Abstract

Parasites can influence host population dynamics and genetic structure; thus, it is crucial to study parasite distribution among and within host populations. We explored spatial infection patterns in Daphnia longispina complex populations from 11 reservoirs, habitats characterized by longitudinal gradients in abiotic and biotic conditions. We hypothesized that sites with different environmental conditions should have divergent infection patterns. We sampled three stations along each reservoir (upstream, middle and downstream) in the summer and autumn of 2 years. Four groups of parasites were analysed for their spatial variation: the protozoan Caullerya mesnili, the yeast Metschnikowia sp., microsporidia and oomycetes. The microsporidia were evenly distributed between seasons, as well as among- and within-reservoirs. In contrast, C. mesnili and Metschnikowia sp. were more prevalent in autumn than in summer, whereas the prevalence of oomycetes was higher in summer. On a within-reservoir scale, C. mesnili was significantly more prevalent in upstream parts of the reservoirs, whereas Metschnikowia sp. and oomycetes were more prevalent in the central and downstream areas. In summary, the within-reservoir distribution of certain Daphnia parasites seems to be affected by environmental gradients (most probably food abundance and predation pressure). Thus, environmental heterogeneity should be considered in future studies of host-parasite interactions, even on a local scale.

INTRODUCTION

Diseases are often spatially structured, with parasite prevalence differing drastically among populations of a single host species. This has been shown across a variety of host systems, for example, in birds (e.g. Moyer et al., 2002), fish (e.g. Karvonen et al., 2005), humans (e.g. Foley et al., 2003) and plants (e.g. Carlsson-Graner and Thrall, 2002). Spatial variation in parasite pressure may lead to differences in host population dynamics (e.g. Hudson et al., 1992) and host genetic structure (Summers et al., 2003). The observed variation in parasite prevalence across geographically separated populations can be caused by differences in local biotic and abiotic environmental conditions (e.g. Ebert et al., 2001), often operating at multiple scales (e.g. Johnson et al., 2007).

Individuals within a single host population may also be exposed to different environmental conditions and, consequently, to different risks of infection. For example, the prevalence of avian malaria in blue tits varied between woodland sections, with local estimates of prevalence ranging from 10 to 60% over as little as 1km (Wood et al., 2007). Furthermore, statistical analyses revealed complex associations between avian malaria infection and both landscape and host characteristics. In another study, on perennial plants, sites with high and low fungus infection were observed only a few metres apart (Ovaskainen and Laine, 2006). This heterogeneous infection pattern was caused by differences in the availability of sheltered conditions which favoured local parasite survival, particularly over winter (Ovaskainen and Laine, 2006). If parasites are heterogeneously distributed within a single host population, this may lead to differences in the strength of selection pressure. Also, host life history traits may be affected, as variation in parasite load can influence the allocation of limited resources between host defences and other fitness components (Moyer et al., 2002).

Good examples of host populations facing heterogeneous environments are small planktonic crustaceans of the genus Daphnia (Cladocera) inhabiting large canyon-shaped reservoirs in Central Europe. In such habitats, environmental gradients are pronounced in both horizontal and vertical dimensions (Straškraba, 1998). River inflow, the main source of nutrients for primary producers, contributes substantially to the emergence of longitudinal (horizontal) gradients. High nutrient load in upstream parts of the reservoirs, which gradually declines to nutrient-poorer conditions downstream, is then reflected by changes in phytoplankton densities (e.g. Seda and Devetter, 2000). Further, fish are more abundant in upstream areas of the reservoirs (e.g. Prchalova et al., 2008). These factors jointly affect the distribution, taxonomic composition and size structure of zooplankton (Urabe, 1990; Seda et al., 2007b), including Daphnia (Petrusek et al., 2008). At the same time, vertical environmental gradients in stratified standing waters include changes in temperature, light intensity and nutrient and oxygen concentrations. These vertical clines affect the distribution of primary producers (phytoplankton) as well as other members of the trophic cascade (Lampert and Sommer, 1999). If such environmental gradients lead to differences in parasite prevalence, then these habitats provide a unique opportunity to study the effects of gradients in parasite prevalence on host population dynamics and genetic structure over a small geographical scale.

Daphnia populations of Central European reservoirs are dominated by members of the Daphnia longispina complex (Seda et al., 2007b; Petrusek et al., 2008), taxa often infected by a variety of parasite species (Wolinska et al., 2007). Furthermore, it has been shown in Daphnia-microparasite laboratory studies, that parasite virulence and spread are influenced by environmental conditions (reviewed in Wolinska and King, 2009). Thus, we hypothesized that sites with different environmental conditions should have divergent patterns of parasite prevalence. In particular, knowing the importance of resource availability (e.g. Hall et al., 2009) and strength of predation pressure (e.g. Duffy and Hall, 2008) on parasite prevalence in aquatic systems, these should result in heterogeneous parasite distribution along the horizontal as well as vertical lake axes. We tested this prediction by assessing the spatial patterns of parasite prevalence for four groups of microparasites infecting reservoir Daphnia.

METHOD

Study sites

The study was carried out in 11 canyon-shaped reservoirs located in the Czech Republic (their basic limnological parameters and location are given in Supplementary data, Table S1). All the reservoirs are elongated with a retention time >25 days, ensuring the development of horizontal environmental gradients (Straškraba, 1998). In addition, the depth near the reservoir dam is sufficient to develop temperature stratified conditions with epi-, meta- and hypolimnion (Petrusek et al., 2008). The position of the reservoirs in respect to other water bodies makes any substantial import of Daphnia hosts or their parasites unlikely. Daphnia populations in the reservoirs studied consist mainly of the D. longispina complex: Daphnia galeata, Daphnia longispina longispina, Daphnia cucullata and their respective hybrids (Seda et al., 2007b; Petrusek et al., 2008).

Sample collection

We collected zooplankton samples from the reservoirs in two consecutive summer and autumn seasons (July and October in 2003 and 2004). On each sampling date, zooplankton samples were collected at three stations along the reservoir's horizontal axis: near the river inflow (upstream), in the centre of the reservoir (middle) and near the dam (downstream). At the upstream and middle sampling stations, we hauled a plankton net (mesh size 170 μm) through the entire water column. When sampling during the period of thermal stratification, separate zooplankton samples were collected at the downstream site (which is the deepest station) by a closing net from each of the three vertical layers of the stratified water column (epi-, meta- and hypolimnion). In cases where the water column was not stratified (often in autumn), a vertical haul was done instead. We determined the borders of the epi-, meta- and hypolimnion from vertical profiles of temperature and dissolved oxygen (see Petrusek et al., 2008). In addition, water transparency was measured with a Secchi disk at each sampling station. Secchi disk depth corresponds with chlorophyll concentrations in the reservoirs studied (Seda and Petrusek, unpubl. data), and the samples were never taken at periods when suspended particles other than phytoplankton would substantially affect the water transparency (e.g. after storms or floods).

From each sampling site/vertical layer of the water column, a sample from one or two hauls was preserved in 4% formaldehyde solution for reliable analyses of Daphnia density. Afterwards, material from several hauls was pooled together to obtain a sufficient amount of Daphnia for further analyses, poured through a sieve and split to two subsamples. One was preserved in 96% ethanol (for analyses of Daphnia parasites); another subsample was deep-frozen in liquid nitrogen for subsequent enzyme electrophoresis (Seda et al., 2007b). The latter preservation method has least effect on the shape or brood of the animals, so this particular subsample was used for size and fecundity estimates (the body size and fecundity of ca. 40 randomly chosen adult Daphnia females from each sampling point was recorded).

Parasite screening

We screened all ethanol-preserved samples for signs of parasite infection using a stereomicroscope at ×250 magnification; whenever possible, ca. 200 adult Daphnia females were analysed. Parasites which could not be identified to species level were classified into functional groups according to their pathology and morphological traits (Green, 1974). The prevalence of a given parasite species (or functional group) was estimated as the proportion of infected individuals per sample. Although ethanol preservation eliminates the possibility of reliably detecting some bacterial parasites (Duffy, pers. comm.), a detection level of a variety of other parasites, including protozoans, fungi, microsporidians and oomycetes, is similar to that from the live samples (Wolinska, pers. obs.).

In previously published molecular studies of parasites infecting D. longispina complex (Wolinska et al., 2009; Lohr et al., 2010a), we evaluated taxonomic composition within the four commonly encountered functional parasite groups (protozoan, fungal, oomycete and microsporidian), as assigned by morphology. We sequenced the nuclear small-subunit (SSU) and internal transcribed spacer (ITS) rDNA regions of 181 parasite isolates (representatives of each functional group and many sampling locations, including the reservoirs). The protozoan and fungal infections were each represented by a single parasite taxon: Caullerya mesnili and Metschnikowia sp., respectively. However, we detected four different microsporidian taxa and five taxa of oomycetes (Wolinska et al., 2009). Within these parasite groups, the different taxa were otherwise indistinguishable by microscopic examination. In the present work, all analyses were performed at the level of parasite functional groups. Although in this way we pool multiple microsporidian and oomycete species, this approach can be substantiated. First, a single species of microsporidium, Berwaldia schaefernai, dominated the samples analysed, accounting for 82% of all microsporidian infections of the D. longispina complex identified by molecular methods (Wolinska et al., 2009). Second, all detected microsporidian taxa were closely related, as revealed in a SSU rDNA phylogeny (Wolinska et al., 2009). Finally, oomycete taxa tend to have similar epidemiology and effects on their Daphnia hosts (Wolinska et al., 2008).

Statistical analyses

Distribution of parasites among-reservoirs

First, we tested spatial variation in parasite prevalence among the reservoirs using one-way analysis of variance (in these and all other tests prevalence data were arcsin transformed to bring their distribution closer to normal and to stabilize their variation, Sokal and Rohlf, 1995). Then, we tested the effect of specific reservoir descriptors on parasite prevalence. The list of descriptors included three ordinal variables (each with three levels; low, medium and high): (i) lake trophy (i.e. oligo-, meso- and eutrophic), (ii) overall strength of horizontal environmental gradients affecting Daphnia populations (assessed by Petrusek and Seda from field observations and available information on nutrient levels, phytoplankton, zooplankton and fish) and (iii) overall strength of fish predation pressure in the reservoir; and six continuous variables: (i) age, (ii) altitude, (iii) length, (iv) maximum depth, (v) retention time and (vi) surface area of the reservoir. The three best descriptors were tested in a model. Thus, Holm's correction (Holm, 1979) was applied to type I error estimates, to adjust for the selection from nine reservoir descriptors.

Distribution of parasites within-reservoirs

To test for temporal and spatial variation in parasite prevalence within-reservoirs, we used a linear mixed effect model in the lme package of the R program version 2.8 (R Development Core Team, 2008). The analyses were done separately for each parasite group; thus, only reservoirs containing a given parasite group were included in the respective model. The year, season and horizontal or vertical stations were modelled as fixed effects, while reservoir identity was included as a random effect. For the analyses of parasite horizontal gradients, the total parasite prevalence at the dam was calculated by weighing parasite prevalence at different vertical layers (if applicable) by the proportion of the Daphnia population present in those layers (the latter was based on estimated Daphnia densities from formaldehyde preserved samples). The analyses of parasite vertical gradients at downstream stations were only done for the summer samples, as the autumn sampling coincided with the end of the stratification period. To test if the horizontal and vertical gradients in parasite prevalence were not simply a function of reservoir length or depth (ranging from 3.5 to 29 km and from 17 to 58 m, respectively), we performed linear regressions (separately for each parasite group) on the distance between sampling sites (in kilometres) or vertical layers (in metres) and on the gradient in parasite prevalence (calculated as a ratio) between sites or vertical layers. Only the distances between adjacent points were considered (i.e. upstream-middle and middle-dam, but not upstream-dam). For the calculation of ratios in prevalence gradients, only cases with parasite prevalence exceeding 4% in at least one sample were included. Denominators with a zero value were avoided by computing a surrogate low-prevalence value, based on the assumption that an infection would have been found if one more Daphnia had been included into the random sample.

Finally, we tested whether parasite prevalence was related to water transparency (a proxy for phytoplankton biomass), host density and the overall state of the host population (expressed as Daphnia body size and fecundity, traits that may be affected by various environmental conditions, including food conditions and predation pressure). We used linear mixed effect models, with a random effect of reservoir, year and season. Body size and fecundity were tested in a single model, performed on median values calculated per sample. Medians were chosen instead of means, as some processes, such as size-selective predation, may result in substantial deviations from normal distribution.

RESULTS

Identification of parasites

In the 11 canyon-shaped reservoirs, we observed (by microscopic examination) four groups of Daphnia parasites: C. mesnili, Metschnikowia sp., microsporidia and oomycetes. Caullerya mesnili is a protozoan living in the Daphnia gut epithelium (Green, 1974) and belongs to the class Ichthyosporea (Lohr et al., 2010a). Metschnikowia sp. is a yeast parasite belonging to the class Hemiascomycetes (Wolinska et al., 2009) which forms massive amounts of spores that fill the host body cavity (Green, 1974). The microsporidia taxa infected the body cavity whereas oomycetes infected either the brood pouch or the entire body cavity of their hosts (Wolinska et al., 2009). The C. mesnili and Metschnikowia sp. infections were detected in a subset of the reservoirs (8 and 4, respectively), whereas microsporidia and oomycetes infections were observed in all the reservoirs studied (Fig. 1).

Fig. 1.

Prevalence of four parasites among Daphnia populations inhabiting 11 canyon-shaped reservoirs, along horizontally separated sampling stations (u, upstream; m, middle; d, downstream) and sampling seasons (grey lines, summer 2003 and 2004; blacklines, autumn 2003 and 2004). If less than four lines are shown, the parasite was not detected in the respective season; ‘none’ indicates that the parasite was not detected in a given lake. Note the difference in scale on the y-axes.

Fig. 1.

Prevalence of four parasites among Daphnia populations inhabiting 11 canyon-shaped reservoirs, along horizontally separated sampling stations (u, upstream; m, middle; d, downstream) and sampling seasons (grey lines, summer 2003 and 2004; blacklines, autumn 2003 and 2004). If less than four lines are shown, the parasite was not detected in the respective season; ‘none’ indicates that the parasite was not detected in a given lake. Note the difference in scale on the y-axes.

Distribution of parasites: temporal variation

Prevalence of three out of four parasite groups changed significantly between seasons (Table I, Fig. 1). Caullerya mesnili and Metschnikowia sp. were more prevalent in autumn than in summer. Specifically, Metschnikowia sp. was exclusively detected in autumn (in four reservoirs), whereas C. mesnili was detected in eight reservoirs in autumn and only in three of these in summer. In contrast, the prevalence of parasitic oomycetes was significantly higher in summer, and the prevalence of microsporidia did not differ between seasons. None of the four parasite groups differed significantly in prevalence between the 2 years.

Table I:

Effect of year, season, horizontal and vertical position on the prevalence of four parasites in Daphnia populations inhabiting 11 canyon-shaped reservoirs

Parasite F-statistic P-value df Response 
Year 
Caullerya mesnili 3.12 0.0807 84 NA 
Metschnikowia sp. 1.89 0.1769 40 NA 
 Microsporidia 2.57 0.1120 106 NA 
 Oomycetes 0.14 0.7050 106 NA 
Season 
Caullerya mesnili 14.70 0.0002 84 Higher in autumn 
Metschnikowia sp. 20.99 <0.0001 40 Higher in autumn 
 Microsporidia 0.02 0.8855 106 NA 
 Oomycetes 15.61 0.0001 106 Higher in summer 
Horizontal gradient 
Caullerya mesnili 4.43 0.0148 84 Highest at upstream 
Metschnikowia sp. 2.89 0.0675 40 (Highest at middle and downstream) 
 Microsporidia 1.54 0.2201 106 NA 
 Oomycetes 4.48 0.0135 106 Highest at downstream 
Vertical gradient 
 Oomycetes 4.14 0.0218 49 Highest at hypo- and then metalimnion 
Parasite F-statistic P-value df Response 
Year 
Caullerya mesnili 3.12 0.0807 84 NA 
Metschnikowia sp. 1.89 0.1769 40 NA 
 Microsporidia 2.57 0.1120 106 NA 
 Oomycetes 0.14 0.7050 106 NA 
Season 
Caullerya mesnili 14.70 0.0002 84 Higher in autumn 
Metschnikowia sp. 20.99 <0.0001 40 Higher in autumn 
 Microsporidia 0.02 0.8855 106 NA 
 Oomycetes 15.61 0.0001 106 Higher in summer 
Horizontal gradient 
Caullerya mesnili 4.43 0.0148 84 Highest at upstream 
Metschnikowia sp. 2.89 0.0675 40 (Highest at middle and downstream) 
 Microsporidia 1.54 0.2201 106 NA 
 Oomycetes 4.48 0.0135 106 Highest at downstream 
Vertical gradient 
 Oomycetes 4.14 0.0218 49 Highest at hypo- and then metalimnion 

Significant P-values are given in bold. NA, not applicable.

Distribution of parasites: spatial variation

Among-reservoirs

Parasite prevalence differed significantly among lakes for C. mesnili (F10,121 = 4.14, P < 0.0001), Metschnikowia sp. (F10,121 = 3.48, P = 0.0005) and oomycetes (F10,121 = 3.03, P = 0.0018), but not for microsporidia (F10,121 = 1.87, P = 0.0554). Among the nine environmental descriptors tested, the best predictor of C. mesnili prevalence was maximum depth (F1,7 = 16.94, P = 0.0405), with the incidence of this parasite increasing with reservoir depth. The prevalence of oomycetes significantly increased with the overall strength of horizontal environmental gradients (F1,7 = 19.65, P = 0.0270). No relationship was found between the environmental descriptors analysed and the prevalence of Metschnikowia sp. or microsporidia.

Within-reservoirs (horizontal gradient)

Prevalence of C. mesnili was higher in the upstream stations, whereas the oomycetes were more abundant in the downstream parts of the reservoirs (Table I, Fig. 1). For Metschnikowia sp., there was also a trend of increased prevalence at middle and downstream stations (it was completely absent from the upstream part of two of the three reservoirs in which it was most abundant); however, the pattern was not significant with the given sample size (P = 0.067). In contrast, the prevalence of microsporidia remained unchanged across the sampling stations. Linear regression analyses showed that the strength of the gradient in parasite prevalence between sampling sites was not a function of reservoir length (C. mesnili: r2 = 0.170, n = 18, P = 0.09; Metschnikowia sp.: r2 = 0.006, n = 10, P = 0.83; microsporidia: r2 = 0.152, n = 7, P = 0.39; oomycetes: r2 = 0.035, n = 9, P = 0.63).

Prevalence of three out of the four parasites studied correlated with water transparency (Table II). With increasing transparency, the prevalence of C. mesnili decreased whereas prevalence of Metschnikowia sp. and oomycetes increased. Additionally, the prevalence of C. mesnili increased with increasing host density. None of the parasites prevalence was significantly related to Daphnia body size and only in the case of oomycetes did prevalence decrease with increasing host fecundity (Table II).

Table II:

Effect of water transparency, host population density (calculated per litre) and overall Daphnia state (measured as body size and fecundity) on the prevalence of four parasites in Daphnia populations inhabiting 11 canyon-shaped reservoirs

Parasite F-statistic P-value df Response 
Water transparency 
Caullerya mesnili 8.91 0.0040 63 Decreases with increasing transparency 
Metschnikowia sp. 13.24 0.0010 31 Increases with increasing transparency 
 Microsporidia 0.59 0.4438 79 NA 
 Oomycetes 6.77 0.0110 79 Increases with increasing transparency 
Density 
Caullerya mesnili 4.91 0.0303 63 Increases with increasing density 
Metschnikowia sp. 0.25 0.6217 31 NA 
 Microsporidia 1.40 0.2407 79 NA 
 Oomycetes 2.10 0.1515 79 NA 
Body sizea 
Caullerya mesnili 1.34 0.2508 62 NA 
Metschnikowia sp. 1.01 0.3231 30 NA 
 Microsporidia 1.54 0.2180 78 NA 
 Oomycetes 1.05 0.3079 78 NA 
Fecunditya 
Caullerya mesnili 0.44 0.5090 62 NA 
Metschnikowia sp. 0.90 0.3501 30 NA 
 Microsporidia 0.73 0.3943 78 NA 
 Oomycetes 10.21 0.0021 78 Decreases with increasing fecundity 
Parasite F-statistic P-value df Response 
Water transparency 
Caullerya mesnili 8.91 0.0040 63 Decreases with increasing transparency 
Metschnikowia sp. 13.24 0.0010 31 Increases with increasing transparency 
 Microsporidia 0.59 0.4438 79 NA 
 Oomycetes 6.77 0.0110 79 Increases with increasing transparency 
Density 
Caullerya mesnili 4.91 0.0303 63 Increases with increasing density 
Metschnikowia sp. 0.25 0.6217 31 NA 
 Microsporidia 1.40 0.2407 79 NA 
 Oomycetes 2.10 0.1515 79 NA 
Body sizea 
Caullerya mesnili 1.34 0.2508 62 NA 
Metschnikowia sp. 1.01 0.3231 30 NA 
 Microsporidia 1.54 0.2180 78 NA 
 Oomycetes 1.05 0.3079 78 NA 
Fecunditya 
Caullerya mesnili 0.44 0.5090 62 NA 
Metschnikowia sp. 0.90 0.3501 30 NA 
 Microsporidia 0.73 0.3943 78 NA 
 Oomycetes 10.21 0.0021 78 Decreases with increasing fecundity 

Significant P-values are given in bold. NA, not applicable.

aThese two parameters were tested in a single model.

Within-reservoirs (vertical gradient)

We compared the prevalence of oomycetes among the different layers of the stratified water column (epi-, meta- and hypolimnion). This was the sole parasite group which was sufficiently abundant at the downstream sites where we collected separate samples along the vertical gradient. The oomycetes were more prevalent in the lower layers of the water column (meta- and hypolimnion) than in the epilimnion (Table I, Fig. 2). The strength of the gradient between the vertical layers was not a function of reservoir depth (r2 < 0.001, n = 14, P = 0.93).

Fig. 2.

Prevalence of oomycete parasites among Daphnia populations inhabiting 11 canyon-shaped reservoirs along vertically separated sampling stations at downstream sites (white dots, epilimnion; grey dots, metalimnion; black dots, hypolimnion). Absence of a black dot indicates that the prevalence could not be assessed because the Daphnia density was extremely low (<1 L−1) at the deepest station.

Fig. 2.

Prevalence of oomycete parasites among Daphnia populations inhabiting 11 canyon-shaped reservoirs along vertically separated sampling stations at downstream sites (white dots, epilimnion; grey dots, metalimnion; black dots, hypolimnion). Absence of a black dot indicates that the prevalence could not be assessed because the Daphnia density was extremely low (<1 L−1) at the deepest station.

DISCUSSION

In contrast to the scarcity of work regarding spatial, within-population patterns of Daphnia diseases, the among-population variation in parasite prevalence has been a frequent research topic (e.g. Ebert et al., 2001; Caceres et al., 2006; Wolinska et al., 2007). These previous studies found that environmental conditions can, to some extent, explain the frequently observed heterogeneous patterns in parasite prevalence. In the present study, for two out of four parasites tested, we observed some relationship between parasite prevalence and reservoir characteristics. Specifically, C. mesnili was more abundant in deeper reservoirs and oomycetes in reservoirs with strong environmental gradients. Although Metschnikowia sp. was also present in a limited set of reservoirs (4 out of 11), none of the parameters measured explained its scattered geographical pattern.

Microsporidia were the only parasite group equally abundant in all study reservoirs. Their homogenous distribution across the reservoirs agrees with their proposed transmission mode; it is likely that common microsporidia infecting Daphnia have a mobile secondary host (Refardt et al., 2002). The most dominant microsporidium species in the reservoirs, B. schaefernai, is presumably transmitted by an insect with aquatic larvae (Wolinska et al., 2011). This should facilitate spread of such parasite species and result in their more homogeneous distribution in comparison with C. mesnili or Metschnikowia sp., which are transmitted directly from Daphnia to Daphnia by a waterborne stage (Lohr et al., 2010b). The presumed differences in transmission mode of these parasite groups are supported by patterns of population genetic structure of C. mesnili and the microsporidian B. schaefernai co-occurring in the same reservoirs, suggesting substantially higher amount of gene flow for the latter (Wolinska et al., 2011).

Seasonal parasite prevalence dynamics are common in nature and seem to be the rule rather than the exception (see references in Lass and Ebert, 2006). We observed seasonal fluctuations in prevalence for three of the four parasites studied. The dynamics of C. mesnili, with epidemics starting in autumn, are consistent with data collected from a long-term field survey of a single lake (Wolinska et al., 2006). Moreover, in agreement with its emergence in autumn, when the water temperature has decreased, we have experimental evidence that the prevalence of this parasite is higher in colder water (Schoebel et al., 2011). In general, it has been shown, across several systems, that the seasonality of infectious diseases is often driven by environmental factors, with temperature playing an especially important role (see Lass and Ebert, 2006).

So far, spatial variation in parasite prevalence within host populations (i.e. within-lake approach) has been tested in a single study of Daphnia parasites (Hall et al., 2005). In contrast to our work, in which we observed strong spatial variation in parasite prevalence within-reservoirs, Hall et al. (2005) found little systematic evidence of strong parasite aggregation. However, the lakes studied by Hall et al. (2005) were relatively small (∼500 m in length), so that the environment was less heterogeneous on a horizontal scale than canyon-shaped reservoirs (Straškraba, 1998).

Which environmental conditions could have caused the strong variation in parasite prevalence on a within-population scale, as observed in our study? One important factor, which may impact parasite distributions via an influence on the host's physiological state, is the gradient in the abundance of phytoplankton, the main food source for Daphnia. It has been shown across a variety of systems that host resources strongly influence disease epidemics (see references in Hall et al., 2009). Changes in algal biomass in the reservoirs were reflected by a consistent increase in water transparency from upstream towards the dam regions (data not shown), which was especially pronounced in reservoirs with higher trophic levels. Indeed, the gradients in phytoplankton abundance have been well documented in the reservoirs (e.g. Hejzlar and Vyhnalek, 1998; Seda and Devetter, 2000). The prevalence of Metschnikowia sp. and oomycetes increased with increasing transparency; implying higher parasite prevalence under low-food conditions. This corresponds well with data from laboratory experiments; the prevalence of Metschnikowia sp. (Hall et al., 2007, 2009) and oomycetes (Seymour et al., 1984) were higher among Daphnia under low-food conditions. Within-reservoirs, we also observed a negative relationship between oomycete prevalence and Daphnia fecundity, which itself is a good proxy for food quality (Lampert, 1978). In the latter case; however, it is difficult to disentangle cause and consequence between parasite prevalence and fecundity effects. We treated Daphnia fecundity as an explanatory variable, but parasite prevalence can also affect host fecundity (Decaestecker et al., 2005; Wolinska et al., 2007).

Further, although the observed relationships between parasite prevalence and water transparency or host fecundity may suggest an influence of Daphnia food conditions, the same pattern could also result from an alternative (but not mutually exclusive) process. It has been documented across several host-parasite systems that predators selectively preying on infected hosts can inhibit the spread of disease (e.g. Packer et al., 2003). Thus, another important factor that might contribute to the observed gradients in parasite prevalence is fish predation. Most important Daphnia predators in the studied reservoirs are visually orienting, size-selective planktivorous fish. Any condition that increases the conspicuousness of the prey (including infection by certain parasites) therefore increases prey vulnerability to predation (Duffy and Hall, 2008). The gradient in fish predation pressure within-reservoirs tends to follow the same pattern as food conditions for Daphnia, decreasing downstream (e.g. Prchalova et al., 2008, 2009). The observed gradients are caused by higher zooplankton abundance in upstream regions and, for some fish species, better spawning conditions (Prchalova et al., 2008).

Infections by three parasite groups included in our study, Metschnikowia sp., microsporidia and oomycetes, substantially increase host opacity (Wolinska, pers. obs.). As a result, infected hosts might be preferentially selected by visually hunting fish, which is well documented for Metschnikowia sp. (Duffy and Hall, 2008). In agreement with this prediction, the prevalence of Metschnikowia sp. and oomycetes was higher in the middle and downstream stations, where fish predation is relatively low. The distribution of C. mesnili, which was more abundant in the upstream region, also fits this scenario. As C. mesnili only infects the Daphnia gut (Lohr et al., 2010a), which does not reduce host transparency, this should not lead to selective removal of infected hosts. On the other hand, we did not find any significant relationship between Daphnia body size (a potential proxy for fish predation pressure, Seda and Kubecka, 1997) and the prevalence of any of the parasite groups studied. It is therefore possible that the parasite distribution results from the combination of both bottom-up (food supply) and top-down (predation pressure) processes that impact the Daphnia hosts.

Finally, we found that the prevalence of C. mesnili increased with host density. In agreement with this, it has been shown in both theoretical (e.g. Anderson and May, 1978) and empirical studies, including the DaphniaC. mesnili system (Bittner et al., 2002), that the transmission of parasites is density dependent. The lack of such an association can easily be explained for both Metschnikowia sp. and microsporidia. Metschnikowia sp. can spread only when the host dies. However, Daphnia dying of infection may often sink to the bottom of the lake before releasing spores (Caceres et al., 2006). Thus, the shape of the lake basin might be more important for the dispersal of this parasite taxon than host density. Specifically, the lake morphology determines the strength of the mixing of water layers, a process that re-suspends parasite spores in the water column (Caceres et al., 2006; Hall et al., 2010). For microsporidians infecting the D. longispina complex, it is suspected that they need passage through a secondary host in order to complete their life cycle (Refardt et al., 2002; Wolinska et al., 2011). Thus, the spread of infection should not be directly related to Daphnia density.

In addition to horizontal variation, the prevalence of oomycetes also varied among vertical layers of the stratified water column. Several explanations could account for this pattern. Fels et al. (2004) showed under laboratory conditions for three microsporidian and one bacterium species that infection affected depth selection of their Daphnia hosts; however, the cause of these differences remains unknown. In other host-parasite systems it is well documented that parasites may manipulate host behaviour to increase their transmission rate (e.g. Moore, 2002) or, in contrast, selection may favour host behaviour that decreases parasite transmission (e.g. Lowenberger and Rau, 1994). We found that the prevalence of oomycetes was higher in deeper water layers. If visually oriented hunting fish represent a dead-end for parasite transmission, the migration of infected hosts to deeper water layers, where fish density and therefore predation pressure is much lower (Prchalova et al., 2008), would increase parasite fitness. On the other hand, if fish have infectious faeces (Duffy, 2009) or serve as an alternative host for oomycetes, the escape of infected Daphnia to deeper layers may cause a decrease in parasite transmission. Indeed, Daphnia and fish seem to share some oomycete parasites, as suggested by identical sequences of SSU and ITS rDNA regions detected in several oomycete isolates from Daphnia and fish (Wolinska et al., 2008, 2009). Alternatively, the influence of parasites on Daphnia behaviour may be non-adaptive: some infected hosts in poor physiological states show reduced swimming activity (Petrusek, pers. obs.), which could result in their passive sinking to deeper layers. The higher prevalence of oomycetes observed in deeper layers of the water column may also be explained without behavioural alterations to the host. Visual fish predation may selectively remove infected (and therefore conspicuous) individuals from the upper, well-lit layer (Duffy and Hall, 2008), eventually leading to increased parasite prevalence in the deeper layers. In any case, Daphnia vertical migration as a fish-avoidance strategy may result in migrating to layers with increased disease prevalence, therefore adding potential extra cost to this antipredator strategy (as shown in shallow waters for D. magna by Decaestecker et al., 2002). If the lake bottom contains a source of parasite spores (as suggested for bacterial parasites of Daphnia, Decaestecker et al., 2002), the deep hypolimnetic fraction of the pelagic Daphnia population (which can be found in reservoirs, Seda et al., 2007a) may be at a higher risk of infection from this source, therefore further contributing to the heterogeneous patterns of disease prevalence.

Could differences in host taxon composition account for the observed gradients in parasite prevalence? Previous field surveys of these 11 reservoirs showed that the taxonomic structure of the D. longispina complex often differed among reservoirs as well as within-reservoirs, across the horizontally and vertically separated sampling stations (Seda et al., 2007b; Petrusek et al., 2008). We believe, however, that this factor is not particularly important for parasite prevalence patterns. First, we have shown in previous field studies, where infected Daphnia were identified with molecular markers, that C. mesnili and oomycete parasites were able to infect all relevant taxa within the D. longispina complex (i.e. parental species as well as their interspecific hybrids, Wolinska et al., 2007), and any preferences of parasites towards infecting a particular host taxon are temporally variable and unstable (Wolinska et al., 2006). Moreover, certain within-reservoir locations, characterized by a specific parasite composition (for example, C. mesnili being highly abundant in the upstream stations), were often occupied by different Daphnia taxa in different reservoirs (Seda et al., 2007b; Petrusek et al., 2008). Specifically, C. mesnili was highly prevalent in the upstream stations that were solely occupied by D. galeata (Stanovice, Trnávka, Římov and Želivka) as well as in those that were dominated by D. cucullata (Brno and Vír). Moreover, in the present survey, we also included reservoirs dominated entirely by a single Daphnia species (Řimov, Stanovice and Trnávka, Seda et al., 2007b; Petrusek et al., 2008), which still showed strongly pronounced parasite gradients.

The spatial heterogeneity of diseases may be caused by various factors, from the fixed genetic background influencing the outcome of host-parasite encounters via host behavioural traits, to habitat properties altering parasite spread. Understanding which environmental factors determine parasite prevalence within a host population is of increasing interest, given the rising incidence of many infectious diseases (e.g. Ostfeld, 2009). As the within-population parasite gradients observed in this work might also be a common scenario in other host-parasite systems, a key challenge for future host-parasite studies is to focus research at the appropriate spatial scale. Otherwise, important ecological (e.g. changes in host densities) and evolutionary (e.g. changes in host genetic structure) signals may be missed. Further studies should try to follow spatial dynamics in parasite prevalence in combination with host population dynamics, thus linking together the causes and effects of parasite-driven patterns.

SUPPLEMENTARY DATA

Supplementary data can be found online at http://plankt.oxfordjournals.org.

FUNDING

The study was supported by the German Science Foundation (WO 1587/1-1 and 2-1), the Czech Science Foundation (project 206/04/0190), the Grant Agency of the Academy of Sciences of the Czech Republic (IAA600960901) and the Czech Ministry of Education (MSM0021620828 and MSM6007665801).

ACKNOWLEDGEMENTS

We thank Štěpánka Dlouhá and Ivana Vaníčková for help with sampling, Jana Hubova for Daphnia measurements; and Jennifer Lohr, Meghan Duffy and anonymous referees for very useful comments.

REFERENCES

Anderson
R. M.
May
R. M.
Regulation and stability of host-parasite population interactions. 1. Regulatory processes
J. Anim. Ecol.
 , 
1978
, vol. 
47
 (pg. 
219
-
247
)
Bittner
K.
Rothhaupt
K. O.
Ebert
D.
Ecological interactions of the microparasite Caullerya mesnili and its host Daphnia galeata
Limnol. Oceanogr.
 , 
2002
, vol. 
47
 (pg. 
300
-
305
)
Caceres
C. E.
Hall
S. R.
Duffy
M. A.
, et al.  . 
Physical structure of lakes constrains epidemics in Daphnia populations
Ecology
 , 
2006
, vol. 
87
 (pg. 
1438
-
1444
)
Carlsson-Graner
U.
Thrall
P. H.
The spatial distribution of plant populations, disease dynamics and evolution of resistance
Oikos
 , 
2002
, vol. 
97
 (pg. 
97
-
110
)
Decaestecker
E.
De Meester
L.
Ebert
D.
In deep trouble: Habitat selection constrained by multiple enemies in zooplankton
Proc. Natl. Acad. Sci. USA
 , 
2002
, vol. 
99
 (pg. 
5481
-
5485
)
Decaestecker
E.
Declerck
S.
De Meester
L.
, et al.  . 
Ecological implications of parasites in natural Daphnia populations
Oecologia
 , 
2005
, vol. 
144
 (pg. 
382
-
390
)
Duffy
M. A.
Staying alive: the post-consumption fate of parasite spores and its implications for disease dynamics
Limnol. Oceanogr.
 , 
2009
, vol. 
54
 (pg. 
770
-
773
)
Duffy
M. A.
Hall
S. R.
Selective predation and rapid evolution can jointly dampen effects of virulent parasites on Daphnia populations
Am. Nat
 , 
2008
, vol. 
171
 (pg. 
499
-
510
)
Ebert
D.
Hottinger
J. W.
Pajunen
V. I.
Temporal and spatial dynamics of parasite richness in a Daphnia metapopulation
Ecology
 , 
2001
, vol. 
82
 (pg. 
3417
-
3434
)
Fels
D.
Lee
V. A.
Ebert
D.
The impact of microparasites on the vertical distribution of Daphnia magna
Arch. Hydrobiol
 , 
2004
, vol. 
161
 (pg. 
65
-
80
)
Foley
D. H.
Torres
E. P.
Mueller
I.
, et al.  . 
Host-dependent Anopheles flavirostris larval distribution reinforces the risk of malaria near water
Trans. R. Soc. Trop. Med. Hyg
 , 
2003
, vol. 
97
 (pg. 
283
-
287
)
Green
J.
Parasites and epibionts of Cladocera
Trans. Zool. Soc. London
 , 
1974
, vol. 
32
 (pg. 
417
-
515
)
Hall
S. R.
Duffy
M. A.
Tessier
A. J.
, et al.  . 
Spatial heterogeneity of daphniid parasitism within lakes
Oecologia
 , 
2005
, vol. 
143
 (pg. 
635
-
644
)
Hall
S. R.
Sivars-Becker
L.
Becker
C.
, et al.  . 
Eating yourself sick: transmission of disease as a function of foraging ecology
Ecol. Lett
 , 
2007
, vol. 
10
 (pg. 
207
-
218
)
Hall
S. R.
Knight
C. J.
Becker
C. R.
, et al.  . 
Quality matters: resource quality for hosts and the timing of epidemics
Ecol. Lett
 , 
2009
, vol. 
12
 (pg. 
118
-
128
)
Hall
S. R.
Smyth
R.
Becker
C. R.
, et al.  . 
Why are Daphnia in some lakes sicker? Disease ecology, habitat structure, and the plankton
Bioscience
 , 
2010
, vol. 
60
 (pg. 
363
-
375
)
Hejzlar
J.
Vyhnalek
V.
Longitudinal heterogeneity of phosphorus and phytoplankton concentrations in deep-valley reservoirs
Int. Rev. Hydrob.
 , 
1998
, vol. 
83
 (pg. 
139
-
146
)
Holm
S.
A simple sequentially rejective multiple test procedure
Scand. J. Stat.
 , 
1979
, vol. 
6
 (pg. 
65
-
70
)
Hudson
P. J.
Newborn
D.
Dobson
A. P.
Regulation and stability of a free-living host-parasite system - Trichostrongylus tenuis in red grouse. 1. Monitoring and parasite reduction experiments
J. Anim. Ecol.
 , 
1992
, vol. 
61
 (pg. 
477
-
486
)
Johnson
P. T. J.
Chase
J. M.
Dosch
K. L.
, et al.  . 
Aquatic eutrophication promotes pathogenic infection in amphibians
Proc. Natl. Acad. Sci. USA
 , 
2007
, vol. 
104
 (pg. 
15781
-
15786
)
Karvonen
A.
Bagge
A. M.
Valtonen
E. T.
Parasite assemblages of crucian carp (Carassius carassius)—is depauperate composition explained by lack of parasite exchange, extreme environmental conditions or host unsuitability?
Parasitology
 , 
2005
, vol. 
131
 (pg. 
273
-
278
)
Lampert
W.
Field study on dependance of fecundity of Daphnia spec. on food concentration
Oecologia
 , 
1978
, vol. 
36
 (pg. 
363
-
369
)
Lampert
W.
Sommer
U.
Limnoökologie
 , 
1999
Suttgart, New York
Georg Thieme Verlag
Lass
S.
Ebert
D.
Apparent seasonality of parasite dynamics: analysis of cyclic prevalence patterns
Proc. R. Soc. B
 , 
2006
, vol. 
273
 (pg. 
199
-
206
)
Lohr
J.
Laforsch
C.
Koerner
H.
, et al.  . 
A Daphnia parasite (Caullerya mesnili) constitutes a new member of the Ichthyosporea, a group of protists near the animal-fungi divergence
J. Eukaryot. Microbiol
 , 
2010
, vol. 
53
 (pg. 
328
-
336
)
Lohr
J.
Yin
M.
Wolinska
J.
Prior residency does not always pay off—coinfections in Daphnia
Parasitology
 , 
2010
, vol. 
137
 (pg. 
1493
-
1500
)
Lowenberger
C. A.
Rau
M. E.
Plagiorchis elegans—emergence, longevity and infectivity of cercariae, and host behavioural modifications during cercarial emergence
Parasitology
 , 
1994
, vol. 
109
 (pg. 
65
-
72
)
Moore
J.
Parasites and the Behavior of Animals
 , 
2002
Oxford
Oxford University Press
Moyer
B. R.
Drown
D. M.
Clayton
D. H.
Low humidity reduces ectoparasite pressure: implications for host life history evolution
Oikos
 , 
2002
, vol. 
97
 (pg. 
223
-
228
)
Ostfeld
R. S.
Climate change and the distribution and intensity of infectious diseases
Ecology
 , 
2009
, vol. 
90
 (pg. 
903
-
905
)
Ovaskainen
O.
Laine
A. L.
Inferring evolutionary signals from ecological data in a plant-pathogen metapopulation
Ecology
 , 
2006
, vol. 
87
 (pg. 
880
-
891
)
Packer
C.
Holt
R. D.
Hudson
P. J.
, et al.  . 
Keeping the herds healthy and alert: implications of predator control for infectious disease
Ecol. Lett
 , 
2003
, vol. 
6
 (pg. 
797
-
802
)
Petrusek
A.
Seda
J.
Machacek
J.
, et al.  . 
Daphnia hybridization along ecological gradients in pelagic environments: the potential for the presence of hybrid zones in plankton
Philos. Trans. R. Soc. Lond. B
 , 
2008
, vol. 
363
 (pg. 
2931
-
2941
)
Prchalova
M.
Kubecka
J.
Vasek
M.
, et al.  . 
Distribution patterns of fishes in a canyon-shaped reservoir
J. Fish Biol
 , 
2008
, vol. 
73
 (pg. 
54
-
78
)
Prchalova
M.
Kubecka
J.
Cech
M.
, et al.  . 
The effect of depth, distance from dam and habitat on spatial distribution of fish in an artificial reservoir
Ecol. Freshw. Fish
 , 
2009
, vol. 
18
 (pg. 
247
-
260
)
R Development Core Team
R: A Language and Environment for Statistical Computing
 , 
2008
Vienna, Austria
R Foundation for Statistical Computing
Refardt
D.
Canning
E. U.
Mathis
A.
, et al.  . 
Small subunit ribosomal DNA phylogeny of microsporidia that infect Daphnia (Crustacea: Cladocera)
Parasitology
 , 
2002
, vol. 
124
 (pg. 
381
-
389
)
Schoebel
C. N.
Tellenbach
C.
Spaak
P.
, et al.  . 
Temperature effects on parasite prevalence in a natural hybrid complex
Biol. Lett
 , 
2011
, vol. 
7
 (pg. 
108
-
111
)
Seda
J.
Devetter
M.
Zooplankton community structure along a trophic gradient in a canyon-shaped dam reservoir
J. Plankton Res
 , 
2000
, vol. 
22
 (pg. 
1829
-
1840
)
Seda
J.
Kubecka
J.
Long-term biomanipulation of Rimov Reservoir (Czech Republic)
Hydrobiologia
 , 
1997
, vol. 
345
 (pg. 
95
-
108
)
Seda
J.
Kolarova
K.
Petrusek
A.
, et al.  . 
Daphnia galeata in the deep hypolimnion: spatial differentiation of a "typical epilimnetic" species
Hydrobiologia
 , 
2007
, vol. 
594
 (pg. 
47
-
57
)
Seda
J.
Petrusek
A.
Machacek
J.
, et al.  . 
Spatial distribution of the Daphnia longispina species complex and other planktonic crustaceans in the heterogeneous environment of canyon-shaped reservoirs
J. Plankton Res.
 , 
2007
, vol. 
29
 (pg. 
619
-
628
)
Seymour
R.
Cowgill
U. M.
Klecka
G. M.
, et al.  . 
Occurrence of Aphanomyces daphniae infection in laboratory cultures of Daphnia magna
J. Invertebr. Pathol
 , 
1984
, vol. 
43
 (pg. 
109
-
113
)
Sokal
R. R.
Rohlf
F. J.
Biometry
 , 
1995
San Francisco
W.H. Freeman and Co
Straškraba
M.
Limnological differences between deep valley reservoirs and deep lakes
Int. Rev. Hydrob
 , 
1998
, vol. 
83
 (pg. 
1
-
12
)
Summers
K.
McKeon
S.
Sellars
J.
, et al.  . 
Parasitic exploitation as an engine of diversity
Biol. Rev.
 , 
2003
, vol. 
78
 (pg. 
639
-
675
)
Urabe
J.
Stable horizontal variation in the zooplankton community structure of a reservoir maintained by predation and competition
Limnol. Oceanogr
 , 
1990
, vol. 
35
 (pg. 
1703
-
1717
)
Wolinska
J.
King
C. E.
Environment can alter selection in host-parasite interactions
Trends Parasitol
 , 
2009
, vol. 
25
 (pg. 
236
-
244
)
Wolinska
J.
Bittner
K.
Ebert
D.
, et al.  . 
The coexistence of hybrid and parental Daphnia: the role of parasites
Proc. R. Soc., B.
 , 
2006
, vol. 
273
 (pg. 
1977
-
1983
)
Wolinska
J.
Keller
B.
Manca
M.
, et al.  . 
Parasite survey of a Daphnia hybrid complex: host-specificity and environment determine infection
J. Anim. Ecol
 , 
2007
, vol. 
76
 (pg. 
191
-
200
)
Wolinska
J.
King
K. C.
Vigneux
F.
, et al.  . 
Virulence, cultivating conditions, and phylogenetic analyses of oomycete parasites in Daphnia
Parasitology
 , 
2008
, vol. 
135
 (pg. 
1667
-
1678
)
Wolinska
J.
Giessler
S.
Koerner
H.
Molecular identification and hidden diversity of novel Daphnia parasites from European lakes
Appl. Environ. Microbiol
 , 
2009
, vol. 
75
 (pg. 
7051
-
7059
)
Wolinska
J.
Spaak
P.
Petrusek
A.
, et al.  . 
Transmission mode affects the population genetic structure of Daphnia parasites
J. Evol. Biol.
 , 
2011
, vol. 
24
 (pg. 
265
-
273
)
Wood
M. J.
Cosgrove
C. L.
Wilkin
T. A.
, et al.  . 
Within-population variation in prevalence and lineage distribution of avian malaria in blue tits, Cyanistes caeruleus
Mol. Ecol.
 , 
2007
, vol. 
16
 (pg. 
3263
-
3273
)

Author notes

Present Address: Georg-August Universität, Dept. Neuroimmunologie, Institut Für Multiple Sklerose Forschung, Waldweg 33, D-37073 Göttingen, Germany
Corresponding editor: Beatrix E. Beisner