2
views
0
recommends
+1 Recommend
1 collections
    0
    shares
      • Record: found
      • Abstract: found
      • Article: found
      Is Open Access

      Gastrointestinal and external parasites of the Variable hawk Geranoaetus polyosoma (Accipitriformes: Accipitridae) in Chile Translated title: Parasitas gastrointestinais e externos do Falcão Variável Geranoaetus polyosoma (Accipitriformes: Accipitridae) de Chile

      Read this article at

      Bookmark
          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

          Abstract

          Abstract Information about parasites associated with diurnal raptors from Chile is scarce. Between 2006 and 2017, a total of 15 specimens of the Variable hawk, Geranoaetus polyosoma (Quoy & Gaimard, 1824) were collected, 14 of them from different localities in the Biobío region and one specimen from the Valparaíso region. An external examination of the plumage was made to collect ectoparasites, and necropsies were performed, focusing primarily on the gastrointestinal tract. Chewing lice (Phthiraptera) were found on five (33.3%) of the birds corresponding to three species: 97 specimens of Degeeriella fulva (Giebel, 1874), six specimens of Colpocephalum turbinatum Denny, 1842 and nine belonging to an unidentified species of the genus Craspedorrhynchus Kéler, 1938. Endoparasites found in three (20%) of the birds included round worms (Nematoda) of the genus Procyrnea Chabaud, 1958, and spiny-headed worms (Acanthocephala) of the genus Centrorhynchus Lühe, 1911. The species Colpocephalum turbinatum and the genera: Craspedorrhynchus sp., Procyrnea sp. and Centrorhynchus sp. are new records for the Variable hawk.

          Translated abstract

          Resumo No Chile, informações sobre parasitas associados a aves de rapina diurnas são escassas. Entre os anos 2006 e 2017, um total de 15 espécimes do Falcão Variável Geranoaetus polyosoma (Quoy & Gaimard, 1824) mortos, foram examinados, 14 deles provenientes de diferentes localidades da região do Biobío e um espécime na região de Valparaíso. Um exame externo da plumagem foi feito para coletar os ectoparasitas e necropsias do tracto gastrointestinal para coleta de endoparasitas. Cinco aves (33,3%) foram positivas para três espécies de piolhos (Phthiraptera): 97 espécimes de Degeeriella fulva (Giebel, 1874), seis espécimes de Colpocephalum turbinatum Denny, 1842 e nove espécimes não identificados do gênero Craspedorrhynchus Keler, 1938. Endoparasitas foram encontrados em três aves (20%), incluindo vermes redondos (Nematoda) do gênero Procyrnea Chabaud, 1958, e vermes achatados (Acanthocephala) do gênero Centrorhynchus Lühe, 1911. As espécies Colpocephalum turbinatum e os dos gêneros Craspedorrhynchus, Centrorhynchus e Procyrnea corresponderam a novos registros para o Falcão Variável.

          Related collections

          Most cited references 47

          • Record: found
          • Abstract: found
          • Article: not found

          The sixth mass coextinction: are most endangered species parasites and mutualists?

          The effects of species declines and extinction on biotic interactions remain poorly understood. The loss of a species is expected to result in the loss of other species that depend on it (coextinction), leading to cascading effects across trophic levels. Such effects are likely to be most severe in mutualistic and parasitic interactions. Indeed, models suggest that coextinction may be the most common form of biodiversity loss. Paradoxically, few historical or contemporary coextinction events have actually been recorded. We review the current knowledge of coextinction by: (i) considering plausible explanations for the discrepancy between predicted and observed coextinction rates; (ii) exploring the potential consequences of coextinctions; (iii) discussing the interactions and synergies between coextinction and other drivers of species loss, particularly climate change; and (iv) suggesting the way forward for understanding the phenomenon of coextinction, which may well be the most insidious threat to global biodiversity.
            Bookmark
            • Record: found
            • Abstract: found
            • Article: found
            Is Open Access

            Parasites Affect Food Web Structure Primarily through Increased Diversity and Complexity

            Introduction Ecological network research is a powerful framework for assessing ecosystem organization, dynamics, stability, and function, topics that are central to ecology [1]–[7]. For example, comparative studies of food web structure have revealed regularities in how consumer–resource interactions (Box 1) among species are organized [8]–[12], produced successful simple models to characterize such structure [13]–[16], and supported research on the robustness (Box 1) of food webs to species loss [17]–[20]. These and other insights, however, have been largely based on analyses of interactions among free-living species, and have generally neglected parasites. Parasites comprise a significant part of the earth's biodiversity [21], can achieve substantial biomass in some ecosystems [22], can have similar abundance and productivity to free-living species of comparable body size and trophic level [23], and likely extend the generality of the metabolic theory of ecology [24]. Further, in terms of their trophic relations, parasites have consumer–resource body-size ratios inverse to those of most free-living predators [23], which enhances their ability to regulate host species abundances [25]; they have durable physical intimacy with their hosts [26]; they often have complex life cycles, sometimes requiring multiple phylogenetically distant hosts of widely varying body sizes over a lifetime [27]; they may have different patterns of trophic specialization than free-living predators [28]; they may differentially associate with hosts in different topological positions in food webs [29],[30]; and their manipulation of hosts can reorganize communities and alter ecosystem function [31]. These and other ecological factors might alter how parasites fit into, and affect the structure of, food webs compared to free-living organisms. For example, although some parasites appear to be trophic generalists (Box 1), when their hosts are aggregated over their whole life cycle, they are actually temporal serial specialists (Box 1), with particular hosts at particular life stages [32]. Taking this into account increases the likelihood that primary species loss will lead to secondary extinction of such parasites and also decreases the robustness of the food web in question [32]–[35]. In general, the great diversity and unique habits and roles of parasites suggest that their explicit inclusion in food webs may alter our understanding of species coexistence and ecosystem structure, stability, and function [35]–[40]. Box 1. Glossary Complexity: In most food web studies, complexity refers to simple relationships between the number of feeding links L and the number of taxa S in a web, particularly link density (L/S) and connectance (C) (Table 1). Consumer–resource interaction: An interaction whereby an individual of species A (the consumer) feeds on an individual of species B (the resource), resulting in a transfer of biomass from B to A. It includes all types of feeding interactions, such as predator–prey, herbivore–plant, parasite–host, and detritivore–detritus. Concomitant links: Trophic links from a free-living consumer to the parasites of its resources [38],[45],[66]. Degree distribution (cumulative): The proportions of species P(k) that have k or more trophic links in a food web [8],[10]. This study focuses on the resource distribution, the numbers of links to resource taxa (i.e., numbers of resource taxa per consumer), and the consumer distribution, the numbers of links to consumer taxa (i.e., numbers of consumer taxa per resource). The resource distribution reflects the balance of specialists and generalists in a food web, while the consumer distribution reflects the balance of invulnerable and vulnerable species in a food web. Diversity: In most food web studies, diversity is measured as species richness S, the number of taxa (nodes) in the web. Food web: The network of feeding interactions among co-occurring taxa in a particular habitat. Generalist: A consumer taxon that feeds on multiple resource taxa. Generality: How many resource taxa a consumer taxon has. MaxEnt model: A model that generates the least biased probability distributions by maximizing the information entropy for a system after applying information-containing constraints [71]. In the current study, it is applied to degree distributions to provide a null expectation for the shape of food web consumer and resource distributions [62]. Motifs: In this study, the 13 unique link patterns (including both single- and bidirectional links) that can occur among three taxa, excluding cannibalistic links. The frequency of a motif in an empirical web is compared to its frequency in an ensemble of randomized webs to determine whether the motif is under- or overrepresented in the empirical web or a set of model webs [11]. Network structure: The patterns of how links are arranged among nodes in a network. In food webs, it refers to patterns of trophic interactions among taxa. Niche model: A simple one-dimensional model of food web structure. S and C (Table 1) are used to specify the number of trophic species and links in a model web. Each species i is assigned a niche value ni drawn randomly and uniformly from the interval [0,1], and it consumes all species within a feeding range ri that is a segment of the interval, which is placed on the interval such that its center ci is equal to or lower than the niche value ni [13]. The niche model is notable for assuming a contiguous trophic niche for consumers. Probabilistic niche model: A model that parameterizes the niche model directly to an empirical food web dataset [63],[64]. It produces an MLE of the fundamental niche model parameters (ni , ri , ci ) for each species i in a given web. This allows computation of the probability of each link in an empirical web according to the model, and the overall expected fraction of links (ƒL) predicted correctly (Table 1, Metric 22). It can be extended to more than one dimension. Scale dependence: The empirically well-corroborated hypothesis that most food web structure metrics (Table 1, Metrics 6–22) and properties such as degree distribution change in systematic and predictable ways with the diversity (S) and/or complexity (L/S, C) of a food web (Table 1, Metrics 1–5). This scale dependence is built into models such as the MaxEnt and niche models through their use of S and C as the fundamental parameters. In addition, the fit of models to observed food webs also displays scale dependence, tending to decrease with increasing diversity or complexity. Specialist: A consumer taxon that has very few possible resource taxa. In its strongest sense it refers to species that have specialized feeding on one other species. Robustness: The proportion of primary extinctions that leads to a particular proportion of total extinctions, equal to primary plus secondary extinctions [17],[61]. A consumer species goes secondarily extinct if it loses all of its resource species. When assessed just based on food web network structure, robustness may be referred to more specifically as structural robustness. Trophic species: Groups of taxa within a food web that share the same set of consumers and resources [65]. A trophic species web is generated from an original species web (i.e., the original dataset) by aggregating such taxa into single nodes. Most comparative food web structure studies focus on trophic species webs to reduce bias due to uneven resolution of taxa within and across food web datasets and to focus analysis and modeling on functionally distinct taxa. Vulnerability: How many consumer taxa a resource taxon has. Consistent with these types of expectations, prior studies of the network structure of food webs that include parasites have suggested that adding parasites alters food web structure [41]–[49]. This type of thinking is rapidly becoming conventional wisdom, as evidenced by a statement in a 2013 paper in Trends in Ecology and Evolution that “recent advances have shown that native parasites dramatically alter food web structure” [50]. However, there are two problems with this assertion. First, prior studies of parasites in food webs do not distinguish between changes in diversity and complexity and changes to network structure (Box 1). In food web studies, measures of diversity, such as species richness (S), and of complexity, such as link density (links per species, L/S) and connectance (the proportion of possible links actually observed, C), provide simple ways to characterize the numbers of nodes and links in those networks (Table 1, Metrics 1–4). However, in the general [51] and ecological [6] network literature, network structure refers to patterns of how links are distributed among nodes. As noted in a recent perspective in Science, “Network approaches to ecological research emphasize the pattern of interactions among species (the way links are arranged within the network)” [6]. While adding parasites, or any species, to food webs necessarily increases the numbers of species and links and can alter link density and connectance [45], such changes to diversity and complexity should not be characterized as changes in food web structure. Second, while adding parasites and their links generally does alter network structure properties, as noted by prior studies for a few metrics [41]–[49], there is usually an assumption that such changes result from unique aspects of parasite biology. However, those studies did not account for generic structural effects of adding any type of species and their links to a food web. One of the key insights of the last dozen years of comparative food web research regards the scale dependence (Box 1) of food web structure, which refers to the empirically well-supported hypothesis that most aspects of network structure change systematically with changes in the diversity and complexity of food webs, regardless of the identity of the species in the webs [52]–[56]. Thus, the overall hypothesis we test is whether changes to network structure arising from the addition of parasites to food webs are attributable to the unique trophic roles that parasites play in food webs, or, alternatively, are generic effects of adding any type of species and links to webs. We conducted comparative analyses of the structure of seven highly resolved food webs that include detailed metazoan parasite data [42],[57]–[60]. The food webs are from coastal areas and include a variety of habitats including estuaries, salt marshes, tidal basins, and mudflats. We assessed many metrics of food web structure (Table 1, Metrics 6–22) as well as degree distributions (Box 1) and motifs (Box 1), most of which have not been evaluated previously for food webs with parasites. To our knowledge, this is the broadest set of food web structure properties yet evaluated in a single study. Together they provide a wide range of ways to understand network structure, from system-level properties to types of taxa present in the system to local structure to the occurrence of specific links. We did not analyze robustness (Box 1) [17],[61], as it has been explored extensively for food webs with parasites elsewhere [32]–[34], including an analysis of the seven food webs studied here [35]. That literature includes the only other study known to us that sought to disentangle generic from unique effects of parasites on network structure, by analyzing “whether the reduction in food web robustness after the inclusion of parasitism is due to factors associated with the characteristics of parasites, or simply an inevitable artefact of the addition of new nodes and links to an existing network” [34]. By comparing models with similar species richness (S) and connectance (C), that study showed that only those models that incorporated parasite life-cycle constraints resulted in substantial reductions in robustness as well as higher vulnerability of parasites to random species loss. Thus, the general finding of reduced robustness of food webs with parasites to species loss [32]–[35] was attributed to the complex life cycles of many parasites, rather than to generic changes in S and C [17],[54]. We also used a model-based strategy to assess whether changes in food web properties due to the addition of parasites are attributable either to their unique trophic roles or to generic effects of adding any species. The MaxEnt model for degree distributions [62], the niche model [12],[13], and the probabilistic niche model [63],[64] (see Box 1 for brief definitions of the three models) incorporate scale dependence. In particular, the MaxEnt and niche models use S and C as input parameters, while the probabilistic niche model matches S and C of empirical webs. The scale dependence of structure implicit in those models has been corroborated by analyses that show that these and related models generate networks with structure similar to that observed in empirical food webs [13]–[16],[62],[64]. The current study uses these models as a normalization tool—they provide a way to meaningfully compare the structural properties of empirical webs with different numbers of species and links, and they have been critical in identifying generalities in food web structure across space and time [10],[11],[54],[55]. In addition, these models display a fit to empirical data that is scale dependent, with decreasing model fit associated with food webs that have greater diversity and complexity. This second form of scale dependence of food web structure provides another way to assess whether parasites have generic or unique impacts on structure. 10.1371/journal.pbio.1001579.t001 Table 1 Food web metrics. Metric Number Metric Name Definition 1 S Species richness Number of taxa (nodes) in a food web. 2 L Trophic links Number of feeding interactions (links or edges) between taxa in a food web. Trophic links are directional, such that “A feeds on B” is a separate link from “B feeds on A.” 3 L/S Link density Mean number of links per species. 4 C Connectance Proportion of possible trophic links that are realized. The most conventional algorithm is “directed connectance,” C = L/S 2, where S 2 is the number of possible links among S taxa, and L is the observed number of links [70]. 5 C adj Adjusted connectance An alternate connectance measure, C adj = L/(F•S), where F is the number of free-living species, used to measure connectance in food webs when excluding links from free-living to parasite species [45]. 6 Top Top taxa Fraction of taxa that lack consumers. 7 Int Intermediate taxa Fraction of taxa that have both consumers and resources. 8 Bas Basal taxa Fraction of taxa that lack resource taxa. 9 Herb Herbivores Fraction of taxa that feed only on basal taxa. This includes detritivores, taxa that feed on detritus (non-living organic matter). 10 Omn Omnivores Fraction of taxa that feed on resource taxa that occur on more than one trophic level. 11 Can Cannibals Fraction of taxa that feed on individuals from the same taxon. 12 Loop Species in loops Fraction of taxa that occur in loops, excluding cannibals, e.g., when A eats B, B eats C, and C eats A, all three taxa occur in a loop. 13 LinkSD Link number standard deviation Standard deviation of the number of links per species. 14 GenSD Generality standard deviation Standard deviation of the number of resources per species. 15 VulSD Vulnerability standard deviation Standard deviation of the number of consumers per species. 16 TL Trophic level A measure of how many steps energy must take to get from an energy source to a focal taxon. Basal taxa are assigned TL = 1, obligate herbivores thus have TL = 2, and higher level consumers have TL averaged across the multiple food chains connecting them to basal taxa. The algorithm used here is “short-weighted trophic level,” the average of a consumer's shortest trophic level (1+shortest chain to a basal taxon) and its prey-averaged trophic level (1+the mean TL of all of its resources) [94]. 17 MaxSim Mean maximum similarity The mean of all species' largest similarity index, which is calculated as the number of consumers and resources shared in common divided by the pair's total number of consumers and resources [13]. 18 Path Mean shortest path length Mean of the shortest chain of feeding links (regardless of link direction) connecting each pair of taxa in a food web [8],[9]. A simple measure of how quickly effects can spread throughout a food web. 19 Clus Clustering coefficient Average fraction of pairs of species one link away from a particular species also linked to each other [8]. 20 ƒ G Degree distribution goodness of fit Goodness of fit of a degree distribution, where ƒG≤0.95 indicates that an empirical degree distribution is not significantly different from the model distribution at the 95% confidence interval [62]. 21 W 95 Degree distribution relative width Relative width of a degree distribution, where −1≤W 95≤1 indicates that an empirical distribution is neither significantly narrower (W 95 1) than the distribution predicted by a model at the 95% confidence interval [62]. 22 f L Fraction of links Fraction of specific links in an empirical food web predicted correctly by a model [63],[64]. To summarize, our study improves on prior studies in the following ways: it distinguishes changes in diversity and complexity from changes in network structure; it accounts for the generic effects of the addition of species and links on food web structure; it examines a wide range of local to system-level structural properties; it uses trophic species aggregation (Box 1) [65], which is a necessary step for model-based comparative analysis [10]–[16]; it considers the role of concomitant links (Box 1), the numerous trophic links that occur when a predator concurrently eats parasites infecting its prey [38],[47],[66]; and it analyzes seven highly resolved webs, compared to the one to five webs of previous studies, some of which lacked high resolution and/or comprehensiveness. Our results underpin a more comprehensive assessment than previously undertaken of whether adding parasites alters food web structure in unique ways and whether parasites play similar or different roles compared to other consumers and resources in ecological networks. Teasing apart the generic effects of increased diversity and complexity on observed food web structure from the specific effects of the unique topological roles of parasites, or other types of organisms not considered here, is an important and necessary step for developing a fundamental understanding of ecological networks that includes a more detailed accounting of the full diversity of ecosystems. Results Diversity and Complexity We analyzed three versions of each web, one without parasites, one with parasites but no concomitant links (Box 1), and one with parasites and concomitant links. Each original species web version was aggregated into a trophic species web (Box 1), used as the basis for comparative network structure analyses. Species richness (S; Table 1, Metric 1) of the seven trophic species webs without parasites ranged from 56 to 117 (Table 2). The number of trophic links (L; Table 1, Metric 2) in the webs ranged from 358 to 1,085 (Table 2). Adding parasites increased S 1.2 to 1.9 times (range of 109 to 185) and L 1.4 to 3.4 times (range of 576 to 2,838), while adding concomitant links increased L 1.8 to 5.7 times (range of 1,252 to 4,671). S was reduced by seven to 33% and L by four to 51% in trophic species webs compared to original species webs (Table S1). The majority of the metazoan parasites (72% to 100%) in the original species webs have complex life cycles, where the parasites use two or more sequential hosts [27]. Those trophic shifts are often accompanied by an abrupt ontogenetic change in parasite morphology [67]. The use of sequential hosts by many of the metazoan parasites in these webs contrasts with the high degree of trophic specialization (i.e., only one host) reported for parasitoids in other ecological networks [68],[69]. In addition, the current webs have a large number of trematode parasites that tend to have relatively low specificity for the final host. 10.1371/journal.pbio.1001579.t002 Table 2 Basic properties of trophic species food webs. Food Web–Type S L L/S C C adj S Free S Par S Bas Fals–Free 80 527 6.59 0.082 — 1.00 0.00 0.11 Fals–Par 141 1,792 12.71 0.090 0.138 0.65 0.35 0.06 Fals–ParCon 142 3,006 21.17 0.149 — 0.65 0.35 0.06 Carp–Free 91 761 8.36 0.092 — 1.00 0.00 0.10 Carp–Par 154 1,982 12.87 0.084 0.131 0.64 0.36 0.06 Carp–ParCon 154 3,350 21.75 0.141 — 0.64 0.36 0.06 Punt–Free 106 1,085 10.24 0.097 — 1.00 0.00 0.08 Punt–Par 185 2,838 15.34 0.083 0.131 0.63 0.37 0.05 Punt–ParCon 185 4,671 25.25 0.136 — 0.63 0.37 0.05 Flens–Free 56 358 6.39 0.114 — 1.00 0.00 0.11 Flens–Par 109 846 7.76 0.071 0.114 0.62 0.38 0.06 Flens–ParCon 109 1,252 11.49 0.105 — 0.62 0.38 0.06 Otag–Free 94 751 7.99 0.085 — 1.00 0.00 0.03 Otag–Par 117 1,054 9.01 0.077 0.090 0.85 0.15 0.03 Otag–ParCon 118 1,354 11.47 0.097 — 0.85 0.15 0.03 Sylt–Free 117 993 8.49 0.073 — 1.00 0.00 0.05 Sylt–Par 147 1,708 11.62 0.079 0.098 0.80 0.20 0.04 Sylt–ParCon 149 2,680 17.99 0.121 — 0.79 0.21 0.04 Ythan–Free 81 394 4.86 0.060 — 1.00 0.00 0.05 Ythan–Par 122 576 4.72 0.039 0.056 0.69 0.31 0.03 Ythan–ParCon 122 1,284 10.52 0.086 — 0.69 0.31 0.03 Fals, Carp, Punt, Flens, Otag, Sylt, and Ythan refer to the food webs for Bahia Falsa, Carpinteria Salt Marsh, Estero de Punta Banda, Flensburg Fjord, Otago Harbor, Sylt Tidal Basin, and Ythan Estuary, respectively. “Free” refers to webs with free-living species only; “Par” refers to webs with parasites but not concomitant links; “ParCon” refers to webs with parasites and concomitant links. S, L, L/S, C, and C adj are defined in Table 1 (Metrics 1–5). S Free, S Par, and S Bas refer to the fraction of taxa that are free-living, parasite, and basal, respectively. Parasites comprised 15%–28% of taxa and were involved in 22%–74% of links, while free-living species were involved in 91%–100% of links in trophic species webs (Table S2), similar to original species webs (Table S3). Links can be divided into four categories based on the different possible relationships between free-living species (FL) and parasite species (Par): classic predation (FL-FL), classic parasitism (Par-FL), parasites consuming parasites (Par-Par), and predation of parasites (FL-Par) (Table S2). In trophic species webs with parasites, classic predation comprised 42%–78% of links, classic parasitism comprised 13%–38%, parasites consuming parasites comprised 0) or underrepresentation ( 0) or underrepresentation ( 1,500 links (i.e., most of the webs that include parasites), a minimum ƒL of ∼0.50 appeared to hold (Figure 3C). A possible lower bound on ƒL in relation to L was suggested in an earlier study [64]. Using maximum likelihood estimates (MLEs) of niche model parameters, we ordered consumers by the position of their feeding range (ci ) along the x-axis in Figure 5, with their resources ordered by their niche value (ni ) along the y-axis, and then marked documented links at the intersection of consumers and resources. This provides visualization of whether the resources of generalists tend to be dispersed along the niche axis or are concentrated with a near-contiguous core (referred to hereafter as “trophic niche structure”), and whether parasite feeding ranges tend to clump or disperse along the niche axis (Figure 5). The trophic niche structure of generalists in the web without parasites showed that their resources' most likely niche values tended to arrange in a nearly contiguous core interval of niche space (Figure 5A), with gaps (i.e., discontinuities in a column of links) occurring more frequently towards the edges of the consumer's trophic niche, consistent with previously studied webs [64]. When parasites were added, the most likely feeding range positions of most parasites tended to group together (Figure 5B). The parasites with multiple hosts also displayed a core trophic niche structure, but compared to those of generalist free-living consumers, parasites' links to resources spread across a larger interval of niche space, there were more gaps in their trophic niches, and in some cases there appeared to be secondary trophic niches separated from the main trophic niche. When concomitant links were added (Figure 5C), the parasites with multiple hosts displayed similar patterns, and the breadth of trophic niches of generalist free-living species expanded greatly but still appeared to have a single nearly contiguous core. All seven webs displayed qualitatively similar patterns (Figures 5 and S4, S5, S6). 10.1371/journal.pbio.1001579.g005 Figure 5 Visualization of trophic niches of species in Estero de Punta Banda food webs. MLE values for consumer niche position (c) are on the x-axis and for resource niche value (n) are on the y-axis. (A) Results for the web with free-living species only. (B) Results for the web with parasites but not concomitant links. (C) Results for the web with parasites and concomitant links. Red dots show the resource links for free-living consumers, and blue dots show the resource links for parasite consumers. Discussion Prior claims that parasites affect food web structure differently from free-living consumers either focused on changes to diversity and complexity when parasites were added, or did not control for the effects of increases in diversity and complexity on network structure properties. Our study clarifies the distinction between changes in food web diversity and complexity and changes in food web structure, which consists of the patterns of how feeding links are distributed among species [6]. We assessed both aspects of change in food webs when parasites were added, as discussed separately below. Our most novel and important findings concern network structure, and whether observed changes in structure result from increases in diversity and complexity when parasites are included, or instead are attributable to the unique roles that parasites play in food webs. In particular we show how the addition of parasites to food webs changes most aspects of local to system-level structure in ways primarily attributable to the generic effects of increases in diversity and complexity, regardless of the identity or type of species and links being added. However, our analyses identify two ways in which parasites do appear to play unique topological roles in food webs. First, in their roles as resources, they have close physical intimacy with their hosts, and thus are concomitant resources for the same predators. Second, in their roles as consumers, they can have complex life cycles and inverse consumer–resource body-size ratios, different from many free-living consumers. These unique roles of parasites in food webs resulted in alteration of the frequency of motifs in the case of their roles as resources, and differences in the breadth and contiguity of trophic niches between parasites and free-living species in the case of their roles as consumers. These findings can be added to one other rigorously identified unique effect of parasites—their impact on robustness. Several studies have reported that the addition of parasites reduces food web robustness to species loss [32]–[35]. One study found that reductions in robustness associated with parasite additions are not explained by species richness and connectance, known to affect robustness [17],[61], but are explained by parasites' complex life cycles [34]. That study and the current study highlight the importance of disentangling the generic structural effects of adding species and links to food webs from the unique effects attributable to the characteristics of parasites, or any other type of species being investigated. Diversity and Complexity Our analyses corroborate previous findings for how parasites alter diversity and complexity of food webs [45]. As occurs with the addition of any species to food webs, adding parasites to the trophic networks studied here increased the number of species (S) and links (L), and also usually increased link density (L/S). Increases in links and link density were especially dramatic with the inclusion of concomitant links, the numerous links from predators to the parasites of their prey. Adding parasites also increased connectance (C) in most of the food webs analyzed here, especially when concomitant links were included or when connectance was adjusted to account for the non-inclusion of those links [45]. However, our study offers clarification of a prior finding that parasites “dominate” food web links, based on a comparison of classic parasitism links to classic predation links in an earlier version of the Carpinteria Salt Marsh web [45]. For the current seven webs, classic predation links outnumbered classic parasitism links in most cases, including in the Carpinteria Salt Marsh web. Overall, parasites were sometimes involved in >50% of food web links, particularly as prey when concomitant links were included, but free-living taxa were always involved with >90% of links because the vast majority of parasite links included free-living species. Thus, strictly speaking (and by necessity), free-living species are involved in more food web links than are parasites. However, parasites are involved in substantial fractions of food web links, and if excluded, datasets would often account for less than 50% of the links in a given food web. It is important to note that any particular observation of the proportions of types of taxa and links, and thus the relative “dominance” of particular types of taxa or links, can be strongly influenced by the levels of taxonomic and trophic resolution [70] and sampling intensity [68],[73],[74] of the ecological networks in question. For example, in the current seven food web datasets, free-living bacteria and protozoa are either absent or highly aggregated. However, parasitic bacteriophages and protozoa are also absent. When we consider that worldwide, ∼60,000 vertebrate species may host ∼300,000 parasite species [21], undersampling likely leads to greater underestimates of parasites and their links than of free-living species. Network Structure: Generic Changes Prior studies have shown that variability in the raw values and distributions of network structure properties, as observed for food webs with and without parasites, often masks generalities in ecological network structure. Such generalities emerge only after appropriate normalization for diversity and complexity [8],[10],[53]. The MaxEnt, niche, and probabilistic niche models (Box 1) are used in this study as tools that provide normalizations that allow comparison of the structure of webs with different numbers of species and links. These models have previously performed well, revealing generalities in the structure of food webs [10]–[13],[54],[62],[64]. In this study, the models generally did a worse job describing the structure of food webs with parasites than food webs without parasites. This would seem to corroborate prior assertions that adding parasites alters food web structure in unique ways [41]–[48]. However, the webs with parasites in this study have species richness values of 109 to 185, greater than that of most webs without parasites previously studied. Each of the models used to evaluate network structure in our study has known scale dependence with diversity and complexity, such that the fit of the models decreases in relation to S, L, L/S, or C of the empirical web being analyzed [12],[62],[64]. When the current seven webs without parasites are compared to prior webs that lack parasites, significant scale dependencies of model fit are corroborated and extended: the width of the consumer distribution narrows with C and L/S; the absolute mean niche ME increases with S and L; and the fraction of links correctly predicted by the probabilistic niche model decreases with S and L (Table 3). The network structure of webs with parasites is in most cases consistent with these scale dependencies observed in webs without parasites (Figures 2 and 3). This suggests that apparent differences in several commonly studied aspects of network structure for webs with and without parasites are not attributable to special topological roles that parasites might play in food webs. Instead, they appear to result from generic changes in network structure due to the increasing diversity and complexity of food webs when parasites are added. Specifically, we found that changes in consumer and resource distributions, 14 commonly studied food web metrics, food web motifs (when concomitant links are excluded), and link probabilities are consistent with generic changes in food web structure associated with changes in diversity and complexity, regardless of species identity. Also, in prior work, relative nestedness, a measure of network structure not considered in the current analysis, was found to change very little with inclusion of parasites and classic parasitism links [45]–[47], but it increased greatly with the further inclusion of concomitant links in the Carpinteria Salt Marsh web [45]. This change may be attributable to a positive relationship of nestedness with connectance [74],[75], which increases with the addition of concomitant links. This should be investigated more explicitly with regard to scale dependence in future research. Our findings suggest that many aspects of previously identified generalities in food web structure across habitats and deep time [10],[11],[54],[55] likely extend from free-living species food webs to those that include parasite species. This is consistent with macroecological patterns showing that parasites and free-living species play by similar rules when it comes to the relationship between body size, abundance, and trophic level [23], in addition to similarities observed in other aspects of the metabolic theory of ecology [24]. Our analyses do highlight some patterns that need clarification with more data in the future. Specifically, a possible lower bound on the fraction of links correctly predicted by the probabilistic niche model (ƒL∼0.50) at ∼1,500 links, as suggested by webs with parasites, needs to be examined for other webs without parasites, but with high numbers of links. Also, the rate of decrease in the width of consumer distributions with increasing connectance needs to be clarified with additional data for webs with C>0.1. In general, because the scale dependencies based on webs without parasites reflect ranges of species richness and numbers of links lower than those for webs with parasites, additional data for more diverse webs without parasites, as well as highly resolved webs with parasites from other habitats, will allow more rigorous assessment of the scale dependence of model fit and whether webs with parasites are as consistent with those trends as initially indicated by this study. This brings us to another important point—our analyses reveal limitations of current simple models of food web structure. The majority of webs used to evaluate network structure thus far generally have trophic species richness less than 100. The simple models used here and elsewhere appear to fit the structure of food webs with S 1) than the distribution predicted by the model at the 95% confidence interval. A distribution is considered well fit by a model when both criteria are met: ƒG≤0.95 and −1≤W 95≤1. We calculated link density (L/S) and directed connectance (C = L/S 2) for each web, as well as adjusted connectance (C adj = L/F•S) (Table 1, Metrics 3–5) for webs with parasites but no concomitant links, to account for exclusion of such links in those web versions [45]. We calculated 14 network structure properties [12],[55] for each web (Table 1, Metrics 6–19): the fractions of top, intermediate, and basal species (Top, Int, Bas); the fractions of cannibals, herbivores, omnivores, and species in loops (Can, Herb, Omn, Loop); the standard deviations of normalized total links, generality, and vulnerability (LinkSD, GenSD, and VulSD); the mean short-weighted trophic level of all species (TL); the mean maximum trophic similarity of species (MaxSim); the mean shortest number of links between species pairs (Path); and the mean clustering coefficient (Clus). We generated 1,000 niche model webs with the same S and C as the 21 webs, and for each property for each web, calculated ME, the normalized difference between the model's median value and the empirical value [12]. ME>|1| indicates that the empirical property falls outside the most likely 95% of model values, with negative and positive MEs indicating model underestimation and overestimation of the empirical value, respectively. We investigated over- and underrepresentation of the 13 unique motifs (Box 1) that can occur among three species [11]. Motifs S1 to S5 include only single-directional links between taxa pairs, while motifs D1 to D8 include bidirectional links (i.e., mutual predation) between at least one species pair. The frequency of a motif in an empirical food web was compared to the same in an ensemble of randomized webs, yielding a z-score for each motif i that measures the degree that the empirical web deviates from the null hypothesis. We used two randomizations: “standard,” in which all links are shuffled, with the restriction that single-directional and bidirectional links are only shuffled with each other [11], and “compartmented,” which proceeds in the same fashion but with the additional restriction that links are shuffled only with those of the same type (links between free-living taxa, between parasites and free-living hosts, etc.). For a given web, we quantified the motif structure with a vector of z-scores Z = {zi }, which has one component for each of the 13 three-species motifs. To compare webs, we plotted the normalized profile, the vector of z-scores normalized to length 1. This aids in graphical comparison because larger and more densely connected webs tend to exhibit more pronounced patterns of motif representation. The occurrence of motifs in empirical webs was compared to niche model expectations. We used a probabilistic niche model (Box 1) [63],[64] based on maximum likelihood methods [16] to parameterize the niche model directly against each empirical food web. The probabilistic niche model tests the overall model fit to the data rather than to partial aspects of structure. It produces a MLE of the niche model parameters for each species i in a given web: its niche position ni , position of feeding range ci , and feeding range (or “trophic niche”) ri . This allows computation of the probability of each link in a web according to the model, and the overall expected fraction of links (f L) in a web predicted correctly by the model (Table 1, Metric 22). The one-dimensional probabilistic niche model outperforms [64] other recently proposed structural models [15],[16]. We calculated f L for one- and two-dimensional versions of the model and compared their performance for each web using the Akaike Information Criterion [93]. The MLE parameter sets were used to explore the trophic niche structure of parasite and free-living species. Supporting Information Figure S1 Cumulative resource distributions. The cumulative degree distributions for links to resources are presented in log-linear format. The link data are normalized (divided) by the mean number of links per species (L/S) in each web. The seven food webs are Bahia Falsa (Fals), Carpinteria Salt Marsh (Carp), Estero de Punta Banda (Punt), Flensburg Fjord (Flens), Otago Harbor (Otag), Sylt Tidal Basin (Sylt), and Ythan Estuary (Ythan). (TIF) Click here for additional data file. Figure S2 Cumulative consumer distributions. The cumulative degree distributions for links to consumers are presented in log-linear format. The link data are normalized (divided) by the mean number of links per species (L/S) in each web. See Figure S1 legend for food web names. (TIF) Click here for additional data file. Figure S3 Motif analysis using compartmented randomization. The representation of three-node motifs in three versions each of seven food webs. (A) Results for webs with free-living taxa only. (B) Results for webs with parasites but not concomitant predation links. (C) Results for webs with parasites and concomitant predation links. Motif labels and graphics are shown at the top of the figure, with arrowheads pointing from resources to consumers. The data points show the normalized profile overrepresentation (>0) or underrepresentation ( 0) or underrepresentation ( 1) than the distribution predicted by the model at the 95% confidence interval. Bold indicates ƒG or W 95 values that differ significantly from model expectations. (DOCX) Click here for additional data file. Table S5 Basic properties of 28 previously studied food webs used for scale dependence analyses. S, L, L/S, and C are defined in Table 1 (Metrics 1–4). An “x” indicates the subset of ten webs utilized in analyses of scale dependence of absolute niche ME (|ME|) [12]. All 28 webs were used in assessments of relative width of the consumer distribution (W 95 Cons) and fraction of links correctly predicted by the probabilistic niche model (f L). The 28 webs represent a subset of overlapping webs from [62],[64], with the following webs eliminated: webs with S |1| are shown in bold and indicate a poor fit of the niche model prediction to the empirical value. Negative MEs indicate niche model underestimation of the empirical value; positive MEs indicate niche model overestimation of the empirical value. (DOCX) Click here for additional data file. Table S7 Niche model errors for web structure properties. See Table S1 for food web naming conventions. The values show the niche MEs for properties related to types of species in the web. The properties are defined in Table 1 (Metrics 13–19). Values of ME>|1| are shown in bold and indicate a poor fit of the niche model prediction to the empirical value. Negative MEs indicate niche model underestimation of the empirical value; positive MEs indicate niche model overestimation of the empirical value. (DOCX) Click here for additional data file. Table S8 Probabilistic niche model results. See Table S1 for food web naming conventions. f L-1D and f L-2D indicate the fraction of links in an empirical web predicted correctly by the one-dimensional and two-dimensional versions of the probabilistic niche model (Box 1), respectively. AIC-1D and AIC-2D give the Akaike Information Criterion values [93] for the performance of the one-dimensional and two-dimensional versions of the probabilistic niche model. (DOCX) Click here for additional data file.
              Bookmark
              • Record: found
              • Abstract: found
              • Article: not found

              Parasites alter community structure.

              Parasites often play an important role in modifying the physiology and behavior of their hosts and may, consequently, mediate the influence hosts have on other components of an ecological community. Along the northern Atlantic coast of North America, the dominant herbivorous snail Littorina littorea structures rocky intertidal communities through strong grazing pressure and is frequently parasitized by the digenean trematode Cryptocotyle lingua. We hypothesized that the effects of parasitism on host physiology would induce behavioral changes in L. littorea, which in turn would modulate L. littorea's influence on intertidal community composition. Specifically, we hypothesized that C. lingua infection would alter the grazing rate of L. littorea and, consequently, macroalgal communities would develop differently in the presence of infected versus uninfected snails. Our results show that uninfected snails consumed 40% more ephemeral macroalgal biomass than infected snails in the laboratory, probably because the digestive system of infected snails is compromised by C. lingua infection. In the field, this weaker grazing by infected snails resulted in significantly greater expansion of ephemeral macroalgal cover relative to grazing by uninfected snails. By decreasing the per-capita grazing rate of the dominant herbivore, C. lingua indirectly affects the composition of the macroalgal community and may in turn affect other species that depend on macroalgae for resources or habitat structure. In light of the abundance of parasites across systems, we suggest that, through trait-mediated indirect effects, parasites may be a common determinant of structure in ecological communities.
                Bookmark

                Author and article information

                Contributors
                Role: ND
                Role: ND
                Role: ND
                Role: ND
                Role: ND
                Role: ND
                Role: ND
                Journal
                rbpv
                Revista Brasileira de Parasitologia Veterinária
                Rev. Bras. Parasitol. Vet.
                Colégio Brasileiro de Parasitologia Veterinária (Jaboticabal, SP, Brazil )
                0103-846X
                1984-2961
                July 2019
                : 28
                : 3
                : 376-382
                Affiliations
                Mar del Plata Buenos Aires orgnameUniversidad Nacional de Mar del Plata Argentina
                Chillán Bío-Bío orgnameUniversidad de Concepción orgdiv1Facultad de Ciencias Veterinarias orgdiv2Laboratorio de Enfermedades y Parásitos de Fauna silvestre Chile
                Missoula Montana orgnameHelm West Lab United States of America
                Concepción Bío-Bío orgnameUniversidad de Concepción orgdiv1Facultad de Ciencias Naturales y Oceanográficas Chile
                Concepción Santiago de Chile orgnameUniversidad Santo Tomás orgdiv1Escuela de Medicina Veterinaria Chile
                Article
                S1984-29612019000300376
                10.1590/s1984-29612019045

                This work is licensed under a Creative Commons Attribution 4.0 International License.

                Page count
                Figures: 0, Tables: 0, Equations: 0, References: 55, Pages: 7
                Product
                Product Information: website
                Categories
                Original Article

                Comments

                Comment on this article