Introduction

Water fleas constitute a major component in freshwater food webs. They are keystone consumers of phytoplankton and bacteria, and an important prey for fish, amphibians, and invertebrate predators (Lampert, 1987; Lampert & Sommer, 1997). Their presence or absence often has important ecosystem-level consequences (Scheffer, 1999; Mergeay et al., 2006). Cladocerans are host to numerous ecto- and endoparasites belonging to diverse taxonomic groups (Green, 1974; Ebert, 2005). The most commonly encountered parasites are bacteria and microsporidia, though fungi, nematodes and cestodes have been recorded as well (Green, 1974; Ebert, 2005). Parasites of Daphnia are well-studied (Green, 1974; Ebert, 2005). They have been shown to increase mortality, reduce growth, lower reproductive output, and change migratory behavior of their hosts. As a result, they can regulate the density and, in extreme cases, even lead to the extinction of natural host populations (Hudson et al., 1998; Ebert et al., 2000). Moreover, naturally occurring levels of parasite infections in cladocerans may reduce the quality of food available to secondary consumers, with potentially important effects on food webs (Forshay et al., 2008). Surveys of parasites in cladocerans have only rarely been carried out and never outside Europe and North America. Here, we report the results of a survey in Israel.

The last survey of Cladocera in Israel (including Eastern Sinai which now belongs to Egypt) was conducted in 1993 in 53 sites, and recorded 21 species from the family Daphniidae, of which 11 belong to the genus Daphnia (Bromley, 1993). However, the sources of several sites described by Bromley (1993) were actually much older museum samples (live specimens were unavailable). Since then many freshwater bodies changed or even became uninhabitable for water fleas, because of increased dryness, elevated salinity and the introduction of other species. Habitat changes and loss can significantly affect species diversity (Forró et al., 2008). For example, the introduction of the fish Mugil cephalus and Sarotherodon aureus into Lake Kinneret in the late 1950s/early 1960s appears to have caused the extinction of Daphnia lumholtzi (Gophen, 1979). Observed patterns of species diversity also depend on the selection of habitats. For instance, a comparative survey of cladoceran composition and diversity of habitats in Belgium revealed that small water bodies, such as wheel tracks, pools, ponds and ditches, contribute more to the total cladoceran richness than lakes (De Bie et al., 2008). This was mainly because small water bodies are more abundant. A new Daphnia species from the Golan Heights (a relatively unexplored region in northern Israel) has recently been discovered (A. Petrusek, personal communication). These results stress the importance of sampling a diversity of water bodies of different size, permanence, and flow regimes.

Most studies of parasites in cladocerans have focused primarily on European and North American populations of Daphnia from rock pools and lakes (Bengtsson & Ebert, 1998; Ebert et al., 2001; Decaestecker et al., 2005; Hall et al., 2005; Rodrigues et al., 2008; Wolinska et al., 2009; Duffy et al., 2010; Yin et al., 2012). Although some studies surveyed cladoceran populations in the Levant (a stretch of land adjacent to the eastern shore of the Mediterranean Sea, about 800 km long and approximately 150 km wide, e.g., Güher, 2000; Altındağ et al., 2011), surveys of parasites of Cladocera in the Levant have never been conducted. This study attempts to fill this gap by analyzing cladoceran communities in 204 water bodies throughout Israel. The survey focuses mainly on planktonic cladocerans from the family Daphniidae and covers a sizable fraction of all standing freshwater habitats in the region, ranging from small ephemeral pools to permanent reservoirs.

Materials and methods

The survey was carried out between October 2010 and May 2011, which includes the rain season in Israel. We surveyed 204 permanent water bodies and rain pools (defined as pools with water only after strong rain, and extended period of summer draught). The selection of water bodies for this survey was partly based on sites sampled by Bromley (1993) as well as sites known to inhabit cladocerans (A. Gasith, personal communication and various reports by the Israel Nature and Parks Authority). Animals were sampled by towing a plankton net (opening diameter: 27 cm, mesh size: 200 μm) 10 times in both the pelagial and littoral zones. Each time the net was towed across 3 m. The sampled cladocerans were kept in 0.5 l of water from the sampled water body and transported to the laboratory in a cool box with ice.

All cladoceran specimens were examined alive in the laboratory. We focused on Daphniidae and identified them to the species level as per Benzie (2005). The rest of the taxa were classified at the family level. Parasite infections in each sample were determined in two ways. First, where possible at least 100 live adult individuals of each species were examined for evidence of parasitism using a dissecting microscope at 10×–160× magnification. Most types of infections are clearly visible as an alteration of the color and/or transparency of the host (Green, 1974; Ebert, 2005). Second, at least 20 individuals of each species were crushed and their gut and tissue contents examined under 100×–1,000× magnification using a phase-contrast microscope. The presence of any particular parasite taxon was recorded and classified according to pathological and morphological traits. Sites were categorized as permanent water bodies or rain pools, and by annual rainfall (<400 mm, between 400 and 600 mm, or >600 mm).

Several ecological correlates were examined in this study. First, we compared the frequency of rain pools with cladocerans/endoparasites with that of permanent pools. We then contrasted the mean number of cladoceran/endoparasites taxa per body of water among different rainfall levels. Finally, we checked for correlations between the number/average body size of cladoceran host species and the number of endoparasites, and between the number of parasites per host taxon and host abundance. For statistical analysis, SPSS for Windows v. 19.0.0.1 (SPSS Inc., 2010) was used. Mean values are shown with ±SE. All probabilities are two-tailed. The equality of variances assumption for ANOVA was verified using Levene’s test of homogeneity of variances. The presence/absence of cladocerans in 204 sites and the presence/absence of endoparasites in 128 sites containing cladocerans were used separately as dependent variables in a binary logistic regression, with the categorical predictors permanent water bodies vs. rain pools and annual rainfall as independent variables. These categorical predictors were “dummy” recoded into uncorrelated dichotomous variables (indicator contrasts), before entering them into a regression model (Hosmer & Lemeshow, 2000). We initially constructed a full model that included all variables. We then removed variables with non-significant coefficients one-by-one, thereby deriving a reduced model (Hosmer & Lemeshow, 2000).

Results

Cladoceran abundance and correlates

One hundred and twenty-eight of 204 water bodies (62%) were found to be inhabited by cladocerans. The cladocerans were identified to 14 taxa: five species of Daphnia, two of Simocephalus, two of Moina, one of Ceriodaphnia, one of Megafenestra, one of Ilyocryptus and two more taxa from the families Macrothricidae and Chydoridae. The number of sites in which each taxon was found is presented in Fig. 1. The most abundant taxon was Daphnia magna Strauss (60 sites, 47%), followed by Moina spp. (37 sites, 29%) and Ceriodaphnia sp. (31 sites, 24%). Daphnia chevreuxi Richard, Ilyocryptus sp. and Megafenestra aurita Fischer were only found in one site each. More than half of the sites (56%) were inhabited by two or more taxa. The frequency of rain pools with cladocerans was significantly higher than that of permanent pools (72.5 vs. 36.4%; binary logistic regression, df = 1, P < 0.001). The mean number of cladoceran taxa per body of water was about twice as high in sites where annual rainfall >400 mm than in sites where annual rainfall is lower (F 2,125 = 4.2, P = 0.016; Fig. 2).

Fig. 1
figure 1

Abundance of cladocerans in 128 sites in Israel during the winter 2010/11. The combined abundance of the two Moina species M. brachiata and M. macrocopa is shown

Full size image
Fig. 2
figure 2

Mean (±SE) number of cladoceran taxa inhabiting water bodies in regions with different amount of average yearly rainfall

Full size image

Endoparasites abundance

In 33 out of 128 sites (26%) some cladocerans were found to be infected by endoparasites. Twenty-one parasite taxa were identified (two fungi, four bacteria, 11 microsporidia and four unclassified parasites; Table 1). The number of sites in which each endoparasite taxon was found is presented in Fig. 3. The most commonly found endoparasites were the bacteria Spirobacillus cienkowskii Metchnikoff and the microsporidium Glugoides intestinalis Chatton, which were found in 11 and 9 different sites, respectively. In 14 out of the 33 sites (42%) infected by parasites more than one parasite taxon was found (Fig. 4). Endoparasites were found in seven cladoceran host species (Fig. 5). Of them, D. magna was host to the largest number of endoparasite taxa (10), followed by Simocephalus vetulus O. F. Muller with seven endoparasite taxa and Moina macrocopa Strauss with three endoparasite taxa. We found no correlation between the average body size of host species and the number of parasites infecting it (Pearson r = 0.28, P = 0.34).

Table 1 Parasite taxa found in this survey, their target tissue, transmission mode, abundance (as the percentage of samples from all pools inhabited by Cladocera and the percentage of infected pools from all pools containing the specific hosts of the parasite), hosts, taxonomic classification (called group here), climatic distribution in Israel and known distribution elsewhere
Full size table
Fig. 3
figure 3

Abundance of cladoceran endoparasites in 33 sites in Israel

Full size image
Fig. 4
figure 4

Distribution of single and multiple occurrences of cladoceran parasites per site

Full size image
Fig. 5
figure 5

Number of parasite taxa and corresponding host average body size per host species. (Pearson r = 0.28, P = 0.34)

Full size image

Endoparasites correlates

The number of parasites per host taxon positively correlated with host abundance (Pearson r = 0.76, P = 0.002). The frequency of rain pools with endoparasites was marginally lower than that of permanent pools (binary logistic regression, df = 1, P = 0.08). The largest number of endoparasites was found in sites with intermediate levels of annual rainfall (between 400 and 600 mm; binary logistic regression, df = 2, P = 0.03). Moreover, there was no significant correlation between the number of cladoceran host species and the number of endoparasites within the 128 sites containing cladocerans (Pearson r = 0.15; P = 0.09) and within the 33 sites containing endoparasites (Pearson r = 0.15; P = 0.41). Figure 6 shows the probability of encountering a site with infected animals for each cladoceran taxon, assuming the species inhabited the site. For D. magna and S. vetulus the chance to find an infected population is approximately 40 and 55%, respectively. In 10 sites (eight cases with D. magna, two cases with S. vetulus) we recorded hosts infected by multiple parasite taxa simultaneously (between 2 and 4 parasite taxa per host).

Fig. 6
figure 6

Percentage of infected populations per host species. The combined percentage of the two Moina species M. brachiata and M. macrocopa is shown

Full size image

Discussion

Cladocerans

In the present survey, we found considerably fewer species from the cladoceran family Daphniidae than in previous studies (9 vs. 13), see Fig. 7; Yaron, 1964; Bromley, 1993). The abovementioned species count is after correcting for changes in the taxonomic statue of some of the species recorded in Israel and the resulting unification of several species into one (e.g., D. ulomskyi = D. atkinsoni, M. rectirostris = M. brachiata; Goulden, 1968). We also detected changes in the distribution of some Daphniids in comparison to past studies. For instance, the most abundant Daphniid in our survey was D. magna, which was found in 47% of the sites inhabited by cladocerans. Bromley (1993) found that D. magna was the most abundant Daphniid as well, although it was previously found in only 24% of the sites in Israel. Daphnia curvirostris Eylmann was found in only 4% of the sites sampled by Bromley (1993) and in none of the sites sampled by Yaron (1964), yet it was the second most common Daphnia in this study (20% of sites). Both Daphnia similis Claus and Daphnia atkinsoni Baird were also more abundant now than in previous studies. M. aurita was first reported (as Scapholeberis aurita) from the Hula Reserve in Northern Israel (Bromley, 1981), but in this survey it was only found in the central coastal plain of Israel.

Fig. 7
figure 7

Comparison of Daphniidae diversity between this study and the survey by Bromley (1993)

Full size image

There are several possible explanations for these differences in cladoceran abundance and distribution. First, only a quarter of the sites sampled in the last two comprehensive surveys by Yaron (1964) and Bromley (1993) were sampled in this study. This is because some sites no longer exist while others are now part of Egypt and thus inaccessible. However, since we sampled more than 190 sites that were not sampled previously, we believe that local extinction of cladocerans is a more likely explanation, given that our survey encompassed almost fourfold more sites than the surveys of Yaron (1964) and Bromley (1993). Second, Gophen (1979) reported the disappearance of Daphnia lumholtzi G. O. Sars from Israel due to the intentional introduction of fish to Lake Kinneret. Habitat loss of wetland area is considered a major cause of species decline (Lehtinen et al., 1999; Brooks et al., 2002). Jenkins et al. (2003) found that conversion of the wet prairie to agriculture may have reduced crustacean metapopulations to isolated populations that are more vulnerable to habitat loss. In Israel wetlands are the most endangered habitats. Levin et al. (2009) estimated that since the nineteenth century, over 75% of the swamps and rain pools that existed in the central coastal plain of Israel have disappeared due to drainage, transformation to agricultural lands and built up areas. In addition to the reduction in wetland numbers, human activity has also diminished wetland size. In the past rainy winters wetland areas in the coastal plain were an order of magnitude larger than they are today. Finally, a major contributor to cladoceran richness in Israel was the Hula wetland, which was largely drained in the 1950s (Bromley, 1993).

Interestingly, the mean number of cladoceran taxa per body of water was much lower in arid areas (rainfall <400 mm; Fig. 2). To the best of our knowledge, cladocerans have not been sampled from such arid habitats in Europe and North America. These habitats present substantial challenges to cladoceran communities due to long-term draughts. They also undergo fluctuating periods of rain and dryness during a single winter season, because of low rainfall levels coupled with extreme weather conditions. Consequently, it is likely that these conditions lead to increased competition and constrain coexistence of species.

Endoparasites abundance and richness

This study documents for the first time the endoparasites of Cladocera in Israel and their distribution. We found 21 species in 26% of the water bodies inhabited by cladocerans, including 13 species that were not reported previously (two bacteria species, seven microsporidia species, and four unidentified parasites). The largest number of parasites found in one site was six, which is a considerably lower than in other studies (Stirnadel & Ebert, 1997; Decaestecker et al., 2005). We may have very well underestimated parasite species richness because the water bodies in this survey were only sampled once. Although the largest source of variation in parasite prevalence in Daphnia populations is usually attributed to the species identities of host–parasite pairs and to their spatial variation (Duffy et al., 2010; Wolinska et al., 2011), variations in host density (Vidtmann, 1993) and predation pressure (Duffy et al., 2005), as well as seasonal and temporal variations (Green, 1974; Stirnadel & Ebert, 1997) also affect parasite richness and prevalence. The focus of this study, however, was on identifying occurrence in one particular season from many locations.

Parasite richness among hosts and sites varied strongly. Only five Daphniid species and two Moina species were found to be infected by endoparasites. D. magna was infected by the largest number of parasites, followed by S. vetulus, Moina spp. and Ceriodaphnia sp. D. atkinsoni and D. curvirostris had the lowest parasite richness. A probable explanation for the large number of parasites infecting D. magna is that it was by far the most common cladoceran and therefore its encounter rate with parasites is expected to be higher. However, this does not explain the high richness of parasites in S. vetulus, which was found in only nine sites but was infected by five different parasite taxa. Nor does it explain the very low richness of parasites in Ceriodaphnia sp., which was quite abundant (two parasites and 31 sites), and the fact that we never found infected D. similis in over 20 sites. An alternative explanation would be differences in body size and behavioral feeding strategies of the seven host species. Previous studies suggested that being a generalist feeder (benthic and planktonic feeding) and being of large size increases the risk of infection (Green, 1957; Vidtmann, 1993; Stirnadel & Ebert, 1997). However, we did not find such correlation (i.e., between the average body size of host species and the number of parasites infecting it). Host–parasite specificity and genetic differences in host susceptibility also play a role in determining parasite richness and infection success (Carius et al., 2001; Duffy et al., 2010; Luijckx et al., 2012).

Endoparasites distribution and host range

Eight of the parasites observed in this study have already been described. However, since most studies of parasitism in Cladocera were carried out in North and Central Europe and North America, in cooler and wetter climate zones than Israel, this survey extends their known geographic distribution, and in many cases the number of recorded host species.

For example, the southernmost edge of the distribution of the bacterium Pasteuria ramosa was Maryland, USA (Ebert, 2005), and it was never found to infect D. curvirostris as it did in Israel. The bacterium S. cienkowskii infects the hemolymph of a wide range of cladoceran hosts and has been described from sites in Africa, North America, and Europe (Ebert, 2005). In Israel it was found to infect D. magna in all climate zones, including arid areas with <150 mm annual rainfall, as well as D. atkinsoni and S. vetulus.

The microsporidium Hamiltosporidium tverminensis Haag (formerly Octosporea bayeri) is known from Finland (Haag et al., 2011) and infects the adipose tissue, gonads and hypoderm of D. magna. It is transmitted both vertically and horizontally. In six different locations along the Mediterranean and in arid areas of Israel, we found an H. tverminensis “like” parasite. The microsporidium Hamiltosporidium magnivora Larsson (formerly Flabelliforma magnivora) is known from Western Europe (Larsson et al., 1998; Haag et al., 2011). It infects D. magna and resembles H. tverminensis in its pathology and target tissue, but differs in its transmission strategy, i.e., in the laboratory it is transmitted only vertically. In Israel, we found H. magnivora “like” microsporidium in two locations in the Golan Heights. G. intestinalis is a gut microsporidium that infects D. magna and D. pulex in Europe and North America (Green, 1974). In Israel, we found a G. intestinalis “like” microsporidium with the same pathology and a wider range of hosts (D. magna, D. curvirostris, and S. vetulus) and habitats (Mediterranean and semi-arid areas). Ordospora colligata Larsson is gut microsporidium that infects D. magna in Western and Northern Europe (Larsson et al., 1997). It resembles G. intestinalis but the spores are pyriform and slightly larger. We found O. colligata in three sites in the arid, semi-arid, and Mediterranean areas of Israel. Agglomerata cladocera Pfeiffer is another microsporidium that infects the hypoderm of D. magna (Larsson et al., 1996). In this study, we found a microsporidium that resembles A. cladocera in its size, shape and pathology, and infected D. magna. A. cladocera is known only from Europe (Larsson et al., 1996), but in this survey we extend its distribution to the Mediterranean regions of Israel.

A fungal brood parasite that has not been identified taxonomically was found to infect D. dentifera Forbes in Southwest Michigan, USA (Hall et al., 2005; Duffy et al., 2010). A similar parasite was found in this survey in two proximate sites, but in different hosts: D. curvirostris and S. vetulus. This parasite appears as a white mass in the brood pouch of its host and destroys the eggs. A female that dumps the infected brood is able to reproduce again. Another brood pouch parasite of Daphnia was reported from Europe by Green (1974) and Tellenbach et al. (2007), but it differs in its pathology from the one found in this study.

Metschnikowia bicuspidata Metschinkoff is an endoparasitic Ascomycete that infects the hemolymph of its host (Green, 1974). It has been recorded from D. magna, D. pulex, and D. longispina as well as from a number of other crustaceans in Europe and North America (Ebert, 2005). In this survey, it was found to infect D. magna in five different locations along the coastal plain of Israel. We also observed it in the copepod Arctodiaptomus similis.

Multiple infections by endoparasites

In many natural populations hosts are commonly found to be infected by more than one parasite species (Lello et al., 2004; Decaestecker et al., 2005; Rutrecht & Brown, 2008). In this survey, co-occurrence of parasites was found in 14 (42%) of the sites inhabited by cladocerans. In 10 (71%) of these sites we found multiply infected host individuals. Multiple infections play a significant role in the pathogenic processes occurring within the host. They can complicate the diagnosis and the prognosis of specific diseases, and are relevant to the evolution of both hosts and disease processes (Petney & Andrews, 1998). Future studies should focus on understanding temporal, spatial and between-host dynamics of multiple infections by different parasite species in natural populations. Such multiple infections may have important consequences for community structure as well as host–parasite coevolution (Ben-Ami et al., 2008, 2011).