at weekly or less intervals from April 1992 to September 1995, and determined population abundances for all cladoceran species. For the Daphnia population (Daphnia ambigua + Daphnia parvula, we measured clutch size and length for all individuals. We used average water column temperature (where dissolved oxygen is >1 mg l^-1) to estimate egg development time from an empirical model. Estimates of Daphnia population birth and death rate were thus generated from abundance, egg ratio and temperature/dissolved oxygen data. We compared observed birth rate (bobs) with expected birth rate (bexp>95% CI; predicted for food-saturated conditions at ambient temperature). For variable (1-13 week) periods between later November and March, 1992-1995, water temperature was the primary factor constraining Daphnia population growth (bobs = bexp). From about April to early November, bobs < bexp suggested food-limited population growth. In spring, summer and early fall (March-October), population densities were several-fold lower than in late fall and winter (November-February). However, all else being equal, egg ratio and population birth rate data would have predicted that Daphnia abundance fluctuates over equivalent ranges in spring and fall. We interpret this discrepancy as evidence for increased rates of extrinsic mortality during the growing season and a seasonal shift in the relative importance of resource and predator regulation. The duration of predator suppression of crustacean population abundance in Lake Oglethorpe and other warm-latitude lakes (36N-27°S) is longer (3-7 months) than that observed in north temperate lakes (1-2.5 months; 41-52°N).      

Causes of Daphnia midsummer decline in two deep meromictic subalpine lakes

1. Daphnia are key organisms in pelagic food webs, acting as a food resource for fish and predatory zooplankton and regulating phytoplankton through grazing. Its population dynamic follows regular seasonal patterns, with spring peaks followed by summer population declines (midsummer declines, MSDs). Midsummer declines show high inter-annual variation, which has been attributed to different causes. However, the mechanisms controlling the MSD remain poorly understood, especially in deep stratified lakes.
2. We tried to disentangle the factors causing Daphnia MSDs in Lake Lugano and Lake Iseo (in Switzerland and Italy), two deep peri-alpine lakes with similar trophic status and vertical mixing dynamics, characterised by phosphorus accumulation in the hypolimnion and variable mixing during late-winter turnovers.
3. Specifically, we assessed the effects of three different hypothetical pathways according to which: (1) winter air temperature controls MSDs by influencing mixing depth during turnovers and epilimnetic phosphorus replenishment; (2) vernal air temperature influences MSD by accelerating the timing of spring population peak; and (3) summer temperature influences MSDs by increasing fish predation. We assessed the relative strength of these pathways using structural equation modelling on long-term datasets for the two lakes (29 years for Lake Lugano and 19 years for Lake Iseo).
4. Between the hypothesised pathways, the one driven by winter air temperature (through P replenishment) influenced Daphnia abundance in spring in both lakes, but the effects propagated to summer Daphnia abundance only in Lake Lugano. Additionally, summer Daphnia abundance was influenced by the summer air temperature through a positive (although weak) effect. By comparison, vernal air temperature had no detectable effects on summer Daphnia abundance.
5. The results revealed marked differences between the meromictic study lakes and the shallow hypertrophic water bodies that were the focus of previous research on Daphnia MSD, and also between the two study lakes. The influence of epilimnetic P replenishment on the summer Daphnia abundance in Lake Lugano, which was recovering from past eutrophication, may have reflected the greater susceptibility of deep, stratified lakes to P depletion after spring compared to shallow hypertrophic lakes or reservoirs. This effect might not have been detected in Lake Iseo because P was more consistently depleted during the study period (i.e. variance in the predictor was too low to detect an effect).
6. This study highlighted the complexity of the effects of climate variability on Daphnia MSD in deep lakes, showing that the responses can differ even between two neighbouring lakes with similar vertical mixing dynamics and trophic status. At the same time, the results suggest that future increases in winter air temperature, caused by global warming, may cause critically low densities of Daphnia during spring and summer and compromise the ability of zooplankton to control phytoplankton biomass.

     

Seasonal variation of egg size and number in a Daphnia pulex population

Seasonal variation of egg size and number was examined in a Daphnia pulex population inhabiting a vernal pond. In this population, size at maturity declines at midseason, probably as an adaptive response to size-selective predation by larvae of the salamander Ambystoma. The larger early season individuals produce more and larger eggs than the smaller late season individuals. Age at maturity does not vary between seasons. Laboratory experiments indicate that temperature may affect egg size, egg number and size at maturity. However, field data suggest that temperature accounts for only a small fraction of the total variation in egg size and number. Indirect measures of nutrition indicate that food limitation does not cause the seasonal decline in egg size and number. The seasonal change in reproductive traits is well correlated with changes in invertebrate and vertebrate predation. Examination of predator feeding preferences and their impact on Daphnia mortality indicate that variation of reproductive traits is most likely a complex adaptation to changing predation regimes.     

Effects of food limitation,notonectid predation,and temperature on the population dynamics of Daphnia carinata

The relative contributions of invertebrate predation (Notonectidae: Anisops spp.), food limitation, and certain abiotic factors in driving the population dynamics of Daphnia carinata were studied for a two year period in a large farm dam in southern Victoria, Australia. Detailed measurements were made on the population densities of Daphnia and Anisops spp., the amount of food available, the nutritional status, and the size-specific fecundity of Daphnia. The density of the Daphnia population at the field site oscillated closely with water temperature. The amplitude of the population fluctuations varied seasonally, being much greater during the warmer months of the year and switching to fluctuations with low peaks when water temperature dropped below approximately 15°C. Anisops spp. densities were greatest in winter and declined during the spring of each year. Nymphs appeared in late spring and early summer and numerically dominated the population. During the warmer periods of the year, the daphnid population went through a series of rapid growth phases leading to over-exploitation of food resources and subsequent population collapses. Daphnia population densities were not correlated with Anisops numbers suggesting that predation was not a major regulatory factor during the warmer periods of the year. When water temperatures fell below ca. 15°C daphnid population densities remained low despite high food levels. During this period the impact of Anisops may have been greater. Two distinct phases were identified: a warm water period when food limitation was the main regulatory factor, and a cool water period when Anisops predation may have been the paramount factor. Low oxygen concentrations were associated with heavy rainfall in the spring and may have had a limiting effect on Daphnia for short periods. Daphnia may have had an important role in sustaining the Anisops population in the pond over each winter.     

Can overwintering versus diapausing strategy in <Emphasis Type="Italic">Daphnia</Emphasis> determine match–mismatch events in zooplankton–algae interactions?

Mismatches between predator and prey due to climate change have now been documented for a number of systems. Ultimately, a mismatch may have far-reaching consequences for ecosystem functioning as decoupling of trophic relationships results in trophic cascades. Here, we examine the potential for climate change induced mismatches between zooplankton and algae during spring succession, with a focus on Daphnia and its algal food. Whereas the development of an overwintering population of daphnids may parallel shifts in phytoplankton phenology due to climate warming, changes in the photoperiod–temperature interaction may cause the emerging population of daphnids to hatch too late and mismatch their phytoplankton prey. A decoupling of the trophic relationship between the keystone herbivore Daphnia and its algal prey can result in the absence of a spring clear water phase. We extended an existing minimal model of seasonal dynamics of Daphnia and algae and varied the way the Daphnia population is started in spring, i.e., from free swimming individuals or from hatching resting eggs. Our model results show that temperature affects the timing of peak abundance in Daphnia and algae, and subsequently the timing of the clear water phase. When a population is started from a small inoculum of hatching resting eggs, extreme climate warming (+6°C) results in a decoupling of trophic relationships and the clear water phase fails to occur. In the other scenarios, the trophic relationships between Daphnia and its algal food source remain intact. Analysis of 36 temperate lakes showed that shallow lakes have a higher potential for climate induced match–mismatches, as the probability of active overwintering daphnids decreases with lake depth. Future research should point out whether lake depth is a direct causal factor in determining the presence of active overwintering daphnids or merely indicative for underlying causal factors such as fish predation and macrophyte cover. Priority program of the German Research Foundation—contribution 5.     

Climate-driven warming during spring destabilises a Daphnia population: a mechanistic food web approach

Temperature-driven changes in interactions between populations are crucial to the estimation of the impact of global warming on aquatic food webs. We analysed inter-annual variability in two data sets from Bautzen reservoir, Germany. In a long-term data set (1981–1999) we examined the pelagic phenology of Daphnia galeata, a keystone species, the invertebrate predator Leptodora kindtii, phytoplankton and Secchi depth in relation to water temperature and the North Atlantic Oscillation index. In a short-term data set (1995–1998) we examined food web relations, particularly the consumption of D. galeata by young-of-the-year (YOY) percids and L. kindtii and rates of population change of D. galeata (abundance, recruitment pattern and non-consumptive mortality). The start of the clear-water stage (CWS) was correlated with winter temperatures. It started 5.8 days earlier per degree warming after warm winters (mean January–March temperature ≥2.5°C) compared to cold winters (mean temperature <2.5°C). However, the end of the CWS remained relatively constant. Predation by L. kindtii and YOY percids on D. galeata started distinctly earlier, i.e. by 13.0 and 6.5 days per degree warming, respectively, in years when the average May temperature was high (≥14°C) compared to years when it was low (<14°C). Significant reductions of Daphnia abundance in midsummer occurred only in years in which the mean May temperature exceeded 14°C. This temperature regime resulted in a match of over-exploitation of food resources by Daphnia during the CWS and strong predation by YOY percids and L. kindtii. Consumptive mortality increased at higher rates with a rise in temperature than net recruitment, resulting in lower Daphnia densities at the end of the CWS. Our data suggest that even low warming by 1.7°C during a short, but critical seasonal period, resulting in the coincidence of two or more factors adversely affecting a keystone species, such as Daphnia, may induce changes in whole lake food webs and thus alter entire ecosystems. Priority programme of the German Research Foundation—contribution 9.     

Individual Energetics and the Equilibrium Demography of Structured Populations

This paper considers the relationship between the demographic mechanisms of population control and the energetics of the individuals who comprise the population. We examine the equilibrium properties of a class of structured population models in which individuals compete for some environmental resource (such as food) and demonstrate that population demography is independent of the nature of the feedback loop which establishes the equilibrium state. We thus derive general insights into the influence exerted by individual energetic and allocation strategies on population average demographic characteristics. We show that models of individual energetics which produce apparently very similar predictions at the individual level can result in very different behaviour at the population level. In addition, we observe that different models of individual mortality can imply marked differences in population demography and that the common assumption of constant mortality can be responsible for potentially unrealistic model behaviour. Our results emphasise the substantial data requirements for parameterising and testing individual-based models.     

Population dynamics and production of Daphnia hyalina Leydig and Daphnia cucullata Sars in Tjeukemeer

As part of a research programme on the food chains in Tjeukemeer, the Daphnia hyalina and Daphnia cucullata populations were studied for three successive years. To analyse the factors regulating the production of these two species, their population parameters (density, size distribution, fecundity) and population dynamics (birth rate, mortality rate) were studied and related to environmental factors. Since Daphnia in Tjeukemeer shows continuous recruitment, the population dynamics model INSTAR was developed and used to integrate field data with laboratory data on development rates and length-weight relationships. The dynamics of the Daphnia species are mainly regulated by temperature and fish predation, the latter affects both birth rate and mortality. Total annual Daphnia production was 3.1–6.9 g org. dry wt M^–2, and annual P/B ratio ranged from 25 to 40 for D. cucullata and from 45 to 49 for D. hyalina.     

Application of individual growth and population models of Daphnia pulex to Daphnia magna, Daphnia galeata and Bosmina longirostris

Daphnia models for individual growth and population dynamics have been developed in the manner of models developed by Gurney, McCauley, Andersen and others. All or most of the earlier models were parameterized for Daphnia pulex; we have used the D. pulex model as a baseline model for other species of Daphnia such as magna, galeata and also Bosmina longirostris. Because of the lack of ample data for D. magna, D. galeata and B. longirostris, some of the physiological data had to be relied on the other species whose data were available and in some case calibrated. We were able to produce reasonable results for individual growth as well as population dynamics under the controlled laboratory conditions. Most of the results were compared with the available laboratory data for population as well as growth. All the simulations have been done under high and low food concentrations. The animals are assumed to be feeding on green algae (Chlamydomonas reinhardtti) under the laboratory conditions of 18–20°C. The continuous growth until the end of the life was observed in smaller B. longirostris, whereas rapid growth in the beginning and slower after the start of the reproduction was observed in Daphnia species. The smaller species matured earlier than larger species. B. longirostris population sustained better than Daphnia species in medium food concentrations.     

Significance of a low oxygen layer for a Daphnia population in Lake Yunoko,Japan

Population dynamics and vertical migration of Daphnia longispina in Lake Yunoko were studied. The Daphnia population was small in spring and early summer, probably because of high predation pressure by fish. The population grew in midsummer, when thermal stratification developed and the dissolved oxygen became very low in the deeper layer of the hypolimnion. In this season, adults of D. longispina concentrated in the daytime near the lake bottom, where fish were absent because of the anoxic conditions, but ascended at night to the upper layer of the hypolimnion, where food was most abundant. The low oxygen layer near the bottom kept out the predators and protected Daphnia from predation, and consequently contributed to the built-up of its population. However, the low oxygen layer was unfavorable for reproduction of Daphnia, as reflected in the low egg ratio and high percentage of males in the population. The population decreased in the fall, when thermal stratification disappeared and predation pressure seemed to increase.     

Life history of Daphnia galeata in a hypertrophic reservoir and consequences of non-consumptive mortality for the initiation of a midsummer decline

SUMMARY 1. Field and laboratory investigations were combined in a 2‐year study on the initiation of a midsummer decline of Daphnia galeata Sars in a hypertrophic reservoir. Quantitative field samples were taken twice a week, and, adult and juvenile mortality rates were calculated. Patterns of reproduction and survival of daphnids born during spring and early summer under fluctuating food conditions were determined in life‐table experiments. 2. The abundance of Daphnia increased strongly in early May and declined in June 1998 (midsummer decline). In 1999, Daphnia density increased only slowly in spring and remained constantly high throughout the summer. 3. Food conditions (concentrations of POC_{<30 μm}) for daphnids deteriorated in both years in response to increasing Daphnia densities, resulting in a clear‐water phase of about 4 weeks. When Daphnia abundance declined in 1998, POC_{<30 μm} concentrations increased greatly, whereas in 1999 food conditions improved only slightly and Secchi depth remained high. 4. Survival of daphnids in life‐table experiments decreased greatly after food became rare and was strongly reduced in those animals born during the clear‐water phase compared with those born later. In addition, age at first reproduction was retarded during the clear‐water phase, resulting in very low population growth rates. Survivorship patterns in life‐table experiments suggest a strong impact of non‐consumptive mortality on Daphnia population dynamics. 5. Field data of mortality point to differences in mortality patterns between years, probably resulting from different predation impacts of juvenile fish. In both years, however, adult mortality contributed substantially to overall mortality at the end of the clear‐water phase. As bottom‐up effects on D. galeata were very similar in both years, the significance of non‐consumptive mortality on the initiation of midsummer declines appears to depend largely on recruitment patterns before the clear‐water phase. A high impact can be expected when Daphnia populations are dominated by a peak cohort of nearly identical age during the clear‐water phase.     

The North Atlantic Oscillation and plankton dynamics in two European lakes –- two variations on a general theme

Long‐term data on water temperature, phytoplankton biovolume, Bosmina and Daphnia abundance and the timing of the clear‐water phase were compared and analysed with respect to the influence of the North Atlantic Oscillation (NAO) in two strongly contrasting lakes in central Europe. In small, shallow, hypertrophic Müggelsee, spring water temperatures and Daphnia abundance both increased more rapidly than in large, deep, meso/oligotrophic Lake Constance. Because of this, the clear‐water phase commenced approximately three weeks earlier in Müggelsee than in Lake Constance. In Müggelsee, the phytoplankton biovolume during late winter/early spring was related to the NAO index. In Lake Constance, where phytoplankton growth was inhibited by intense downward mixing during all years studied, this was not the case. However, in both lakes, interannual variability in water temperature, in Daphnia spring population dynamics and in the timing of the clear‐water phase, were all related to the interannual variability of the NAO index. The Daphnia spring population dynamics and the timing of the clear‐water phase appear to be synchronized by the NAO despite large differences between the lakes in morphometry, trophic status and flushing and mixis regimes, and despite the great distance between the lakes (～700 km). This suggests that a great variety of lakes in central Europe may possibly have exhibited similar interannual variability during the last 20 years.     

Feeding behaviour of major 0 + fish species in a shallow, eutrophic lake (Tjeukemeer, The Netherlands)

The forage base and the food selectivity of 0+ representatives of six abundant freshwater fish species were studied in a shallow, eutrophic Dutch lake. Most species relied on the zooplankton; the size-selective predation in early summer was directed to the smaller copepods and in late summer to larger cladocerans and copepods than concurrently present in the lake. Daphnia spp. and cyclopoid copepods were the main zooplankton taxa for smelt, perch and pikeperch. Energetically, the large cladoceran, Leptodora kindtii, was especially important for pikeperch. Bream and roach preyed upon smaller zooplankton than the other fish species. The influence of the zooplankton predation by abundant 0+ fish was clear from a small mean Daphnia size in September; this size is to be used as an indicator in fishery management. Neomysis integer, the most important macrofauna species, was consumed by perch, pikeperch and ruffe; pikeperch was most size-selective in this respect. The 0+ ruffe was à typically benthivorous fish. Only the 0+ pikeperch became piscivorous, especially in years when smelt was abundant.     

Impact of food concentration on diel vertical migration behaviour of <Emphasis Type="Italic">Daphnia pulex</Emphasis> under fish predation risk

Vertical migration of Daphnia represents the best-studied predator-avoidance behaviour known; yet the mechanisms underlying the choice to migrate require further investigation to understand the role of environmental context. To investigate the optimal habitat choice of Daphnia under fish predation pressure, first, we selected the individuals exhibiting strong migration behaviour. The animals collected from the hypolimnion during the daytime were significantly larger, being more conspicuous, and in turn performed stronger diel vertical migration (DVM) when exposed to fish cue. We called them strong migrants. Second, we provided the strong migrant D. pulex with food at high and intermediate (1 and 0.4 mg C l⁻¹, respectively) levels, which were well above the incipient limiting level and of high quality. They traded the benefits of staying in the warm water layer and moved down to the cold water in response to fish cue indicating fish predation. The availability of food allowed the animals to stay in the cold hypolimnion. However, at the low food level (0.1 mg C l⁻¹), which is an additional constraint on fitness, Daphnia moved away from the cold hypolimnion. Poor food condition resulted in strong migrant Daphnia to cease migration and remain in the upper warmer water layer. Although temperature is known to be a more important cost factor of DVM than food, our results clearly show that this is only true as long as food is available. It becomes clear that food availability is controlling the direction of vertical positioning when daphnids experience a dilemma between optimising temperature and food condition while being exposed to fish cue. Then they overlook the predation risk. Thus, the optimal habitat choice of Daphnia appears to be a function of several variables including temperature, food levels and fish predation. Handling editor: S. I. Dodson     

Food quality triggers the reproductive mode in the cyclical parthenogen <Emphasis Type="Italic">Daphnia</Emphasis> (Cladocera)

Cyclic parthenogenesis (heterogony) is a widespread reproductive mode found in diverse taxa such as digenean trematodes, gall wasps, gall midges, aphids, cladocerans and rotifers. It is of particular interest as it combines the advantages of asexual reproduction (rapid population growth) and sexual reproduction (recombination). Usually sexual reproduction is initiated when, or slightly before, environmental conditions deteriorate, and often results in the production of resting stages. The optimal timing of diapause induction must thus be under strong natural selection. Using the cladoceran Daphnia as a model system, we show here for the first time that the switch from parthenogenetic to sexual reproduction in a cyclical parthenogenetic organism can be influenced by the chemical composition of food. Under crowding conditions Daphnia reproduced parthenogenetically with subitaneous eggs when fed the algal species Cryptomonas sp., but started the production of resting eggs when fed with the green algal species Scenedesmus obliquus. Supplementation experiments with lipids and especially proteins showed that the induction of resting egg production in two clones of different Daphnia species was due to a dietary deficiency in the green alga. Hence, the low food quality induced a switch in the reproductive mode that may contribute to optimal timing of the sexual reproduction of Daphnia in nature. Furthermore, our results have two other major implications: first, they suggest that protein compounds should be added to the list of diet constituents potentially limiting or influencing Daphnia reproduction. Second, we show that the role of food quality goes far beyond the up to now documented effects of food quality on somatic growth and trophic transfer efficiency of herbivores: due to its effects on sexual reproduction and the production of resting eggs, food quality might influence genetic diversity and long-term persistence of Daphnia in lakes. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Life history consequences of sterol availability in the aquatic keystone species <Emphasis Type="Italic">Daphnia</Emphasis>

The absence of essential biochemical nutrients, such as polyunsaturated fatty acids or sterols, has been considered as a mechanism determining trophic interactions between the herbivore Daphnia and its phytoplankton food source. Here, we experimentally quantify the sensitivity of two Daphnia species to decreasing amounts of dietary sterols by measuring variations in life history traits. The two species Daphnia magna and D. galeata were fed different mixtures of the sterol-containing green alga Scenedesmus obliquus and the sterol-free cyanobacterium Synechococcus elongatus; a higher proportion of Synechococcus in the food is equivalent to a decrease in dietary sterols. To address the significance of sterol limitation, the Daphnia species were also fed Synechococcus supplemented with cholesterol. In both species, somatic and population growth rates, maternal dry mass, the number of viable offspring, and the probability of survival were significantly reduced with the lower availability of sterols. A high correlation between the sterol content of the mixed diet and the somatic and population growth rates was found, and growth on cholesterol-supplemented Synechococcus fitted well into this correlation. Somatic growth of first-clutch neonates grown on 100% Synechococcus exhibited a pattern similar to that of somatic growth of their mothers grown on the different food regimes, which demonstrated the significance of maternal effects for sterol-limited population growth. Daphnia galeata had a twofold higher incipient limiting sterol level than D. magna, which indicated interspecific differences in sterol requirements between the two Daphnia species. The results suggest a strong impact of dietary sterols on life history traits and therefore, population dynamics of the keystone species Daphnia.     

Relative importance of competition with Daphnia (Cladocera) and nutrient limitation on Anuraeopsis (Rotifera) population dynamics in a laboratory study

• 1 In the laboratory, the growth and reproduction of Anuraeopsis fissa were measured when fed on Scenedesmus species grown in nutrient‐sufficient, nitrogen‐limited and phosphorus‐limited media and in the presence or absence of one adult Daphnia longispina per vial.
• 2 Poor food quality may reduce the effect of competition on rotifers. Competition from Daphnia was stronger with nutrient sufficient algae than with nutrient limited algae. P‐limitation significantly reduced Anuraeopsis population growth rate and fecundity. The effect of nutrient limitation on Anuraeopsis was stronger than that of competition with Daphnia. The Anuraeopsis population declined with P‐limited food in both the presence and absence of Daphnia.
• 3 Exploitative competition by Daphnia on Anuraeopsis was stronger in the nutrient‐sufficient treatment than in the N‐limited one. Density, fecundity and population growth rate of Anuraeopsis were negatively affected by Daphnia in the nutrient‐sufficient treatment, while only fecundity was reduced by Daphnia in the N‐limited treatment. Consequently, in the N‐limited treatment, mortality should be lower in the presence of Daphnia. This result could suggest that Anuraeopsis lives longer when short of nitrogen.
• 4 Nutrient limitation may affect to the competitive interactions between zooplankton species. P‐limitation decreased the quality of algae as food for Anuraeopsis while N‐limitation decreased the susceptibility of this rotifer species to exploitative competition by Daphnia.

     

Finding the optimal vertical distribution: behavioural responses of Daphnia pulicaria to gradients of environmental factors and the presence of fish

1. The vertical distribution of zooplankton results from active habitat choice aiming to optimise fitness gain in a system of trade‐offs. 2. Using large, controlled indoor mesocosms (Plön Plankton Towers), we monitored the behavioural response of Daphnia pulicaria to vertical gradients of temperature, food, oxygen and light, in the presence and absence of fish predation. 3. In the absence of fish, Daphnia distributed as predicted by an ideal ‘free distribution with costs’. If the food was distributed homogeneously, they stayed in the warm epilimnion, while they balanced their time dwelling in epi‐ and hypolimnion if the food was concentrated in a deep‐water maximum. 4. However, oxygen depletion in the hypolimnion, representing an additional cost, prevented Daphnia from completely exploiting the hypolimnetic food maximum. Consequently, the proportion dwelling in the hypolimnion was larger if oxygen was not limiting. 5. Fish predation had an overwhelming effect, driving Daphnia into the hypolimnion under all experimental conditions. If permitted by oxygen availability, Daphnia used the whole hypolimnion, but oxygen depletion reduced their possible habitat to the upper hypolimnion with oxygen concentrations above c. 0.7 mg L^?1. As fish were less tolerant of low oxygen, the layer below the thermocline formed a predation refuge for Daphnia.