Observations on the Development of Echinococcus multilocularis in Cats

1086

THE JOURNAL OF PARASITOLOGY, VOL. 89, NO. 5, OCTOBER 2003

J. Parasitol., 89(5), 2003, pp. 1086–1088 q American Society of Parasitologists 2003

Observations on the Development of Echinococcus multilocularis in Cats R. C. A. Thompson, P. Deplazes*, and J. Eckert*, WHO Collaborating Centre for the Molecular Epidemiology of Parasitic Infections and Western Australian Biomedical Research Institute, Division of Veterinary and Biomedical Sciences, Murdoch University, Murdoch, Western Australia 6150, Australia; *Institute of Parasitology, University of Zurich, Winterthurerstr 266A, CH-8057 Zurich, Switzerland. e-mail: [email protected] ABSTRACT: The development of a European isolate of Echinococcus multilocularis was compared in cats and dogs at the end of the prepatent period. Echinococcus multilocularis established in all dogs and cats, but worm recovery was significantly greater from dogs than from cats. Overall, worms in cats were not as advanced as those in dogs in terms of development and maturation, but there was no evidence of retarded development or stunted forms. These results confirm that dogs are highly susceptible to E. multilocularis, whereas cats have lower and more variable recovery rates. However, because cats produce thick-shelled eggs of E. multilocularis after experimental and natural infections, they have to be regarded as potential sources of infection both for intermediate and accidental hosts, including humans. However, their general role in the epidemiology of the infection has yet to be determined.

multilocularis based on data of experimental infections (Vogel, 1957; Zehyle and Bosch, 1982; Nonaka et al., 1996), high infection extensities (.60%) in fox populations in Europe and other endemic areas (Eckert, Gemmell et al., 2001), high infection intensities (.1,000 specimens per animal) in about 25% of the population, and massive worm burdens (up to approximately 60,000) in a few individuals (Hofer et al., 2000). Domestic dogs were also shown to be highly susceptible to E. multilocularis in experimental studies. A recent example is the infection of 2 groups of 5 dogs each (approximately 1–2 yr old) with approximately 200,000 and 80,000 protoscoleces of E. multilocularis, resulting in the establishment of intestinal worm populations in all 10 animals and average worm burdens in the 2 groups of 100,725 and 33,575, respectively (Eckert, Thompson et al., 2001). In contrast, a number of studies using isolates of the parasite from Europe, North America, and Japan have suggested that cats are less susceptible to infection with E. multilocularis than dogs, with much lower worm burdens and retarded parasite development (Crellin et al., 1981; Thompson and Eckert, 1983; Kamiya et al., 1985, 1986). Because the potential role of cats in the life cycle of the parasite and in disease transmission to humans is still unclear, we have simultaneously infected dogs and cats with a European isolate of E. multilocularis and compared the developmental status of the worm populations at the end of the prepatent period. Five cats (European Shorthair, 2 males and 3 females, 4–6 mo old) and 3 dogs (Beagle 3 Niederlaufhund, females, approximately 6–7 yr old) were used in the study (Table I). The animals originated from breeding colonies and had never been infected with cestodes or other helminths. They were experimentally infected with E. multilocularis in association with a much larger study investigating the efficacy of a cestodicidal drug (Eckert, Thompson et al., 2001). The animals were each infected with 22,600 protoscoleces of E. multilocularis. Metacestodes of the parasite were collected several years previously in southern Germany (Stuttgart isolate) from a naturally infected vole (Microtus arvalis) and subsequently maintained in laboratory colonies of the same rodent species by intraperitoneal serial passages. The metacestodes used for isolating protoscoleces were obtained from voles dissected 17 wk postinfection (PI). The same batch of protoscoleces was used for the

In central Europe, Echinococcus multilocularis, the causative agent of human alveolar echinococcosis, is predominantly perpetuated in a sylvatic life cycle with red foxes (Vulpes vulpes) as definitive hosts and various species of rodents as intermediate hosts. Domestic dogs and cats can also be involved; these acquire the infection from the sylvatic cycle by ingestion of rodents infected with the metacestode stage of the parasite (Rausch, 1995; Eckert, Gemmell et al., 2001). For example, in an endemic area of eastern Switzerland, where the average prevalence of E. multilocularis in red foxes was 33%, 0.3% of 663 dogs and 0.4% of 283 cats from the normal population were parasite carriers (Deplazes et al., 1999), but a higher local prevalence was recorded from a focus in the western part of the country, where 7% of 86 dogs were infected (Gottstein et al., 2001). Dogs and cats naturally infected with E. multilocularis have been also recorded from other central European countries, such as Germany and France (Eckert et al., 1974; Zehyle et al., 1988; Fesseler et al., 1989; Worbes, 1992; Petavy et al., 2000). A recent study in an endemic area of northern Germany has shown that 2 of 74 (2.7%) wild raccoon dogs (Nyctereutes procyonoides) had very heavy intestinal burdens of E. multilocularis (Thiess et al., 2001). With this finding the number of confirmed definitive hosts of E. multilocularis in central Europe has increased to 4, but the epidemiological significance of dogs, cats, and raccoon dogs is not yet well understood. Red foxes are regarded as highly susceptible definitive hosts of E.

TABLE I. Development of Echinococcus multilocularis in dogs and cats.

Animals Dog 1 Dog 2 Dog 3 Average Cat 1 Cat 2 Cat 3 Cat 4 Cat 5 Average

Total no. of worms

Worm recovery (%)

Total length, mean 6 SD (range) (mm)

6,330 9,750 17,110 11,163 282 6,833 1,475 20 5,765 2,864

29.3 43.1 75.7 49.4 1.2 30.2 6.5 0.09 25.3 12.7

1.57 6 0.2 (1.0–2.2) 1.95 6 0.3 (1.4–2.5) 2.13 6 0.3 (1.3–2.7) 1.88 1.4 6 0.3 (0.6–1.8) 1.74 6 0.4 (0.9–2.5) 1.47 6 0.3 (0.9–2.2) Not done 1.46 6 0.3 (0.8–2.2) 1.24

Segmentation: % worms*

Maturation: % worms†

S11

S12

S13

S14

S15

T 1 FG U 1 C U 1 TE U 1 SE

0.9 0 0

14.6 0 0

81.6 47.2 34.5

2.9 52.8 63.1

0 0 2.4

4.8 0 0

82.4 15.1 14.3

5.9 66.0 28.6

6.9 18.9 57.1

0 0 3.2

0 14.4 42.0

100 85.6 54.8

0 0 0

0 0 0

0 0 3.3

91.7 76.7 96.7

8.3 23.3 0

0 0 0

0

20.6

79.4

0

0

2.9

97.1

0

* S 1 1, scolex with 1 segment; S 1 2, scolex with 2 segments; S 1 3, scolex with 3 segments; S 1 4, scolex with 4 segments; S 1 5, scolex with 5 segments. † T 1 FG, testes containing spermatozoa, ovary, uterine streak, and other female genitalia; U 1 C, developing eggs in the uterus; U 1 TE, ‘‘thin-shelled’’ (partly developed embryophore) eggs with a fully formed oncospheres in the uterus; U 1 SE, thick-shelled eggs in uterus.

RESEARCH NOTES

infection of dogs and cats, but the former received the protoscolex suspension mixed with a small portion of canned meat for spontaneous uptake, whereas cats were orally infected with 1 ml of the suspension by means of a syringe. Dogs and cats were maintained under helminthfree conditions according to the animal welfare regulations and necropsied 25 and 26 days PI, respectively. Procedures for isolation of protoscoleces and worm recovery and processing were as described previously (Thompson and Eckert, 1983; Eckert, Thompson et al., 2001). Worm counts were performed using the dilution technique (Eckert, Thompson et al., 2001), and morphological features were determined in 50 (apart from catalog number 4) 10% formalin-fixed and stained worms per animal, as described previously (Eckert et al., 1989). As seen from Table I, E. multilocularis established in all dogs and cats, but worm recovery was significantly greater from dogs than from cats, although the worm burden of 1 cat (6,833) surpassed the lowest burden (6,630) in 1 of the dogs. Furthermore, the difference between the lowest and highest percent recovery in dogs was only 2.6 times as compared with 335 in cats, indicating a much lower variability of worm establishment in the former than in the latter. There was little difference between the growth, development, and maturation of worms from dog 1 compared with worms from cats 2, 3, and 5. However, with the remaining 2 dogs, worms were generally longer than the worms in cats. Shelled eggs were present on day 25 PI in some worms of all 3 dogs, but on day 26 PI, shelled eggs were present only in the parasites from 2 of 4 cats. Overall, worms in cats were not as advanced as those in dogs in terms of development and maturation, but there was no evidence of retarded development or stunted forms as reported in previous studies (Thompson and Eckert, 1983; Kamiya et al., 1986). Our results confirm that dogs are highly susceptible to E. multilocularis, even at higher ages of approximately 6–7 yr. In contrast, the young cats in our experiment had lower and more variable recovery rates as compared with the 3 dogs infected with the same batch and dose of protoscoleces. In another recent study, each of 10 cats at an age between 7 and 8 mo was infected with 10,000 protoscoleces (Stuttgart isolate) (Jenkins and Romig, 2000). Two of the cats did not acquire the infection, and in the remaining 8 animals the individual worm burdens were rather variable with low worm numbers (5–220) in 5 and higher burdens (815–3,045) in 3 cats, corresponding to recovery rates between 0.05–2.2 and 8.1–30.4, respectively. In a Japanese study (Kamiya et al., 1986), 6 cats (6–12 mo old) were each infected with 70,000 protoscoleces of E. multilocularis (Hokkaido isolate), but only 4 acquired the infection with individual worm burdens between 31 and 833 and recovery rates ranging from 0.04 to 1.2 at day 27 PI. In contrast, in 1 dog infected with the same isolate and dose of protoscoleces, the recovery was 40.0% (Kamiya et al., 1986). In a further Japanese study, it was observed that the recovery of E. multilocularis (Alaska isolate) from 7 experimentally infected cats suddenly decreased after day 10 PI, whereas this rate remained nearly constant in 7 dogs throughout the duration of the experiment until day 30 PI (Kamiya et al., 1985). These data and other studies (Vogel, 1957; Crellin et al., 1981; Thompson and Eckert, 1983) show that cats appear to be less susceptible to E. multilocularis than dogs, resulting in comparatively lower worm burdens in the former. However, our data indicate that in some young cats, growth and development of the worm population during the prepatent period may be very similar to the worms in dogs and that shelled fully embryonated eggs in the worms from cats may be formed within 26 days PI. This is in agreement with findings of naturally infected cats that harbored at least some egg-producing worms (Eckert et al., 1974; Worbes, 1992; Petavy et al., 2000) and observations on fecal egg excretion in experimentally infected cats (Vogel, 1957). On the other hand, there are also well-documented observations that the growth and maturation of E. multilocularis in cats may be retarded as compared with those in dogs (Thompson and Eckert, 1983; Kamiya et al., 1986) and that worm losses may occur during the prepatent period (Kamiya et al., 1985). Because cats may produce fully developed eggs of E. multilocularis after experimental and natural infections, they have to be regarded as potential sources of infection both for intermediate hosts and accidental hosts, including humans (see also Kamiya et al., 1986). However, their general role in the epidemiology of the infection has not yet been studied in detail. This would require simultaneous experimental infections of dogs, cats, and wild definitive hosts with comparative studies on egg excretion and infectivity of the eggs.

1087

LITERATURE CITED CRELLIN, J. R., A. A. MARCHIONDO, AND F. L. ANDERSEN. 1981. Comparison of suitability of dogs and cats as hosts of Echinococcus multilocularis. American Journal of Veterinary Research 42: 1980– 1981. DEPLAZES, P., P. ALTHER, I. TANNER, R. C. A. THOMPSON, AND J. ECKERT. 1999. Echinococcus multilocularis coproantigen detection by enzyme-linked immunosorbent assay in fox, dog, and cat populations. Journal of Parasitology 85: 115–121. ECKERT, J., M. A. GEMMELL, F.-X. MESLIN, AND Z. S. PAWLOWSKI. 2001. WHO/OIE manual on Echinococcosis in humans and animals: A public health problem of global concern. World Health Organisation for Animal Health, Paris, France, 265 p. ———, B. MU¨LLER, AND A. J. PARTRIDGE. 1974. The domestic cat and dog as natural definitive hosts of Echinococcus (Alveococcus) multilocularis in Southern Federal Republic of Germany. Tropenmedizin und Parasitologie 25: 334–337. ———, R. C. A. THOMPSON, H. BUCKLAR, B. BILGER, AND P. DEPLAZES. 2001. Pru¨fung der Wirkung von Episprantel (Cestext) gegen Echinococcus multilocularis bei Hunden und Katzen. Berliner und Mu¨nchener Tiera¨rztliche Wochenschrift 114: 121–126. ———, ———, S. A. MICHAEL, L. M. KUMARATILAKE, AND H. M. ELSAWAH. 1989. Echinococcus granulosus of camel origin: Development in dogs and parasite morphology. Parasitology Research 75: 536–544. FESSELER, M., E. SCHOTT, AND B. MU¨LLER. 1989. Zum Vorkommen von Echinococcus multilocularis bei der Katze. Untersuchungen im Regierungsbezirk Tuebingen. Tiera¨rztliche Umschau 44: 766–775. GOTTSTEIN, B., F. SAUCY, P. DEPLAZES, J. REICHEN, G. DEMIERRE, A. BUSATO, CH. ZUERCHER, AND P. PUGIN. 2001. Is high prevalence of Echinococcus multilocularis in wild and domestic animals associated with disease incidence in humans? Emerging Infectious Diseases 7: 408–412. HOFER, S., S. GLOOR, U. MU¨LLER, A. MATHIS, D. HEGGLIN, AND P. DEPLAZES. 2000. High prevalence of Echinococcus multilocularis in urban red foxes (Vulpes vulpes) and voles (Arvicola terrestris) in the city of Zurich, Switzerland. Parasitology 120: 135–142. JENKINS, D. J., AND T. ROMIG. 2000. Efficacy of Droncitt Spot-on (praziquantel) 4% w/v against immature and mature Echinococcus multilocularis in cats. International Journal for Parasitology 30: 959– 962. KAMIYA, M., H.-K. OOI, AND M. OHBAYASHI. 1986. Susceptibility of cats to the Hokkaido isolate of Echinococcus multilocularis. Japanese Journal of Veterinary Research 48: 763–767. ———, ———, Y. OKU, K. YAGI, AND M. OHBAYASHI. 1985. Growth and development of Echinococcus multilocularis in experimentally infected cats. Japanese Journal of Veterinary Research 33: 135– 140. NONAKA, N., M. IIDA, K. YAGI, T. ITO, H.-K. OOI, Y. OKU, AND M. KAMIYA. 1996. Time course of coproantigen excretion in Echinococcus multilocularis infections in foxes and an alternative definitive host, golden hamsters. International Journal for Parasitology 26: 1271–1278. PETAVY, A. F., F. TENORA, S. DEBLOCK, AND V. SERGENT. 2000. Echinococcus multilocularis in domestic cats in France: A potential risk factor for alveolar hydatid disease contamination in humans. Veterinary Parasitology 87: 151–156. RAUSCH, R. L. 1995. Life cycle patterns and geographic distribution of Echinococcus species. In Echinococcus and hydatid disease, R. C. A. Thompson and A. J. Lymbery (eds.). CAB International, Wallingford, Oxon, U.K., p. 88–134. THIESS, A., R. SCHUSTER, K. NO¨CKLER, AND H. MIX. 2001. Helminthenfunde beim einheimischen Marderhund Nyctereutes procyonoides (Gray, 1834). Berliner und Mu¨nchener Tiera¨rztliche Wochenschrift 114: 273–276. THOMPSON, R. C. A., AND J. ECKERT. 1983. Observations on Echinococcus multilocularis in the definitive host. Zeitschrift fu¨r Parasitenkunde 69: 335–345. ¨ ber den Echinococcus multilocularis Su¨ddeutschVOGEL, H. 1957. U lands. Zeitschrift fu¨r Tropenmedizin und Parasitologie 8: 404–454. WORBES, H. 1992. Echinococcus granulosus and Echinococcus mul-

1088


tilocularis in Thu¨ringen. Angewandte Parasitologie 33: 193– 204. ZEHYLE, E., M. ABEL, AND W. FRANK. 1988. Untersuchungen zum Vorkommen von Echinococcus multilocularis bei End- und Zwischenwirten in der bundesrepublik Deutschland (1974–1985). 13 Tagung Deutsche Gesellschaft fur Parasitologie; 22–25 March 1988; Neuchate, Germany.

———, AND D. BOSCH. 1982. Comparative experimental infections of cats and foxes with Echinococcus multilocularis. Zentralblatt fur Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene Abteilung 1, Referate 277: 117–118.

J. Parasitol., 89(5), 2003, pp. 1088–1090 q American Society of Parasitologists 2003

The Influence of Habitat on the Distribution and Abundance of Metacercariae of Macravestibulum obtusicaudum (Pronocephalidae) in a Small Indiana Stream Eric J. Wetzel and Eric W. Shreve, Department of Biology, Wabash College, 301 West Wabash Avenue, Crawfordsville, Indiana 47933. e-mail: [email protected] ABSTRACT: Snails (Elimia livescens) from a central Indiana stream were examined for encysted metacercariae of Macravestibulum obtusicaudum (Pronocephalide) to determine the distribution and abundance of this parasite on its second intermediate host. Five samples of snails were collected, with 2 samples being restricted to high-flow (riffle) or low-flow (pool) areas of the stream. Snails (n 5 386) were measured for shell length; the shell and the inner and outer surfaces of the operculum (in most samples) were examined for metacercariae. Seventyfive percent of snails (overall) had encysted metacercariae (range, 52– 97%), primarily on the opercula. A significantly lower proportion of snails from riffles were infected, and these snails had significantly fewer cysts as well. Snails collected from pools showed up to 10 times as many metacercariae than those from riffles, although there was no significant difference in the proportion of snails with intramolluscan infections of M. obtusicaudum. The inner opercular surface appeared to be the preferred site of encystment in both flow regimes. Differences in microhabitats, in terms of both snails in the stream and metacercariae on the snails, clearly must be considered when evaluating the infection patterns of this trematode.

Macravestibulum obtusicaudum is a digenetic trematode infecting freshwater turtles (Mackin, 1930). Elimia (5Goniobasis) livescens (Pleuroceridae) are infected as first intermediate hosts, which then shed cercariae (Horsfall, 1930). Motile cercariae emerge and encyst upon these snails (though not necessarily the same individuals) and form metacercariae (5cysts). The life cycle is completed when turtles (the definitive host) ingest snails bearing cysts (Horsfall, 1935). Little is known about the infection dynamics of M. obtusicaudum metacercariae on their second intermediate hosts (snails). Horsfall (1935), also working with E. livescens, noted that ‘‘cercariae encysted upon the outside of the operculum of any convenient snail or on any hard surface but not on vegetation,’’ but neither reported any additional observations on site-specific encystment on snails nor any data on patterns of infection in wild snail populations. Hsu (1937) studied the life history of M. eversum in the same snail species in Michigan and reported that cercariae would encyst on the inner surface of the operculum (with no mention of the outer surface) and rarely on the shell; however, no population-level information was given on the infection dynamics, and the maximum number of metacercariae observed on a snail was only 18. Because the distribution and abundance of metacercariae in the field can potentially be an important determinant of infection patterns in the definitive host, e.g., Bush et al. (1993), and because so little ecological work has been done with the metacercariae of this trematode (on snails), we examined the pattern of infection of metacercariae of M. obtusicaudum on E. livescens taken from a central Indiana stream. More specifically, we addressed the following questions. (1) What are the patterns of distribution and abundance of M. obtusicaudum metacercariae in this E. livescens population? (2) Are these snails used as second intermediate hosts simply those which are infected with intramolluscan stages, i.e., shedding cercariae? (3) Do metacercariae exhibit

microhabitat preferences (on snails) that are related to the snail’s habitat, e.g., riffles versus pools? Snails (E. livescens) were collected by hand or with a kicknet from Little Sugar Creek, upstream of Bridge No. 62 on 550 E, east of Crawfordsville, Indiana (408019580N, 868489030W). At the sites of collection, this is a third-order stream characterized by dominant substrata of cobble, gravel, and sand, with generally well-developed riffles and pools. Unless otherwise specified, snails were collected from rocks and the benthos in riffle areas as well as in more sandy-bottomed, slower-flowing regions (pools) of the stream; samples were taken from water #0.75 m in depth. Except in the case of samples 4 and 5 (see below), snails were collected haphazardly. Samples 1 (collected in February and March 2002; n 5 126) and 2 (collected in August 2002; n 5 100) were returned to the laboratory at Wabash College, held for up to 3 days in individual plastic jars filled with stream water, measured for shell length (apex to tip of aperture, in mm), and examined for metacercariae of M. obtusicaudum. A subsample (n 5 50) of snails from sample 2 was crushed to investigate whether snails with encysted metacercariae were only those with intramolluscan infections of this fluke. The number of metacercariae (or ‘‘cysts’’) on the outside of the snail’s operculum and shell was counted. The number of metacercariae on the inside of the operculum was determined by counting the cysts as viewed through the operculum. Sample 3 (collected in September 2002; n 5 100) was returned to the laboratory, and snails were isolated and maintained as above. However, within 2 hr of collection, the operculum of each snail was removed and placed with stream water into an individual well in a 32-well culture plate that was matched to the container number of the snail from which it was removed. Although no snail in this sample was crushed, each snail was measured (in mm) for shell length. Each matching operculum was then examined for the number of metacercariae on its inner (where the snail’s foot attached) and outer surfaces. Samples 4 and 5 (each n 5 30) were collected on the same day in September 2002. Sample 4 was collected from the middle of a fastflowing riffle, whereas sample 5 was collected in the middle of a long run (5pool) with a primarily sandy substratum and a much lower flow rate (E. Wetzel, pers. obs.; no datum on flow rates was collected). These samples were treated as those in sample 3, except that all snails (n 5 60) were crushed and examined for intramolluscan infection. Ecological terminology follows the recommendations of Bush et al. (1997). Differences among prevalences were tested by chi-square test using 2 3 2 contingency tables. Differences in mean abundance among samples were tested using analysis of variance or t-tests, with Tukey– Kramer’s honestly significant difference post-hoc comparisons used with the former. Differences were considered significant when P , 0.05. All statistical tests were run using tha JMP statistical software (version 4.0.2, SAS Institute, Inc., Cary, North Carolina). A total of 386 snails from 5 different samples were examined for infection with M. obtusicaudum. Snails collected in February–March (sample 1) were significantly smaller than all other samples (which were

RESEARCH NOTES

1089

TABLE I. Infection statistics for metacercariae of Macravestibulum obtusicaudum on 5 collections of Elimia livescens. All values with error estimates represent mean abundance 6 standard error. IN and OUT refer to the location of cysts on the inner and outer surfaces of a snail’s operculum, respectively. TOTAL represents all cysts found on the operculum, and SHELL represents only those metacercariae found on the snail’s shell. For each row, different letters (a, b, and c) represent values that are significantly different (at least P , 0.5) as per analysis of variance and post-hoc comparisons. Values in parentheses represent ranges. 1 n Snail size (mm) Prevalence (%) Operculum cysts IN Operculum cysts OUT Operculum cysts TOTAL Cysts SHELL Variance–mean ratio

2

126 10.9 6 0.2 a 52 a 0.7 6 0.5 a (0–7) 0.5 6 0.2 a (0–5) 1.2 6 0.6 a (0–8) No data 2.5

100 13.9 6 0.2 b 87 b 6.7 6 0.6 b (0–34) 1.2 6 0.2 a (0–12) 7.9 6 0.7 b (0–42) 2.3 6 0.5 (0–42) 11.7

collected in late summer), among which there was no difference in snail size (Table I). Prevalence of infection with metacercariae of M. obtusicaudum ranged from 52 to 97%. Prevalences in samples 1 and 4 (52 and 53%, respectively) were significantly lower than in the other samples in which prevalences ranged from 87 to 97% (Table I). Samples 1 and 4 were also similar in the mean abundance of metacercariae found on the inside of opercula of their snail hosts, as well as in the total number of metacercariae found on opercula (inside and out) (Table I). By comparison, snails from samples 2, 3, and 5 had approximately 10 times as many (total) metacercariae on their opercula (Table I). There were significantly more metacercariae (9.3 6 0.6 and 9.6 6 1.0) on the inner surface of the operculum in snails from samples 3 and 5, respectively, when compared with snails from other samples (Table I). In contrast, snails from different samples did not differ in the abundance of metacercariae on the outside of the operculum. Metacercariae were not evenly distributed among snails in the various samples because variance–mean (for the total number of cysts) ratios ranged from 1.9 in sample 4 to 11.7 in sample 2 (Table I), indicating an aggregated dispersion pattern. Snails bearing metacercariae (13.3 6 0.1 mm) were significantly larger than uninfected snails (11.0 6 0.2 mm; t 5 5.3, P , 0.0001). Although there existed in sample 1 a significant linear relationship between snail size (shell length) and abundance of metacercariae on opercula (P 5 0.005), there was no linear relationship between snail size and number of cysts for any of the other samples. When snails from sample 1 were excluded from the analysis, there was no significant difference between infected and uninfected snails (t 5 21.1, P 5 0.259). When snails from sample 4 (riffle) were compared with snails from sample 5 (pool), those snails from the fast-flowing riffle had significantly lower numbers of metacercariae on their opercula than those from the slow-flow pool (Table I). In terms of the proportion (%) of snails TABLE II. Microhabitat differences for metacercariae of Macravestibulum obtusicaudum collected from Elimia livescens in samples 4 (riffle) and 5 (pool). IN and OUT refer to the location of cysts on the inner and outer surfaces of a snail’s operculum, respectively. TOTAL represents all cysts found on the operculum, and SHELL represents only those metacercariae found on the snail’s shell. Except for the proportion of snails with intramolluscan infections, all comparisons between samples 4 and 5 were statistically different; P-values for the corresponding chi-square analyses are shown. 4 Snails Snails Snails Snails Snails Snails

with cysts IN (%) with cysts OUT (%) with cysts TOTAL (%) with cysts SHELL (%) with intramolluscan infection (%) that lost operculum (%)

47 20 53 3 10 0

5 87 57 97 30 13 10

(P (P (P (P (P (P

, , , , . ,

0.005) 0.025) 0.001) 0.025) 0.5) 0.025)

3

4 (riffle)

5 (pool)

100 14.1 6 0.2 b 92 b 9.3 6 0.6 c (0–30) 1.2 6 0.2 a (0–14) 10.4 6 0.7 b (0–32) No data 6.2

30 14.4 6 0.4 b 53 a 0.7 6 1.0 a (0–5) 0.2 6 0.3 a (0–2) 1.0 6 0.2 a (0–5) 0.07 6 1.0 (0–2) 1.9

30 13.3 6 0.4 b 97 b 9.6 6 1.0 b,c (0–31) 1.2 6 0.3 a (0–6) 10.9 6 1.21 b (0–33) 1.0 6 1.0 (0–14) 4.8

bearing metacercariae on their opercula (inside and out), 53% of snails from the riffle (sample 4) bore metacercariae, whereas 97% of snails from the pool (sample 5) carried metacercariae (x2 5 12.8, P , 0.001). Likewise, significantly lower proportions of snails from the riffle (vs. those from the pool) had metacercariae on the inside of the operculum (47 vs. 87%; P , 0.005), on the outside of the operculum (20 vs. 57%; P , 0.025), or on the shell (3 vs. 30%; P , 0.025) (Table II). There was no significant difference (P . 0.5) in the prevalence of intramolluscan infection (with M. obtusicaudum) in snails from the riffle (10% prevalence) versus the pool (13% prevalence) (Table II). Ten percent (10%) of the snails from the pool (sample 5) had lost their operculum at some point in time (based on relative size), whereas no snail from sample 4 (riffle) was found to have lost its operculum. Based on the distribution and abundance of their resultant metacercariae, cercariae of M. obtusicaudum clearly have microhabitat preferences for encystment sites on E. livescens. Moreover, there appears to be an important effect of snail habitat preference (fast- vs. slow-flow areas of the stream) on the overall pattern of infection with M. obtusicaudum metacercariae. Elimia livescens collected early in the year (February–March) were smaller in size (on average) than those collected in late summer (August–September); the presence of (presumably younger) snails on the lower end of the size distribution is likely the reason for the significant, positive linear relationship between snail size and abundance of metacercariae in this early sample. If snails ,9 mm are removed from the analysis, no linear relationship persists. Moreover, snails from samples 1 and 4 were similar in overall prevalence of infection, mean abundance of opercular cysts, and mean abundance of cysts on the inside of the opercula. These snails were similar despite snails from sample 4 being significantly larger (and thus likely older). Even when (for the sake of analysis) the size range of sample 1 was truncated to bring the average snail size into the range of the other samples (to approximately 13.5 mm), the mean total abundance of opercular cysts remained 1.5 6 0.3. These observations suggest that despite the haphazard collection of snails in sample 1, it is likely that many of these snails were collected from faster-flowing regions of the stream. Likewise, infection similarities between snails in samples 3 and 5 suggest that the majority of snails in sample 3 may have been sampled from the pool habitat. Along with the conclusions that can be drawn from samples 4 and 5 (discussed below), it is clear that the selection of sampling sites within a stream can have a potentially significant impact on the assessment of infection levels in this system. This observation is consistent with the existence of foci of infection as seen in other host–parasite systems (Zelmer et al., 1999). Moreover, although obviously not the only potential reason, it is possible that this is why Horsfall (1930) only saw prevalences of 1.2% in her work with M. obtusicaudum versus the 80–90% prevalence reported by Hsu (1937) in her work on M. eversum. Infection foci might also explain the wide range of prevalences (52–97% in the same stream) in this study. Because opercula were not immediately removed from snails in samples 1 and 2, it is quite possible that infection data collected from these hosts underestimate the abundance of metacercariae encysted on the inner side of the operculum. For instance, significantly fewer cysts were counted on the inner opercular surface from sample 2 snails (opercula

1090


attached) when compared with snails in sample 3, from which the opercula had been removed; however, there was no difference between these 2 groups in the abundance of metacercariae found on the outer opercular surface. Originally, our thinking was that the transmission of a swimming cercaria from a patent infection to a second snail that would act as the second intermediate host, i.e., one on which cercariae would encyst to form metacercariae, would be extremely unlikely given the flow dynamics in a stream. Thus, 1 hypothesis we sought to test was that snails with metacercariae were only those infected with rediae, i.e., those snails that would themselves shed cercariae. Clearly, our data do not support this notion. Most snails bearing metacercariae were not infected with intramolluscan stages of the life cycle. On the basis of this observation, we then tested (with samples 4 and 5) the hypothesis that snails from slow-flow regions of the stream (pools) would have greater levels of infection with metacercariae of M. obtusicaudum than snails from sections of the stream with faster-flowing water, reasoning that it would be easier for cercariae to locate and encyst upon a snail in a pool versus in a riffle. Whereas there was no difference in size of snails from these 2 samples, snails from sample 5 (pools) had significantly greater abundances of metacercariae on the inner opercular surface as well as on the operculum as a whole and a significantly greater prevalence of infection than those snails from sample 4 (riffles). Although sample sizes in this test were relatively small (n 5 30 in each sample), there was no difference in the prevalence of intramolluscan infection between snails from pools and snails from riffles. Thus, despite no significant difference in the proportion of snails shedding cercariae, which would then be available to encyst as metacercariae, there were quite dramatic differences in the patterns and intensities of infection on E. livescens as second intermediate hosts for this parasite. It seems clear that the probability of infection for a snail in a pool is much greater than that for a snail in a riffle. Whether snails regularly move between these habitats remains to be seen, although it seems unlikely based on our infection data, i.e., snail movement clearly did not homogenize infection intensities. Interestingly, snails collected from the pool were observed to have lost opercula, in contrast to those collected from riffles. When comparing samples 4 and 5, no snail from sample 4 (riffle) lost its operculum, whereas 10% of the snails from sample 5 had lost their opercula (along with 17% of the snails from sample 2). Based on the data available in this study, it is clear that the operculum of E. livescens is the preferred microhabitat for encystment by cercariae. Some limited laboratory observations of cercaria behavior (data not shown) also support this hypothesis. In experiments in which newly emerged cercariae were introduced to containers with a lone operculum that had been removed from a snail, the cercariae preferentially encysted upon the operculum. Moreover, the inner opercular surface appears to be preferable to the outer surface because metacercariae were typically seen (at least in large numbers) on the outer surface of the operculum only after the inside had

DATE

OF

‘‘filled up.’’ Although Horsfall (1935) observed cercariae of M. obtusicaudum encysting upon ‘‘the outside of the operculum’’ of snails, she did not report metacercariae from the inner surface of the operculum. In contrast, Hsu (1937) reported that cercariae of M. eversum encysted on the inside (but not the outside) of the snail’s operculum. In light of this study, the differences noted in these 2 studies are not likely speciesspecific but rather habitat-specific variability that is effected by the collection of snail hosts from either riffles or pools. Up to 17% of the snails in any given sample were found to have lost their opercula. It is impossible to know why opercula were discharged, but it seems possible that it may be due to infection with high numbers of metacercariae. Although only a qualitative observation, it was clear to us that opercula bearing large numbers of metacercariae were much more easily removed from snails than those with few, or no, metacercariae. Opercula are important in many snails as a morphological defense against predators such as crayfish (Appleton and Palmer, 1988). Although speculative, it is possible that infection of E. livescens with metacercariae of M. obtusicaudum could have a significant impact on the antipredator defenses of this intermediate host. At the least, loss of heavily infested opercula from snails could serve to limit the intensity of infections transferred by way of predation to turtle definitive hosts. LITERATURE CITED APPLETON, R. D., AND A. R. PALMER. 1988. Water-borne stimuli released by predatory crabs and damaged prey induce more predator-resistant shells in a marine gastropod. Proceedings of the National Academy of Sciences of the United States of America 85: 4387– 4391. BUSH, A. O., R. W. HEARD JR., and R. M. Overstreet. 1993. Intermediate hosts as source communities. Canadian Journal of Zoology 71: 1358–1363. ———, K. D. LAFFERTY, J. M. LOTZ, AND A. W. SHOSTAK. 1997. Parasitology meets ecology on its own terms: Margolis et al. revisited. Journal of Parasitology 83: 575–583. HORSFALL, M. W. 1930. Studies on the structure of Cercaria infracaudata n. sp. Journal of Parasitology 17: 43–48. ———. 1935. Observations on the life history of Macravestibulum obtusicaudum Mackin, 1930 (Trematoda: Pronocephalidae). Proceedings of the Helminthological Society of Washington 2: 78–79. HSU, D. Y. M. 1937. The life history and morphology of Macravestibulum eversum sp. nov. (Pronocephalidae, Trematoda). Transactions of the American Microscopical Society 56: 478–504. MACKIN, J. G. 1930. A new pronocephalid monostome from a freshwater turtle. Journal of Parasitology 17: 25–29. ZELMER, D. A., E. J. WETZEL, AND G. W. ESCH. 1999. The role of habitat in structuring Halipegus occidualis metapopulations in the green frog. Journal of Parasitology 85: 19–24.

PUBLICATION

Volume 89, No. 5, was mailed 27 October 2003