ENDOPARASITE INFRACOMMUNITIES OF Hemisorubim platyrhynchos (VALENCIENNES, 1840) (PISCES: PIMELODIDAE) OF THE BAÍA RIVER, UPPER PARANÁ RIVER FLOODPLAIN, BRAZIL: SPECIFIC COMPOSITION AND ECOLOGICAL ASPECTS

Parasites of 136 specimens of Hemisorubim platyrhynchos, popularly called jurupoca, were analyzed. Fourteen parasite species were recorded: four digeneans (Crocodilicola pseudostoma, one Gorgoderidae species, Sanguinicola platyrhynchi, and Sphincterodiplostomum sp.); three cestodes (Goezeella paranaensis, Spatulifer maringaensis, and Mariauxiella piscatorum); five nematodes (Contracaecum Type 1 larvae of Moravec, Kohn, & Fernandes, 1993; Contracaecum Type 2 larvae of Moravec, Kohn, & Fernandes, 1993; Cucullanus (Cucullanus) zungaro; Eustrongylides ignotus; and Goezia sp.); one acanthocephalan (Quadrigyrus machadoi), and one pentastomid (Sebekia sp.). Most of the helminth specimens were found in larval stages, confirming that H. platyrhynchos is a significant source of their transmission. No dominance was reported. With the exception of Contracaecum Type 2 (featuring random dispersion), the species had a clumped pattern of dispersion in the host sample. Evidence of competition among parasite species was not reported. Several species of parasites were correlated with host size and sex. This fact may be explained by increase in food content and possible behavioral modifications of male and female hosts. Infracommunity diversity was not related to host size or sex.


INTRODUCTION
Hemisorubim platyrhynchos (Valenciennes, 1840), popularly called jurupoca, is a pimelodid fish inhabiting large South American rivers (Burguess, 1989).According to Ringuelet et al. (1967), this species can be found in Brazil in the Amazon, Parnaíba, and Paraná rivers and their tributaries.It is considered an important species in the fisheries of the Paraná River basin (Agostinho et al., 1995).In spite of its importance in the region under analysis, the parasite fauna and parasite ecology of this host are still unknown.This paper aims to provide information on the endoparasite infracommunities of H. platyrhynchos of the Baía River (Upper Paraná River floodplain) and deals with aspects that may influence infracommunity structures in this locality.

MATERIAL AND METHODS
One hundred and thirty-six specimens of Hemisorubim platyrhynchos were caught between August 1998 and July 1999 in the Baía River (22º42'-22º44'S and 53º17'-53º20'W), State of Mato Grosso do Sul, Brazil.This river lies within a wide flood region in the upper section of the Paraná River (Maack, 1981).Fish captures were undertaken using gill nets.The total weight, standard length, and sex of each fish were registered.Internal organs and the visceral cavity were analyzed under a stereomicroscope.The collected parasites were treated according to Eiras et al. (2000) and identified based on Travassos et al. (1969), Yamaguti (1971), Pavanelli & Rego (1989), Thatcher (1993), Moravec (1998), andChambrier &Vaucher (1999).
Data analyses were made using the following ecological indexes and statistical tests: Simpson's index (C) to determine dominance trends among parasite species, dominance being accepted when C > 0.25 (Stone & Pence, 1978); the Dispersion index, tested using "d" statistics in accordance with Ludwig & Reynolds (1988), to verify the dispersion pattern of parasite species in sample hosts; Chi-square test to determine the occurrence of associations among species and Spearman's rank correlation (rs) to verify correlations between abundance of the species that formed associated (Ludwig & Reynolds, 1988); Shannon's index (H') to determine infracommunity diversity, computed using PC-ORD software (MacCune & Mefford, 1997); Pearson's linear correlation (r), with previous angular transformation of prevalence (arc sen χ ), calculated to determine the correlation between host standard length and parasite prevalence (Zar, 1996); Spearman's rank correlation (rs) to verify the correlation between standard length and parasite abundance and diversity (Zar, 1996); the "G" loglikelihood test (using the 2x2 contingency table) and Mann-Whitney's test (U), used to determine the effects of host sex in the prevalence and abundance of each species, respectively (Zar, 1996); and Student's t test, calculated to verify differences in the diversity of male and female host infracommunities (Zar, 1996).Statistical analyses were applied to parasite species with over 10% prevalence and the results considered significant when p ≤ 0.05.The ecological terms were suggested by Bush et al. (1997).
Progenetic metacercariae of C. pseudostoma were observed, which facilitated identification at the specific level.Gorgoderidae could not be identified at a lower taxon because of the small number of individuals collected.

Ecological aspects of the endoparasite infracommunity
Among the total number of helminth individuals collected, 48.3% (1638) were adults and 51.6% (1751) were larvae.C. pseudostoma, Sanguinicola platyrhynchi, G. paranaensis, Contracaecum Type 1, and Q. machadoi had the highest prevalence, mean intensity, and mean abundances (Table 1).Simpson's index did not indicate any dominance trend in the endoparasite infracommunities (C = 0.1273).Most parasite species presented a clumped pattern of dispersion in the host sample, as shown by the Dispersion index.Only Contracaecum Type 2 showed random dispersion (Table 2).
Forty-one males and 87 females were studied.It was impossible to determine the sex of eight specimens.There was a significant relationship between host sex and prevalence of G. paranaensis and M. piscatorum (G = 5.395, 0.025 > p > 0.01; and G = 8.587, 0.005 > p > 0.001, respectively), with the females being more infected.A significant relationship between M. piscatorum abundance and host sex was observed (Z = 2.78, 0.01 > p > 0.005), with the highest abundance also occurring in the females.Infracommunities in the female hosts had a mean diversity of 1.262 (0-1.968) and a mean richness of 5.2 (1-9).In the male hosts these values were 1.188 (0-1.704) and 4.7 (1-7), respectively.However, Student's t test indicated that diversity and richness are independent of host sex (t = 0.983, p = 0.3272; t = -1.7261,p = 0.0879, respectively).

Specific composition of the endoparasite fauna
Host feeding habit and diet are important in helminth acquisition.According to Dogiel (1970), the habitat from which the main bulk of food is obtained is as important as the diet in parasite fauna composition.It is a well-known fact that the main food source for fish analyzed in the present paper is of autochthonous origin.Hemisorubim platyrhynchos is a piscivorous species that feeds on small foraging fishes (Hahn et al., 1997) which, according to Eiras (1994), were intermediate or paratenic hosts in the life cycle of parasites such as the digeneans, cestodes, and nematodes observed in this paper.Thus, host feeding habits were an important factor in determining the endoparasite fauna composition in these infracommunities.This composition comprised several species with high infection levels, much like other piscivorous species from the same region, such as Pseudoplatystoma corruscans (Agassiz, 1829) and Hoplias malabaricus (Bloch, 1794) (Machado et al., 1996;Almeida, 1998).The studied endoparasite infracommunities included autogenic and allogenic species, a fact explained by the position of the host in the food web.The fish is a great predator within the aquatic environment; however, it occupies an intermediate position in the food web, being preyed on by land animals.Thus, the host is highly important to the allogenic species (mainly the larvae with high prevalences such as C. pseudostoma, Contracaecum Type 1, and Q. machadoi) in the completion of their life cycle.Leon et al. (1992) also reported the occurrence of the progenetic metacercariae of this species in Rhamdia guatemalensis, another pimelodid fish from Mexico.According to Zdzitowiecki (1997), progenetic metacercariae can mature and produce eggs in intermediate hosts.This was observed in some specimens of C. pseudostoma parasitizing jurupocas.

Ecological aspects of the endoparasite infracommunity
The high abundance of larvae observed in this study suggests a cumulative effect in helminth infection during this stage.Larvae of previous and more recent infections could be found side by side in the same host individual.Santos & Eiras (1995) have also suggested a cumulative effect in larval infection by digenean Lecithochirium furcolabiatum Jones, 1933 from Portugal.Espinosa-Huerta et al. (1996) mentions the same by digenean Posthodiplostomum minimum (MacCallum, 1921) Dubois, 1936 in fishes from Mexico.
Clumped distribution is considered typical in parasites of freshwater fishes and has also been reported in other fish species in the region (Almeida, 1998;Brasil-Sato, 1999;Machado et al., 2000).Individual differences in immunological reaction and susceptibility to infection might have caused the clumped pattern.This dispersion pattern may be increasing reproductive efficiency in some adult species (e.g., cestodes), since it enhances mating opportunities (Holmes, 1990).However, according to this author, such local aggregations may be widely dispersed throughout the intestine.In this case, additional studies on intestinal distribution are suggested.
According to Dobson (1990), clumped distribution tends to magnify the stability of the hostparasite relationship, since regulating mechanisms of parasite populations (such as host mortality, reductions in parasite fecundity, and density-dependent survival) influence a greater proportion of these populations.Only Contracaecum Type 2 showed random dispersion.Its possibly reduced opportunity to colonize the host, as well as pathological effects, must have contributed towards its random dispersion.According Dobson (1990), rare or more pathogenic species are frequently less aggregated.In addition, according to Moravec (1998), some Contracaecum larvae may have pathogenic effects.
Positive associations among cestode species G. paranaensis/S.maringaensis and M. piscatorum/ S. maringaensis, with positive abundance correlations, may demonstrate that they have the same ecological requirements, which must be abundantly available.For example, cestodes have common microhabitats and are known to absorb monosaccharides (sometimes disaccharides) (Holmes, 1990).It is thus probable that food and space are abundant, while the occurrence of competition or other interference mechanisms suggested by Holmes (1990) is less probable.These cestode species may use the same or similar intermediate hosts, which are a frequent item in the fish diet.This fact may explain why none of the cestodes dominated the infracommunities separately.The positive association and correlation shown by G. paranaensis/S.platyrhynchi must be related to infection strategies and not to the direct interactions of these species, since they have different microhabitats.This conclusion agrees with Dobson (1990), i.e., the abundance and distribution of parasites are more a direct result of different life cycle characteristics than of interactions among species (e.g., competition).
Although infracommunities can present great richness (14 species of helminths were found in the host species), they had a maximum of 9 species.Though a species of fish presents high parasite richness, if the opportunities of transmission are low for some representatives, the probability of occurring together in a single fish is very low.According to Kennedy (1990), the number of species in an infracommunity reflects the number of species in the locality, the opportunity of transmission and infection, and, thus, the probability of host infection.Differences in species richness and diversity of the infracommunities may also be the result of individual responses to parasitism by different species of helminths and transmission rates.Interspecific competition may also play an important role in parasite richness, although no evidence of competition has been obtained.
Studies that relate parasitism with host size may indicate how parasite infracommunities structure changes during the life cycle of the host.Community organization may be influenced by host age and size through changes in diet or the volume of ingested food, ontogenetic changes in immunocompetency, and modifications in the probability of contact with intermediate hosts (Esch et al., 1990).
Increase in endoparasite abundance in larger fishes, as verified in this study for the three cestode species, has been attributed by Isaac et al. (2000) to the occurrence of cumulative processes in the infection.Although accumulation of parasites may occur, other processes are also involved.Feeding habit changes or use of intermediate hosts whose ingestion is limited by fish size may cause this infection pattern.However, no records exist of changes in the feeding habit or diet of jurupocas in the analyzed size classes.Moreover, no correlation was verified between prevalence and host size in the case of M. piscatorum and S. maringaensis.Although a positive correlation was observed for G. paranaensis, some small-sized specimens were infected.Thus, the abundance increase in larger fishes might be related to an increase in ingested food volume, as observed by Zelmer & Arai (1998) in yellow perch parasitized by Crepidostomum isostomum and Raphidascaris acus (Bloch, 1770).One may conclude that infection is probably not dependent on host size but the rates in which it occurs do.Although there was an increase in abundance in larger fishes, smaller fishes were also infected.Takemoto & Pavanelli (1994), Machado et al. (1994), andTakemoto &Pavanelli (2000) have also observed cestode infection increases in larger pimelodid fishes on the Upper Paraná River floodplain.
Contracaecum Type 1 showed negative correlation between fish size and prevalence whereas Contracaecum Type 2 had a positive correlation.This fact suggests the possibility of different transmission strategies and the use of different intermediate hosts by these congeneric species.The acquired immunity might justify the lower infection rate of C. Type 1 among the larger fishes.
Changes in behavior or habits due to an increase in host size might influence infracommunity diversity (Esch et al., 1990).However, in the present paper, possible changes in the volume of food ingested by H. platyrhynchos were not responsible for changes in endoparasite diversity among infracommunities in fishes of different sizes.
The prevalence of G. paranaensis and M. piscatorum and the abundance of the latter were dependent on host sex, and female hosts were more parasitized.These results differ from those of Takemoto & Pavanelli (2000), who observed that the cestode species of Sorubim lima, also a pimelodid fish, were more numerous in male hosts.Differences in the infection of the male and female hosts in the present study may be due to their different behaviors and, consequently, to differences in exposure to infection.However, additional investigations are necessary at the biochemical level, since Zaman &Seng (1989) andFolstad &Karter (1992) cited by Poulin (1996) mentioned the influence of sexual hormones on infection.
Segregation of infections by C. Type 1 in smaller fishes and by cestodes in female and larger fishes may have contributed to the clumped pattern of these species.Zelmer & Arai (1998) reported the contribution of host size to parasite aggregation in the yellow perch from Canada.Other factors without any relationship to fish sex or size must be involved in the aggregation of species.Life cycle features of each parasite species or the differentiated immunological response among host individuals may have produced this pattern.