Longitudinal Studies

Table 4.1 gives on overview over all longitudinal field studies on Cladoceran parasites.

Green (1974) conducted a 4-year longitudinal study of Long Water at Hampton Court, UK that included several species of Cladocerans but no Daphnia. It is not known whether this pond contained planktivorous fish. Green observed that some parasite species (also known to infect Daphnia) had a seasonal abundance pattern. For instance, Pasteuria ramosa and Spirobacillus cienkowskii were typically found only between April and December. The abundance patterns of other parasite species were not tied to seasons, however, leading Green to conclude that the distribution of certain parasites is influenced by the severity of winter and spring temperatures. He did not discuss the role of host density as an explanatory factor. He did observe, however, that several parasite species have negative effects on host survival and fecundity.

Brambilla (1983) studied the microsporidium Thelohania sp. in a D. pulex population over a 3-year period in a small, apparently fishless vernal pool in Michigan, USA. He noted that parasite prevalence varied strongly, from 20% to peaks of nearly 100% of adult females. Parasites were present in May and June in all 3 years and were first seen whenever the host density rose above 2-3 animals/liter. One year, however, the parasites disappeared in mid-summer despite high host densities, suggesting that high density alone cannot explain parasite spread. Infection of females with ephippia was never observed. Parasitized animals were usually larger than uninfected hosts but had lower fecundity and survival.

Yan and Larsson (1988) followed the dynamics of two undescribed and very similar microsporidian parasites of Holopedium gibberum in a 32-ha Canadian shield lake (maximum depth, 16 m) from April to October 1985. The lake has several planktivorous fish species. Parasites appeared only in July and reached a prevalence of 4%, which they maintained for the rest of the summer. The parasites appeared when host density was high but did not decline when host density decreased. The authors argued that elevated summer temperatures were not the cause of the seasonal occurrence of the parasites. They further rejected the idea that changes in host resistance influenced the abundance pattern of the parasites. They suggest, instead, that the interplay between host and parasite population dynamics may have caused the seasonal changes in prevalence and that predation by planktivorous fish may have further influenced these changes, because infected hosts may be, they speculate, the preferred target of visually hunting fish. Infected hosts had a lower fecundity than healthy hosts and may have had lower survival. In closing, the authors noted that in a survey of 15 other H. gibberum populations in shield lakes, 3 additional populations were found to be parasitized by microsporidians.

Vidtmann (1993) studied parasitism by the microsporidium Larssonia daphniae (later called Larssonia obtusa (Vidtmann and Sokolova 1994)) in a D. pulex population in a shallow, fishless, eutrophic pond at the Kaunas Zoological Garden in Lithuania. He observed that although microsporidians were present only during times of high host density, they were nonetheless often absent during periods of high host density as well. Prevalence among adult females peaked in summer at a maximum of 52%, but the average prevalence (all age classes) within seasons and across years was much lower: 0.63% in spring, 3.2% in summer, and 2.4% in fall. Prevalence was generally lower in juveniles and in males. Over 3 years, the microsporidians were seen only from late May to early October. Because this period closely overlaps with the presence of the host, this apparent seasonality may be related to the seasonal occurrence of the host. Nevertheless, Vidtmann (1993) speculated that the delayed onset of L. daphniae epidemics in May was a consequence of low spring temperatures.

Schwartz and Cameron (1993) studied an undescribed trematode parasite of D. obtusa from seven seasonal, fishless ponds in southeastern Texas, USA over 4 years. They recorded strong within-season, between-year, and between-pond dynamics in the presence of the parasite. Despite recording maximum prevalences up to 79%, they more typically found prevalences to be around a few percents. Large animals were more often infected than small females. Host fecundity was only reduced in infections with three or more parasites per host.

Stirnadel (1994) and Stirnadel and Ebert (1997) studied parasites of D. magna, D. pulex, and D. longispina in three fishless ponds near Oxford, UK over a period of 1 year (about 10-12 Daphnia generations, 65 samples in total). She assessed host density and fecundity together with parasite prevalence, richness, diversity, and host specificity. Overall parasite prevalence (all species combined) was high throughout the year, averaging 84.7% in adult D. magna, 53.6% in D. pulex, and 38.6% in D. longispina. Overall, 31% of D. magna, 17% of D. pulex, and 11% of D. longispna were infected with more than one parasite species. In all three host species, the fecundity of parasitized females was significantly lower than of uninfected females (>20% reduction in D. magna, >25% reduction in D. pulex, and >7% reduction in D. longispina). Only 2 of the 11 common micro-endoparasites found in these three ponds (17 species in total) showed no specificity within the three Daphnia host species; the other nine common parasites infected either only one or two of the three sympatric host species or differed in their host specificity across the three ponds, indicating that the parasites may be specialized for the pond's current or former predominant host community. A few parasite species showed a seasonal pattern (parallel in all three ponds). For example, the microsporidium Thelohania acuta and the protozoan Caullerya mesnili were never found in winter, whereas other parasites showed no such pattern (Stirnadel 1994).

Bittner (2001; Bittner et al. 2002) investigated the parasites of D. galeata and D. hyalina in Lake Constance in southern Germany. The lake has a surface area of 538 km² and a maximum depth of 252 m and contains several planktivorous fish species. In a 3-year study with regular sampling, eight endoparasites (plus one brood parasite) were found with an average prevalence of 5.6% in D. galeata and 15.6% in D. hyalina. Five of these eight parasites reached peak prevalences of more than 20% (up to 50%) in at least one host species. Most of the prevalence peaks were found in fall and winter. The most common parasite was C. mesnili, which was observed to have a strong negative effect on host fecundity. There was no apparent correlation between host density and parasite prevalence.

Decaestecker (2002; Decaestecker et al. 2005) studied parasitism in D. magna over a period of 2 years (April to December each year) in two shallow, eutrophic ponds in Belgium that contain several planktivorous fish species. Eight endoparasite species and six epibiont species were recorded, with microsporidia being the most common group (four species). The overall prevalence of endoparasites was high (95.5% in 1999 and 69.9% in the following year). Severe reductions in fecundity were observed in females infected with Pasteuria ramosa, White Fat Cell Disease, Flabelliforma magnivora, and Ordospora colligata, but hardly any fecundity reduction was found for infections with epibionts. There were no clear seasonal trends in the temporal dynamics, but the sampling period did not cover the entire year. Daphnia density was observed to be negatively related with overall endoparasite prevalence, whereas epibiont prevalence correlated positively with Daphnia density. Interestingly, parasite species that severely reduced host fecundity did not persist as long in the population and had, on average, lower prevalences than benign species.

Wolinska et al. (2004) studied the parasites of the D. galeata x hyalina species complex in Lake Greifensee in Switzerland. This lake harbors several planktivorous fish species. The prevalence of C. mesnili was as high as 22%, and severe effects on host fecundity were observed. Most interestingly, D. galeata x hyalina hybrids were frequently infected, whereas D. galeata was rarely infected. (The other parental species, D. hyalina, was very rare.) The authors speculated that differential parasitism of parental and hybrid taxa may contribute to their coexistence. There was no correlation between host density and parasite prevalence. The authors also reported the occurrence of a bacterial parasite in the haemocoel, which reached a peak prevalence of 7%.

Duffy et al. (2005) studied the dynamics of Spirobacillus cienkowskii infecting D. dentifera in five lakes during a 5-month period. They recorded a marked prevalence peak (up to 12%) in some of these lakes in fall, which coincided with a drastic drop in the predation rate by bluegill sunfish. An epidemiological model, fitted to the particulars of this system, indicated that the drop in predation rate was enough to account for the occurrence of the S. cienkowskii epidemics. Changes in predation pressure cannot, however, explain the strong decline of the epidemics in late fall. The authors speculated that the reduced temperature may cause the termination of the epidemic, but these speculations are not well-supported. Host density as a causative factor for the termination of the epidemic was not discussed.

Mitchell et al. (2004) followed P. ramosa infections in a D. magna population for a period of 4 months in a small farm pond near the Scottish border in UK. Because their paper concerns the coevolution of this system, little information is given on the epidemiology of the system. Pasteuria prevalence increased drastically in mid-August, reached a peak of nearly 30% in late August, and had disappeared by late September.

Comparative Analyses

Table 4.2 gives on overview over all comparative field studies on Cladoceran parasites.

Green (1957) studied parasites and epibionts of Cladocera in 67 rock pool populations in the Skerry islands of southern Finland. These pools were small (3-4 m in length and up to 0.4 m deep) and fishless. Parasite richness declined from D. magna to D. pulex to D. longispina, suggesting that larger Daphnia species harbor more parasites. Green found that some parasite species lowered host fecundity more than others, and in one case, he observed that ephippial females were overparasitized. The author suggested that certain species of epibionts compete with each other for space on the host and thus exclude each other at the population level.

Brunner (1996) (D. Bruner and D. Ebert, unpublished observations) investigated single samples from 43 Daphnia populations in southern England, mainly west of London. Water bodies ranged from small ponds in parks to large natural ponds and medium-sized drinking-water reservoirs. Most of these ponds were fishless. Ninety-one percent of these populations harbored at least one endoparasitic infection (mainly microsporidians). The average prevalence was rather high. In the more common Daphnia species, parasites had an average prevalence of 43% (n = 17) in D. magna, 69.7% (n = 17) in D. pulex, and 43% (n = 9) in D. longispina (all parasite species combined). Among the D. magna populations, average prevalence was 58.4% (standard error of the mean (SE), 8.4) in permanent ponds, and only 23% (SE, 6.4) in intermittent ponds. This difference was, however, most likely attributable to the smaller size of the intermittent ponds. As seen in other studies (Stirnadel and Ebert 1997; Decaestecker 2002), the most common parasites of D. magna were microsporidian gut parasites.

Bengtsson and Ebert (1998) conducted a similar survey with only one sample per pond in a rock pool metapopulation along the Swedish east coast near Uppsala. In these pools, 24 of 50 (48%) D. pulex populations and 9 of 25 (36%) D. longispina populations investigated harbored at least one parasite species. Across all ponds, the average microparasite prevalence was 15.5% for D. pulex and 9.1% for D. longispina (about 30% and 25% when only populations with at least one parasite species are considered). The infections in the pools were primarily attributable to a single, virulent microsporidium species (possibly Larssonia obtusa (Vidtmann and Sokolova 1994)), which reduced clutch size by 98%.

Ebert et al. (2001) studied D. magna in the same rock pool metapopulation in southern Finland as did Green (1957) (see above and Figure 2.18). Because the ecology of this metapopulation is well known, it was possible to address several aspects of parasite distribution across populations in relation to various pool characteristics. Eight endoparasites and eight epibiont species were found in 137 rock pool populations. The number of endoparasite species per population increased with the age of the Daphnia population. Typically, newer populations founded in the year the survey was conducted had no or few parasite species, whereas older populations had increasingly more. Furthermore, large rock pools with presumably larger and more permanent Daphnia populations were more likely to harbor parasites than smaller pools. The most prevalent parasite in the Finnish rock pools was the microsporidium Octosporea bayeri, which often occurred in a prevalence of 100%. This parasite exclusively infects D. magna and was found in nearly 50% of all populations, with much higher percentages in older populations. Surprisingly, Green (1957) found this parasite in only 8.3% of D. magna populations.

What Can We Learn from Prevalence Estimates?

Prevalence estimates are a common and convenient measure of parasite abundance. They allow the investigator to follow changes in parasite abundance over time and provide a reasonable picture of the degree to which the host population is infested. Prevalence estimates have some limitations that have to be taken into account when doing parasitological research. First, they are usually underestimates, because parasites are only detectable after signs of infection have developed. Infections by the microsporidium O. bayeri are only visible 8-12 days after the infection occurs (Vizoso and Ebert 2004). Infections of C. mesnili take 6 days until the parasite is visible (Bittner 2001). Although it is possible to obtain a more accurate measure by keeping the sampled animals for a few days in the laboratory before dissection, this may lead to losses because of mortality before the animals are investigated. Using this method, I found that prevalence estimated in fresh samples might be underestimated by as much as 30% (personal observation). Because most parasites need about 1 week to show the first symptoms, juvenile Daphnia usually appear to be uninfected, even if they contracted the disease within the first day of life. Therefore, most investigators studying Daphnia parasites concentrate on adult animals.

A second problem regarding prevalence is that it correlates with the expected life span of an infection and therefore, when compared across parasite species, can only provide a rough guideline. The investigator will hardly ever see a parasite that kills its host shortly after it produces the first signs of infection. In contrast, parasites that allow their hosts to stay alive for long periods are observed more often, thus showing higher prevalence. Mathematical modeling has shown that, everything else being equal, the more quickly the parasite kills its host the lower is its prevalence (Anderson 1979).

When comparing reports on parasite prevalence in natural populations, one may want to distinguish between studies that were initiated because the investigator had observed high parasite abundance beforehand and those in which populations were screened at random or for other reasons than to study parasitism per se. I know or suspect that the investigations by Green (1974) (but not those in Green 1957, 1964), Stirnadel and Ebert (1997), Brambilla (1983), Yan and Larsson (1988), Vidtmann (1993), and Mitchell et al. (2004) were initiated because rates of parasitism were known to be high. In contrast, this was not the case in the following studies: Brunner (1996), Bengtsson and Ebert (1998), Ebert et al. (2001), and Bittner (2001). Accordingly, the average prevalence estimates in the later studies are mostly lower than in the earlier listed reports. This does, however, show that parasites can be common even in populations that were not specifically chosen because of known high parasite abundance.

Host Body Size and Parasitism

Studies that investigated more than one Daphnia species in a given habitat found that within populations, the larger species were more strongly parasitized than smaller species (Green 1957; Stirnadel and Ebert 1997). Because transmission for many parasite species occurs after the host ingests spores with its food (see Chapter 8, Epidemiology, subsection on Transmission), this relationship may be explained by the considerably larger volume of water that the larger Daphnia filters. However, alternative explanations, such as differential susceptibility, may contribute to this pattern as well.

Effect of Parasites on Individual Hosts

Several studies looked for the effects of parasites on host fecundity and survival. Because Daphnia carry their offspring in their brood chamber for several days before releasing them, clutch size is the most convenient and most often studied trait in relation to infection status. Several studies reported reduced fecundity of infected hosts (Green 1974; Brambilla 1983; Yan and Larsson 1988; Vidtmann 1993; Stirnadel and Ebert 1997; Decaestecker et al. 2005). The degree to which fecundity is reduced varies strongly among parasites, with certain species showing no effect. Interestingly, the number of eggs in a clutch seems to be affected less often than the presence of a clutch in the brood chamber (Bittner 2001; Decaestecker et al. 2005; Ebert et al. 2004; Stirnadel and Ebert 1997). Thus, in many cases it seems that parasites suppress host fecundity totally rather than reducing fecundity to a variable degree. Furthermore, the effect of parasites on host fecundity seems to vary with environmental conditions. For example, Yan and Larsson (1988) found no significant fecundity effect in a large sample of 401 females, whereas in the following year, a smaller sample revealed a strong effect of parasites on host fecundity. Bengtsson and Ebert (1998) found that the degree of fecundity reduction varied across populations.

At least two studies have reported associations (positive and negative) between the production of resting eggs and parasitism. Brambilla (1983) found that ephippial females were never infected, whereas Green (1957) found that ephippial females were relatively more often infected than parthenogenetic females. The association between gender and parasitism certainly needs further investigation.

The effect of parasites on host survival has been tested in field studies by bringing plankton samples to the laboratory, dividing the individuals into infected and uninfected groups, and then monitoring their survival under controlled conditions. For a number of reasons, I consider this approach to be unsatisfactory. First, infected and apparently uninfected hosts may differ in size, age, and experience. Because parasites often influence growth, it is not possible to correct for these differences easily. Second, the assessment of infection status is often difficult, with strong variation across investigators and among diseases. For example, certain infections may only be recognizable shortly before the death of the host, whereas others can be detected a few days after infection. Thus, comparing infected and uninfected animals from field samples does not allow one to judge the effect of parasitism on host survival in a meaningful way.

Infection Dynamics

All of the longitudinal studies found that prevalence varied dynamically over time, with certain parasite species being seen only over short time intervals. In some cases, the dynamics appear cyclic, with seasonal reoccurrence of parasites (mostly in summer), but for the majority of parasite species, it is unclear what determines abundance patterns. Extreme cases of parasite dynamics have been observed in some of the longer studies, where certain parasites disappeared for extended periods of time and then reemerged without any noticeable reason (Green 1974; Bittner 2001). It is totally unclear whether environmental or evolutionary factors play a role in these extreme dynamics.

A few of the longitudinal studies analyzed the dynamics with respect to host density. Thus far, no study has shown a clear density effect, although density-dependent transmission has been shown in the laboratory (Ebert 1995; Bittner et al. 2002). Some studies observed that parasites first appeared when host density was high, but in contrast to what would be expected if dynamics were driven by density dependence, parasites did not decline when host density declined (Brambilla 1983; Yan and Larsson 1988). This trend has also been observed for parasites of planktonic rotifers (Miracle 1977; Ruttner-Kolisko 1977). Several studies suggested that host density and water temperature are to some degree confounded, because the density of most plankton organisms is high during the warmer periods. Therefore, it is not clear to what degree elevated temperatures play a role in summer epidemics (Green 1974; Brambilla 1983). At least for one microsporidium, it has been suggested that low temperature can hinder transmission (Ebert 1995). There are, however, several reports of parasite occurrence at low (winter) water temperatures, indicating that temperature alone cannot explain the occurrence of epidemics (Stirnadel 1994; Bittner 2001; Decaestecker et al. 2005). A possible explanation could be that parasites do not grow at low temperatures but may be able to persist for some time. The relationship between the spread of parasites in relation to host density and water temperature certainly needs further investigation.

The community-level perspective of Decaestecker et al. (2005) revealed remarkable patterns. Daphnia density was observed to be negatively related with overall endoparasite prevalence, whereas epibiont abundance correlated positively with Daphnia density. Furthermore, parasite species that severely reduced host fecundity persisted for shorter amounts of time in the population and had, on average, lower prevalences than benign species. The data did not allow a fine resolution of these patterns, but the following interpretation may explain these findings. Higher host density allows parasites to spread and thus increases prevalence. Thus, harmless parasites (such as epibionts) are more abundant when host density is high. Here it is the host that governs parasite dynamics. However, harmful parasites may at the same time reduce the host population growth rate so much that their net effect on the host population is a reduction in density. This reduction in host density destabilizes the parasite population, which leads to short parasite persistence times. Thus, for harmful parasites, the epidemiological feedback between host and parasite governs the parasite dynamics.

The strong dynamics of many parasite species also indicate that studies that use only one or few samples per population to estimate the richness of the local parasite community are likely to vastly underestimate parasite richness.

Are There Fewer Parasites in Lakes with Fish?

There seems to be a difference in the degree of parasitism in water bodies with and without planktivorous fish. The lower parasite richness and prevalence estimates in lakes with fish predation are underscored by the fact that there are fewer literature reports of Daphnia parasites from lakes with fish. Several factors may work together to explain this fact.

First, the likelihood of infection increases with body size (Vidtmann 1993; Stirnadel and Ebert 1997), which is probably a result of both higher filtration rates (and thus higher uptake rates of parasite spores) and an accumulation effect with age. In ponds with high adult mortality, as is typical for populations with planktivorous fish, the average life expectancy of a Daphnia is low, and thus, parasites may have a lower chance of completing their development. This reduces not only parasite survival but also parasite transmission, because older infected hosts are those that release most (or even all) of the transmission stages. A prediction of this hypothesis is that parasites found in lakes with high predation pressure should complete their development quickly (short prepatent phase) and thus kill their host early. The most virulent Daphnia parasites have been indeed described from habitats with planktivorous fish (Bittner 2001; Wolinska et al. 2004; Duffy et al. 2005). Another prediction is that parasitism rates in lakes should be higher at times when fish predation is low. Indeed, parasitism in Lake Constance is mainly found in fall and winter (Bittner 2001) when predation is strongly reduced, whereas in fishless ponds and lakes, prevalence peaks in summer (Brambilla 1983; Vidtmann 1993; Stirnadel 1994). Duffy et al. (2005) linked the seasonal occurrence of Spirobacillus epidemics in several North American lakes to a drop in predation rate by bluegill sunfish.

Second, some diseases make their hosts more conspicuous through a reduction in transparency, thus increasing the likelihood of predation by visually hunting predators (Lee 1994; Yan and Larsson 1988) (P.T.J. Johnson, personal communication). Similarly, increased susceptibility to predation was reported for hosts carrying large loads of epibionts (Willey et al. 1990; Allen et al. 1993; Chiavelli et al. 1993; Threlkeld et al. 1993). Consistent with this, Willey and Threlkeld (1993) reported a reduction in the prevalence of clearly visible epibionts after stocking with fish. A prediction of this hypothesis is that parasites found in lakes with visually hunting fish should not make their hosts too visible or, if so, only in the terminal phase of infection. Consistent with this, the main parasites of D. galeata and D. hyalina in Lake Constance are hardly visible with the naked eye (Bittner et al. 1998, 2002; Bittner 2001). However, this hypothesis needs further careful examination. A twist to this hypothesis is that in turbid waters with low visibility, infected hosts may not have a reduced life expectancy relative to uninfected hosts (Decaestecker et al. 2005).

Third, fish predators are typically more common in larger ponds and lakes. A number of factors that go hand-in-hand with the size of lakes may limit the spread of parasites. Summer temperatures in larger water bodies may not rise as high as in smaller lakes in the same region, thus influencing parasite development or shortening the season during which parasites can occur (Ebert 1995). Furthermore, the sediment of larger, and in particular deeper, lakes may be a sink for parasite transmission stages. Parasite spores are known to rest in sediment, where they can be picked up by Daphnia (Ebert 1995; Decaestecker et al. 2002). In deep lakes, Daphnia are less likely to come in contact with lake sediment, thus reducing transmission rates. A prediction of this hypothesis is that parasites that rely exclusively on transmission from dead hosts are less likely to be found in deep lakes (see Chapter 8, Epidemiology, section on Transmission), as for example P. ramosa. In deep lakes, transmission from living hosts (e.g., gut parasites) may be much more important for the persistence of parasites.

Fourth, because Daphnia populations in lakes may not reach the density levels of pond populations, parasite transmission may be reduced. Two factors may account for this situation: a) lakes are often less nutrient rich (eutrophic) than ponds, so that lower rates of primary production may limit the maximum density of zooplankton populations; and b) predation by planktivorous fish may influence Daphnia density, and thus parasite transmission, negatively.

Thus, increased parasite mortality in Daphnia populations with fish predators and unfavorable conditions for parasite transmission in larger water bodies may act together to limit the spread of parasites in these Daphnia populations. One should keep in mind, however, that Daphnia communities in fishless ponds and those in lakes with fish are usually made up of different species. For example, whereas D. magna and D. pulex are more common in fishless water bodies, D. galeata, D. hyalina, and D. cucullata are typically lake-dwelling species. Therefore, the question of whether Daphnia populations in fishless water bodies have more parasites requires further critical scrutiny.