2012 SOUTHEASTERN NATURALIST 11(3):415–422 Prevalence of a Potentially Lethal Parasite of Wading Birds in Natural and Agricultural Wetlands in South Louisiana Margaret C. Luent1,2,*, Melissa Collins3,4, Clinton Jeske5, and Paul Leberg1 Abstract - Gambusia affinis (Western Mosquitofish) were sampled from 18 sites representing marsh, forested wetlands, and agricultural wetlands in south Louisiana to determine distribution and infection parameters of Eustrongylides ignotus, a potentially lethal nematode parasite of wading birds. (n = 400 per site). Overall, prevalence of infection was 0.3%, with significantly higher prevalence in agricultural wetlands than in marshes or swamps. Our findings are similar to work in Florida suggesting parasite prevalence is higher in disturbed wetlands, and suggest that birds foraging in crayfish ponds and rice fields may be at increased risk of exposure. Introduction Epizootics in fish-eating birds (family Ardeidae) from many parts of the United States have been attributed to the nematode Eustrongylides ignotus Jäegerskiold (Franson and Custer 1994, Roffe 1988, Weise et al. 1977). During outbreaks, affected colonies can lose high proportions of their nestlings (Roffe 1988, Spalding et al. 1993). The nematode has a three-host life cycle; fish become infected via consumption of oligochaetes having consumed E. ignotus eggs released with wading bird feces. In Florida, one of the fish with the highest prevalence of infection is Gambusia holbrooki Girard (Eastern Mosquitofish) (Coyner et al. 2002, Spalding et al. 1993). The prevalence of E. ignotus has not been assessed in Gambusia affinis (Baird and Girard) (Western Mosquitofish), a close relative, although infections are known to occur (Deaton 2009). The Western Mosquitofish is probably the most common freshwater fish in Louisiana and thus a good candidate for studying the potential distribution of the parasite in various habitats used by foraging wading birds. Mosquitofish inhabit shallow wetlands and shorelines, making them susceptible to bird predation (Coyner et al. 2001). Even for species of herons that eat larger prey, the Mosquitofi sh is useful for detecting the presence of E. ignotus in the environment. Mosquitofish are more easily sampled in large numbers than larger fish (Coyner 1University of Louisiana at Lafayette, Department of Biology, PO Box 42451, Lafayette, LA 70504. 2Current address - Louisiana Department of Wildlife and Fisheries, 102 Magnate Drive, Suite 201, Lafayette, LA 70508. 3IAP World Services, Inc., US Geological Survey, National Wetlands Research Center, Lafayette, LA 70506. 4Current address - 1056 Sandy Nell Road, Breaux Bridge, LA 70517. 5US Geological Survey, National Wetlands Research Center, Lafayette, LA, 70506. *Corresponding author - mluent@wlf.la.gov. 416 Southeastern Naturalist Vol. 11, No. 3 et al. 2002, Frederick et al. 1996). Additionally, larger piscivoruous fish may accumulate E. ignotus by consuming infected mosquitofish; predators of those larger fish might be more at risk of infection than smaller wading birds (Coyner et al. 2003). Prevalence of E. ignotus in Florida is highest in habitats impacted by disturbed soils, altered hydrology, and exogenous nutrients (Coyner et al. 2002, Spalding et al. 1993). Frederick et al. (1996) hypothesized recent increases in the distribution and prevalence of the nematode may be in response to increased human disturbance of aquatic systems. Wading bird populations increased in Louisiana between 1949–1988, during the same time period that there was an increase in crayfish aquaculture (Fleury and Sherry 1995). If wading birds are food-limited, then crayfish/rice impoundments may contribute to the abundance of wading birds in Louisiana by increasing foraging-site availability (Fleury and Sherry 1995). However, if agricultural impoundments support a large number of infected fish relative to what is observed in natural wetlands, they may serve as an ecological sink diminishing the reproductive success of foraging birds. Due to the importance of Louisiana to continental populations of wading birds (Martin and Lester 1990), we undertook a study to determine the prevalence and intensity of infection of E. ignotus in Louisiana populations of mosquitofish. We tested the null hypothesis that Western Mosquitofish from agricultural wetlands do not have a higher occurrence of E. ignotus than do natural wetlands. We also evaluated the association between parasite prevalence and fish abundance and water quality. Methods We sampled fish from 6 sites representing each of three types of wetlands (Fig. 1). All sampling sites were separated by at least 2 km. Forested sampling sites included the Sherburne Wildlife Management Area (WMA), Lake Fausse Point State Park (SP), Chicot SP, Thistlewaite WMA, Tunica Hills WMA, and Maurepas WMA. Marsh sites included Mandalay National Wildlife Refuge (NWR), Laccasine NWR, Sabine NWR, Cameron Prairie NWR, and Rockefeller State Refuge. Agricultural wetland sites were temporarily flooded impoundments used for both rice and crayfish production (sites located near Cade, Catahoula, and Crowley) and ditches that drained such impoundments (sites located near Kaplan, Maurice, and Ridge). We sampled Gambusia when fledgling concentrations were highest (June– July 2004). To estimate relative abundance of mosquitofish, we used a catchper- unit-effort index. We sampled three areas of similar and suitable habitat (less than 0.5 m deep, partially vegetated) at each site, vigorously sweeping a longhandled dip net for three 3-minute catch intervals. Counts of fish collected were averaged together to obtain a mean relative abundance for a site. This index of abundance was uniform both in terms of number of subsamples and the intensity of the sampling. 2012 M.C. Luent, M. Collins, C. Jeske, and P. Leberg 417 Our objective was to examine 400 fish per site for infection by E. ignotus. If fewer fish were caught, we supplemented those numbers with additional dip-netting in similar habitat. Fish were examined using a dissecting microscope, individual infection status was noted, and nematodes were sent to M. Kinsella, Department of Pathobiology at the University of Florida, for positive identification. Following Coyner et al. (2002), we determined total nitrogen, total phosphorous, dissolved oxygen, and chlorophyll between 6–11 AM at each site in habitat similar to that sampled for fish abundance. Dissolved oxygen was immediately measured in the field using a galvanic sensor (YSI, Inc). Samples for total nitrogen, total phosphorous, and chlorophyll were initially frozen prior to analysis at the National Wetland Research Center in Lafayette, LA, using a CE440 Elemental Analyzer (Leeman Labs, Inc.) or a 10-AU Fluorometer (Turner Designs). All results are reported as mg/L. Figure 1. Locations of six sample sites for each wetland type used to assess parasite prevalence. The number of infected fish found in a sample (n = 400/site) is located next to each site. 418 Southeastern Naturalist Vol. 11, No. 3 We used a nonparametric Kruskal-Wallis (KW) test to evaluate whether E. ignotus prevalence and abundance differed among wetland types. If a KW test was significant, we conducted a Dunn’s multiple comparison to assess differences among wetland types. The same approach was used to assess differences of fish abundance and water quality measures among wetland types. We tested the null hypothesis of no correlation between parasite prevalence and measures of water quality or fish abundance at our 18 sample sites using Pearson’s r. Finally we used Poisson regression to determine if a combination of habitat with the water quality variables or fish abundance better explained the variation in parasite prevalence than did habitat alone. Results and Discussion Of the 7200 fish examined, 25 were infected by nematodes; all were confirmed to be E. ignotus. We did not note the presence of any other parasites, but did not conduct a microscopic evaluation. No fish was infected with more than one E. ignotus. Therefore the prevalence per site provided the same information as parasite abundance per site, so only the former is analyzed. Overall prevalence of E. ignotus in southern Louisiana (0.3%) was slightly lower than levels found in two large surveys in Florida (1.1% in Frederick et al. 1996, and 0.6% in Coyner et al. 2002). However, infected mosquitofish were found at 50% of our sampled sites, which was much higher than the occurrence in Florida (14.7% and 17.1% of sites examined by Frederick et al. 1996 and Coyner et al. 2002, respectively). This difference in the proportion of sites with infected fish between studies is probably not due to differences in sampling intensity. While Frederick et al. (1996) examined fewer fish than we did per site, Coyner et al. (2002) looked at an average of 360 fish per site (which is similar to our sample of 400). The prevalence of infected fish was higher in agricultural wetlands (0.9%) than in forested wetlands or marshes (H = 10.44, P = 0.005; Fig. 2); there was no difference between the two natural wetlands. The parasite was detected in all of the agricultural wetland sites compared to only 1–2 sites within each of the natural wetland types (Fig. 1). In Florida, infection prevalence is higher in altered than in natural wetlands (Coyner et al. 2002, Frederick et al. 1996, Spalding et al. 1993). The man-made habitats in Florida associated with increased parasite prevalence and occurrence, however, were not agricultural wetlands. The reason for increased prevalence of parasitized fish in artificial wetlands is unclear, but it could be due to higher use of agricultural wetlands by wading birds, or to higher oligochaete populations, than in natural wetlands, either of which would enhance the ability of the parasite to complete its life cycle. Unfortunately, to our knowledge, relative habitat use by wading birds or relative abundance in of oligochaetes in natural and agricultural wetlands has not been assessed. No correlation was found between dissolved oxygen, phosphorus, chlorophyll, or fish abundance and nematode prevalence (Table 1); nitrogen 2012 M.C. Luent, M. Collins, C. Jeske, and P. Leberg 419 was below the detection limits across the sites. Furthermore, water quality measures and fish abundance did not differ among wetland types (Table 1). Coyner et al. (2002) observed the higher levels of total nitrogen, total Figure 2. Box plot (25 and 75 quartiles) indicating differences in parasite prevalence among wetland types. The median and mean prevalence for 6 sample sites are indicated by dark horizontal solid and dashed lines, respectively (the median corresponded with the 25th quartile for two wetland types). The vertical bars represent the range of prevalence for each wetland type. Letters above bars are based on the results of a Dunn’s multiple comparison test; prevalence did not differ among wetland types identified with the same letter. Table 1. Water quality measurements1 and fish abundance for three wetland types and Spearman’s correlation (R) coefficient between these measurements and nematode abundance (n = 12). The null hypothesis that the environmental variable did not differ among habitats was evaluated with a Kruskal Wallis (H) test. Environmentalvariable Wetland types Mean SD R P H P Dissolved oxygen (mg/L) Forested 5.59 1.80 Marsh 7.34 3.25 Agricultural 8.52 4.54 0.30 0.22 1.37 0.53 Chlorophyll (mg/L) 0.50 0.80 Forested 32.50 26.38 Marsh 36.50 35.65 Agricultural 58.57 54.98 0.13 0.62 1.14 0.58 Phosphorous (mg/L) Forested 0.12 0.09 Marsh 0.14 0.08 Agricultural 0.25 0.21 0.35 0.15 1.14 0.58 Fish relative abundance (mean number captured in three 3-mimute samples) Forested 87.57 67.15 Marsh 149.17 274.56 Agricultural 46.58 27.01 -0.155 0.53 1.22 0.56 1Nitrogen was also measured but was below detection limits at most sites and so is not presented. 420 Southeastern Naturalist Vol. 11, No. 3 phosphorous, and chlorophyll and lower levels of dissolved oxygen at altered wetland sites compared to natural wetlands in Florida were associated with E. ignotus prevalence. We found no similar correlations or significant differences in water quality among wetland types; however, subtle differences in water quality among wetland types might not have been detected with our sample size of only 18 sites. Poisson regressions to assess the effects of combinations of variables on prevalence lacked sufficient samples to model interactions between environmental variables, including fish abundance, with habitat (based on model convergence statistics). When assessing main effects on parasite prevalence, none of the environmental variables significantly improved on the explanatory power of the model containing only wetland type (χ2 2 =18.58, P = 0.001) except for the model containing both wetland type and DO (χ2 3 = 21.44, P = 0.001). The odds of detecting an additional parasite in a sample increased 1.165 (95% CI = 1.064–1.28) times for each mg increase of dissolved oxygen per liter at a site, when controlling for the effects of wetland type. For comparison, the odds of finding an additional parasite per 400 fish in a sample from agricultural wetlands were 7.79 (95% CI = 1.79–49.40) and 10.50 (95% CI = 2.46–44.79) times higher than in either marsh or forested wetlands, respectively. We found that parasite prevalence was positively associated with DO within wetland types, and was highest in agricultural wetlands, which did not appear to be eutrophic relative to natural wetlands. This finding is different than in Florida where high parasite prevalence was associated with low DO levels found in eutrophic artificial wetlands (Coyner et al. 2002). This difference suggests that the mechanism responsible for increased parasite prevalence in artificial wetlands in Louisiana differs from the one proposed by Coyner et al. (2002) in Florida. We found E. ignotus infects the Western Mosquitofish in Louisiana at similar levels as have been observed in Eastern Mosquitofish in Florida. Outbreaks of E. ignotus associated with high hatchling mortality have been documented in Louisiana and Texas (Franson and Custer 1994, Roffe 1988). Assuming the biology of parasite and hosts is similar to that found in Florida, bird infections in Louisiana may be derived from direct consumption of infected Mosquitofish or from the consumption of larger fish infected by eating Mosquitofish harboring E ignotus. Louisiana leads the US in crayfish aquaculture (McClain and Romaire 2007). Large, shallow ponds with high densities of crustaceans are major foraging sites for wading birds (Fleury and Sherry 1995). Although these ponds provide foraging sites for wading birds, they also harbor parasitic nematodes that are potentially lethal to juvenile wading birds. If parasitic nematodes are present in agricultural wetlands then they could lead to a reduction in wading bird reproductive success. As a precaution, we suggest agricultural wetlands be monitored to assess presence and abundance of parasitic nematodes. Such efforts would benefit from verification that the observed prevalence of parasites is high enough to cause a threat to wading birds (although Spalding and 2012 M.C. Luent, M. Collins, C. Jeske, and P. Leberg 421 Forrester 1991 show that ingestion of 4 larvae can be lethal to juveniles) and that Gambusia is the best choice for a sentinel species. If a particular pond was found to have unusually high densities of infected fish, a rapid drawdown or an increase in water depth might be used to reduce the access of birds to fish (Coyner et al. 2002). Investigations into impoundment designs or management to reduce bird use might also be warranted. Acknowledgments We thank M. Kinsella for identification of nematodes, A. Hitch and J. Larrivere for help with field collections, and T. Doyle, K. Krauss, and T. Michot for editorial comments. Literature Cited Coyner, D.F., S.R. Schaack, M.G. Spalding, and D.J. Forrester. 2001. Altered predation susceptibility of mosquitofish infected by Eustrongylides ignotus. Journal of Wildlife Diseases 37:556–560. Coyner, D.F., M.G. Spalding, and D.J. Forrester. 2002. Epizootiology of Eustrongylides ignotus in Florida: Distribution, density, and natural infections in intermediate hosts. Journal of Wildlife Diseases 39:483–499. Coyner, D.F., M.G. Spalding, and D.J. Forrester. 2003. Epizootiology of Eustrongylides ignotus in Florida: Transmission and development of larvae in intermediate hosts. Journal of Parasitology 89:290–298. Deaton, R. 2009. Effects of a parasitic nematode on male mate choice in a live-bearing fish with a coercive mating system (Western Mosquitofish, Gambusia affinis). Behavioral Processes 80:1–6. Fleury, B.E., and T.W. Sherry. 1995. Long-term population trends of colonial wading birds in the southern United States: The impact of crayfish aquaculture on Louisiana populations. The Auk 112:613–632. Franson, J.C. and T.W. Custer. 1994. Prevalence of Eustrongylidosis in wading birds from colonies in California, Texas, and Rhode Island, USA. Colonial Waterbirds 17:168–172 Frederick, P.C., S.M. McGehee, and M.G. Spalding 1996. Prevalence of Eustrongylides ignotus in Mosquitofish (Gambusia holbrooki) in Florida: Historical and regional comparisons. Journal of Wildlife Diseases 32:552–555. Martin, R.P., and G.D. Lester. 1990. Atlas and census of wading bird and seabird nesting colonies in Louisiana: 1990. Louisiana Department of Wildlife and Fisheries and Louisiana Natural Heritage Program Special Publication No. 3. Baton Rouge, LA. 182 pp. McClain, W.R., and R.P. Romaire. 2007. Procambarid crawfish life history and biology. Southern Regional Aquaculture Center, Stoneville, MS. Publication No. 2403. EUA. 6 pp. Roffe, T.J. 1988. Eustrongylides sp. epizootic in young Common Egrets (Casmerodius albus). Avian Diseases 32:143–147. Spalding, M.G., and D.J. Forrester. 1991. Effects of parasitism and disease on the nesting success of colonial wading birds (Ciconiiformes) in southern Florida. Florida Game and Fresh Water Fish Commission Final Report, Tallahassee, FL. 130 pp. 422 Southeastern Naturalist Vol. 11, No. 3 Spalding, M.G., G.T. Bancroft, and D.J. Forrester. 1993. The epizootiology of eustrongylidosis in wading birds (Ciconiiformes) in Florida. Journal of Wildlife Diseases 29:237–249. Weise, J.H., W.R. Davidson, and V.F. Nettles. 1977. Large-scale mortality of nestling ardeids caused by nematode infection. Journal of Wildlife Diseases 13:376–382.