2009 SOUTHEASTERN NATURALIST 8(2):347–354 Impacts of Nutria Removal on Food Habits of American Alligators in Louisiana Steven W. Gabrey1,*, Noel Kinler2, and Ruth M. Elsey3 Abstract - In southern Louisiana, high population densities of exotic Myocastor coypus (Nutria) have been implicated in causing significant coastal marsh damage through extensive herbivory. Wildlife officials instituted a Nutria removal program in 2002 to reduce this marsh loss. Because Alligator mississippiensis (American Alligator) frequently consume Nutria, concern arose regarding the program’s impacts on alligator food habits. Therefore, we conducted our study to determine if the Nutria removal program affected the frequency of occurrence of Nutria remains in alligator stomachs collected from five parishes in southern Louisiana. Three parishes had high Nutria densities and removal programs; two parishes had low Nutria densities and no Nutria removal. We collected >550 alligator stomachs during three September trapping seasons and examined the contents of each. We used logistic regression to model effects of year (1 year prior to the removal program compared to two years during removal) and parish (three with Nutria removal programs compared to two without) on the probability that an alligator stomach contained Nutria remains. Overall, about one-third of the alligator stomachs contained Nutria remains. Nutria removal appeared to have no effect on the probability of a stomach containing Nutria remains even after two years of Nutria removal. In addition, the probability that an alligator stomach contained Nutria remains was similar among all parishes regardless of whether Nutria removal occurred or not. We recommend that continuance of the Nutria removal program be based on its effectiveness in reducing marsh damage and not on perceived impacts to alligator food habits. Introduction Myocastor coypus Molina (Nutria) was introduced from South America into the marshes of southwest Louisiana in the 1930s (Kinler et al. 1987). In the decades since, this large (>5 kg) and highly fecund aquatic rodent has dispersed throughout the marshes of the entire Gulf Coast region into the Atlantic coast as far north as Maryland and Delaware. In some regions, notably the Barataria Basin of southeastern Louisiana and the Chesapeake Bay area, populations of Nutria have become sufficiently dense that their foraging activities have resulted in “eat-outs” of coastal marsh vegetation (Baroch and Hafner 2002, Marx et al. 2004), areas of marsh that have been completely devegetated and consequently significantly contribute to the high rate of marsh deterioration. To reduce the extent of these eat-outs, the Fur and Refuge Division of the Louisiana Department of Wildlife and Fisheries (LDWF) instituted a Nutria 1Department of Biology, Northwestern State University, Natchitoches, LA 71497. 2Louisiana Department of Wildlife and Fisheries, 2415 Darnall Road, New Iberia, LA, 70560. 3Louisiana Department of Wildlife and Fisheries, Rockefeller Wildlife Refuge, 5476 Grand Chenier Highway, Grand Chenier, LA 70643. *Corresponding author - steveng@nsula.edu. 348 Southeastern Naturalist Vol. 8, No. 2 control program in 2002 with the goal of reducing Nutria populations in those marshes with the highest densities and most extensive marsh damage (Marx et al. 2004). Although the general public readily accepted this program, there has been some concern regarding the impact of Nutria removal on Alligator mississippiensis Daudin (American Alligator) populations because alligators frequently consume Nutria (Elsey et al. 1992, McNease and Joanen 1977, Taylor 1986, Wolfe et al. 1987). In this paper, we compare frequency of occurrence of Nutria in alligator stomachs collected from two parishes with high Nutria densities before and during the first two years of the removal program. We also compare frequency of occurrence of Nutria in alligator stomachs collected from five parishes with differing levels of Nutria density and Nutria removal. Valentine et al. (1972) attributed a decline in the frequency of occurrence of Nutria remains in alligator stomachs from 56% to 5% in four years on Sabine National Wildlife Refuge because a Nutria harvest program reduced the Nutria population by an estimated 30%. Similarly, we expected the occurrence of Nutria remains in our stomach samples to decline over the duration of the Nutria removal program in those parishes with Nutria removal. In addition, we expected the frequency of occurrence of Nutria remains to be lower in parishes with Nutria removal programs than in parishes without removal. Methods Alligator stomach collection and content analysis We collected alligator stomachs from cooperative trappers in Terrebonne and Lafourche parishes (Fig. 1) during the September 2002, 2003, and 2004 alligator-trapping seasons. These two parishes comprise part of the Barataria Basin, the region of the state with the greatest density of Nutria and the most severe marsh damage problems (Marx et al. 2004). Nutria removal coincided with the November to March furbearer season and began in November 2002. Thus, alligator stomachs collected in September 2002 came from marshes with high Nutria density before the start of the removal program; alligator stomachs collected in subsequent years came from the same marshes as in 2002 after one (2003) and two (2004) Nutria removal periods. We expanded the geographic range from which we collected stomachs to include locations in Cameron, Vermilion, and St. Charles parishes (Fig. 1) during the 2003 and 2004 September alligator trapping seasons. Marshes in these three parishes had either small Nutria populations and no Nutria removal (Cameron, Vermilion) or limited Nutria removal (St. Charles). We were unable to collect alligator stomachs from these three parishes in 2002. All marshes from which alligators and Nutria were harvested are classified as fresh marsh except those in Lafourche Parish, which are classified as intermediate (Chabreck 1970). In Louisiana, alligator trappers harvest animals from specific marsh locations (leases) through arrangements made with the appropriate landowner and LDWF. Our cooperating trappers harvested alligators from marshes that were owned privately (Terrebonne, Lafourche, Vermilion parishes), federally (Cameron Parish, Lacassine National Wildlife Refuge), or by LDWF 2009 S.W. Gabrey, N. Kinler, and R.M. Elsey 349 (St. Charles Parish, Salvador Wildlife Management Area). All trappers maintained these same leases throughout the duration of our study. In addition, the Nutria removal protocol required that Nutria hunters identify the locations from which they harvested the Nutria. Consequently, we were able to obtain from records maintained by LDWF (Marx et al. 2004) yearly Nutria removal totals for those leases from which alligator stomachs were collected. Thus, the number of Nutria removed reported here is the number removed from only those leases from which we also collected alligator stomachs. Stomachs were removed and length and sex of alligators recorded as trappers brought them to any of several processing sheds. We collected stomachs only from adult alligators (>1.8 m total length) because trappers do not target juveniles (<1.8 m) and because juveniles rarely consume vertebrates (Elsey et al. 1992, Giles and Childs 1949, Wolfe et al. 1987). Stomachs were frozen until analyzed. We identified prey items to the lowest taxon possible based on state of digestion and recorded percent frequency of all food items. Statistical analyses We used backwards stepwise logistic regression (Agresti 1996) to determine which explanatory variables infl uenced the probability that an alligator stomach contained Nutria remains. All analyses were conducted using Proc Figure 1. Louisiana parishes from which American Alligator stomachs were collected during the September 2002, 2003, and 2004 alligator trapping seasons. Nutria removal programs were in effect in Lafourche, Terrebonne, and St. Charles parishes but not in Cameron or Vermilion parishes. 350 Southeastern Naturalist Vol. 8, No. 2 Logistic (probit link, binary distribution, backwards selection) in SAS/STAT 9.1.3 (SAS Institute, Inc., Cary, NC). The initial model included year, parish, and sex, all possible interactions, and an intercept; non-significant effects (P > 0.05) were removed iteratively until only significant effects remained. Although larger alligators consume more mammals than do smaller alligators (McNease and Joanen 1977), we did not include alligator length in our models for the following reasons: (1) Platt et al. (1990) indicated that alligators change from invertebrate to vertebrate prey once they reach the 1.2–1.8-m size class and only five of the alligators from which we collected stomachs were <1.8 m in length) (range = 1.7–1.8 m); (2) stomachs from three of thirteen 1.8-m long alligators in our sample contained Nutria remains; and (3) Elsey et al. (1992) found Nutria remains in stomachs of several alligators that were between 1.2 and 1.8 m in length. Thus, all alligators in our sample were of sufficient size to capture and ingest Nutria. In addition, Proc Logistic and other maximum-likelihood estimators may fail to find a solution when a continuous explanatory variable with many values (such as length) is included in the model; the appropriateness of the resulting model is then questionable (Agresti 1996). We included sex as an explanatory variable because, except for the brief spring mating season, male and female alligators typically use different habitats— males in deep open water or canals compared to females in vegetated marshes and swamps (Goodwin and Marion 1979, Newsom et al. 1987 and references therein). Thus, any between-sex differences in Nutria consumption could refl ect differences in foraging habitat. Pre-Nutria removal vs. post-Nutria removal. We first compared the probability that an alligator stomach contained Nutria remains among the three years (2002, before Nutria removal; 2003, after one year of Nutria removal; and 2004, after two years of Nutria removal) for stomachs collected from the two parishes (Terrebonne and Lafourche) for which we had pre-removal data. In this model, the parameters of interest were the Year main effect or the parish by year interaction, as they indicated a change in the occurrence of Nutria as a function of Nutria removal. Nutria removal vs. no Nutria removal. We also compared the probability that an alligator stomach contained Nutria remains among parishes with low Nutria densities and no removal (Cameron and Vermilion) and parishes with high Nutria densities and removal programs (Terrebonne, Lafourche, and St. Charles) for the two years (2003 and 2004) for which we have data from all five parishes. We used the same approach and initial model as described above. In this second analysis, the parameters of interest are the Parish main effect or the parish by year interaction as they indicate that Nutria occurrence differed among parishes with or without Nutria removal. Results Characteristics of alligators harvested We collected stomachs from 553 alligators harvested during the September trapping seasons of 2002, 2003, and 2004. Ten stomachs lacking data for 2009 S.W. Gabrey, N. Kinler, and R.M. Elsey 351 sex, length, or parish, and 72 stomachs that contained no food items were excluded from analyses. Of the remaining 471 stomachs, 73% were collected from male alligators and 27% were from females. The mean length of harvested alligators used in analyses was 231.0 cm ± 36.5 SD. Males (mean = 238.5 cm ± 38.3 SD) were larger than females (mean = 209.4 cm ± 17.7 SD; t = 11.2, df = 432.8, P < 0.001). Four percent (n = 5) of female alligators were longer than 2.4 m compared to 42% (n = 143) of males. Occurrence of Nutria in alligator stomachs In Terrebonne Parish, the proportion of alligator stomachs containing Nutria remains ranged from 14% in 2002 to 37% in 2004; during this time, more than 13,000 Nutria were removed from the same leases from which the alligators were harvested (Table 1). In Lafourche Parish, the proportion of alligator stomachs containing Nutria remains ranged from 44% in 2002 to 36% in 2004; during this time, more than 8000 Nutria were removed from the same leases from which the alligators were harvested (Table 1). The proportion of alligator stomachs that contained Nutria remains was relatively constant within each of the other three parishes from 2003 to 2004. During that time, >3300 Nutria were removed from alligator leases in St. Charles Parish; no Nutria were removed from leases in Cameron and Vermilion parishes (Table 1). Pre-Nutria removal vs. post-Nutria removal. Backwards removal of non-significant effects resulted in a logistic regression model that retained the parish by sex interaction and its main effects (Table 2). The -2 log L goodness-of-fit test was not significant (X2 = 12.93, df = 8, P = 0.11), indicating that the model was appropriate. The predicted probability that a stomach from a female alligator contained Nutria remains was higher in Lafourche compared to Terrebonne parishes (0.46 and 0.15, respectively). The predicted probability that a stomach from a male alligator contained Nutria remains was similar between the two parishes (0.36 and 0.32 in Lafourche and Terrebonne parishes, respectively). Because the final regression model did not include the variable of interest (year), we do not discuss this model further. Table 1. Percentage of American Alligator stomachs containing Nutria remains (n = number of stomachs examined) and number of Nutria removed from five parishes in southern Louisiana from 2002 through 2004. Nutria were removed from the same trapping leases from which alligator stomachs were collected. Nov. 2002– Nov. 2003– Sept. 2002 Mar. 2003 Sept. 2003 Mar. 2004 Sept. 2004 % of stomachs No. of Nutria % of stomachs No. of Nutria % of stomachs Parish w/Nutria (n) removed w/Nutria (n) removed w/Nutria (n) Terrebonne 14 (57) 8826 27 (77) 5120 37 (86) Lafourche 44 (27) 3842 37 (46) 4665 36 (49) St. Charles 621 31 (35) 3336 34 (44) Cameron 0 33 (15) 0 35 (20) Vermilion 0 27 (11) 0 21 (14) 352 Southeastern Naturalist Vol. 8, No. 2 Nutria removal vs. no Nutria removal. Backwards removal of nonsignificant effects resulted in a logistic regression model that retained the intercept only, indicating that no interactions or main effects infl uenced the probability that an alligator stomach contained Nutria remains (intercept estimate = -0.44, standard error = 0.07; Wald X2 = 44.04, df = 1, P < 0.01). The intercept-only model predicts the probability that an alligator stomach contains Nutria remains is 0.33. The -2 log L goodness-of-fit test was not significant (X2 = 27.25, df = 19, P = 0.15), indicating that the model was appropriate. Because the final regression model did not include the variable of interest (parish), we do not discuss this model further. Discussion In the early part of the twentieth century, mammals—primarily Ondatra zibethicus L. (Muskrat)—were a frequent but variable component of adult alligator food habits in Louisiana (Arthur 1931, Giles and Childs 1949, Kellogg 1929, O’Neil 1949). Following their introduction into the marshes of southwest Louisiana in the 1930s, Nutria became an increasingly important component, while the significance of muskrats declined (Elsey et al. 1992, McNease and Joanen 1977, Valentine et al. 1972, Wolfe et al. 1987). The changes in occurrence of Nutria in alligator food habits studies parallel changes in abundance of Nutria and muskrats in Louisiana marshes (Baroch and Hafner 2002, Kinler et al. 1987), suggesting that alligators readily adapt to variation in prey availability. Valentine et al. (1972) reported that on Sabine National Wildlife Refuge in southwestern Louisiana, the frequency of occurrence of Nutria in alligator stomachs decreased dramatically as a result of an intensive Nutria-trapping program in which about 30% of the Nutria population was removed during 1961–1965. Thus, there is evidence for possible impacts of Nutria removal programs on alligator food habits, although effects of the Nutria removal on alligator demographics or population size were not considered. We expected the current Nutria removal program in Lafourche and Terrebonne parishes to have similar effects on alligator food habits. However, our analysis shows no corresponding decline in the proportion of alligator stomachs containing Nutria remains despite two Nutria removal seasons Table 2. Maximum likelihood estimates, standard errors, and significance tests of logistic regression parameters predicting the probability that an American Alligator stomach contained Nutria remains following backwards stepwise elimination of non-significant effects. The initial model included sex, two parishes (both of which had Nutria removal), and three years (2002, prior to Nutria removal; 2003, after one Nutria removal season; 2004, after two Nutria removal seasons). See Methods for details. Wald Parameter Value df Estimate Std. Error chi-square P > chi-square Intercept 1 -0.49 0.09 29.30 <0.01 Parish Lafourche 1 0.26 0.09 8.36 <0.01 Sex Female 1 -0.07 0.09 0.53 0.47 Parish x sex Lafourche x female 1 0.20 0.09 5.10 0.02 2009 S.W. Gabrey, N. Kinler, and R.M. Elsey 353 during which more than 8000 and 13,000 Nutria were removed from the alligator leases in Lafourche and Terrebonne parishes, respectively. We also expected to detect differences in the proportion of alligator stomachs containing Nutria remains between parishes with low Nutria densities and no removal (Cameron and Vermilion) and parishes with high Nutria densities and removal (Lafourche, Terrebonne, and St. Charles); however, we found no such difference. One possible explanation for the differences between our results and those of Valentine et al. (1972) is that in our study, two years of Nutria removal in Terrebonne and Lafourche parishes were insufficient to reduce Nutria populations to such low densities that they became unavailable to alligators. The amount of time necessary to see such a decrease in Nutria density and a resultant effect on alligator food habits likely depends on the density of Nutria prior to the removal and the proportion of the Nutria population removed, as well as alligator density prior to and during the removal. We examined alligator stomach contents over two years of Nutria removal compared to four years by Valentine et al. (1972). Thus, it is possible that additional Nutria removal is necessary in our study area for a change in alligator food habits to become apparent. Although the current Nutria removal program has continued every year since the end of our study, we were unable to continue collecting stomachs past 2004 to determine any such long-term impacts of Nutria removal. An annual average of 296,000 Nutria have been removed from throughout Louisiana over the five years of the Nutria removal program (Scarborough and Mouton 2007). During the same period, Louisiana’s coastal alligator population increased from an estimated 34,635 nests in July 2002 to 42,150 nests in July 2007 (LDWF, New Iberia, LA, unpubl. data). This population increase, combined with our data, strongly suggest that the current Nutria removal program has no negative impacts on alligator food habits or on population density. Thus, we recommend that decisions regarding the continuance of the Nutria removal program be based on the efficacy of achieving the objective of reducing the extent of “eat-outs” to reduce marsh damage. Acknowledgments This study was funded by the Louisiana Department of Wildlife and Fisheries as part of the Nutria Damage Control and Vegetative Restoration Research program, Award # NA16F22601 Authority NOAA, OMB Circular A-87 and A-133, 15 CFR Part 24. We thank W. Parke Moore III for administrative assistance in support of this research and L. Morris for statistical guidance. LDWF field personnel, particularly G. Linscombe, E. Mouton, and J. Marx, helped with stomach collections, and provided Nutria harvest data and other guidance. J. Jenkins, K. Franklin, and K. Capshaw separated and identified stomach contents. Numerous undergraduate students from Northwestern State University helped with stomach collections. Three anonymous reviewers provided valuable input on earlier drafts of this manuscript. We are grateful for the cooperation of the many licensed alligator trappers and processors who provided us with logistical help and access to their facilities. 354 Southeastern Naturalist Vol. 8, No. 2 Literature Cited Agresti, A. 1996. An Introduction to Categorical Data Analysis. John Wiley and Sons, New York, NY. 290 pp. Arthur, S.C. 1931. The Fur Animals of Louisiana. Bulletin 18. Louisiana Department of Conservation, New Orleans, LA. 433 pp. Baroch, J., and M. Hafner. 2002. Task I. Biology and natural history of the Nutria, with special reference to the Nutria (Myocaster coypus) in Louisiana. Report prepared by Genesis Laboratories, Wellington, CO, for Louisiana Department of Wildlife and Fisheries, New Iberia, LA. 155 pp. Available online at http://www.brownmarsh.net/ data/IV-1/NutriaReport.pdf. Accessed January 9, 2008. Chabreck, R.H. 1970. Marsh zones and vegetative types in the Louisiana coastal marshes. Ph.D. Dissertation, Louisiana State University, Baton Rouge, LA. 112 pp. Elsey, R.M., L. McNease, T. Joanen, and N. Kinler. 1992. Food habits of native wild and farm-released juvenile alligators. Proceedings of the Southeastern Association of Fish and Wildlife Agencies 46:57–66. Giles, L.W., and V.L. Childs. 1949. Alligator management of the Sabine National Wildlife Refuge. Journal of Wildlife Management 13:16–28. Goodwin, T.M., and W.R. Marion. 1979. Seasonal activity ranges and habitat preferences of adult alligators in a north-central Florida lake. Journal of Herpetology 13:157–164. Kellogg, A.R. 1929. The habits and economic importance of alligators. United States Department of Agriculture Technical Bulletin 147:1–36. Kinler, N.W., G. Linscombe, and P.R. Ramsey. 1987. Nutria. Pp. 327–343, In M. Novak, J.A. Baker, M.E. Obbard, and B. Malloch (Eds.). Wild Furbearer Management and Conservation in North America. Ontario Trappers Association, Toronto, ON, Canada. 1150 pp. Marx, J., E. Mouton, and G. Linscombe. 2004. Nutria harvest distribution 2003–2004 and a survey of Nutria herbivory damage in coastal Louisiana in 2004. Louisiana Department of Wildlife and Fisheries, Fur and Refuge Division, New Iberia, LA. 45 pp. Available online at http://www.Nutria.com/uploads/0304Harvestand04Da mageReport.pdf. Accessed January 9, 2008. McNease, L., and T. Joanen. 1977. Alligator diets in relation to marsh salinity. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 31:36–40. Newsom, J.D., T. Joanen, and R.J. Howard. 1987. Habitat suitability models: American Alligator. United States Fish and Wildlife Service Biological Report 82(10.136). 14 pp. O’Neil, T. 1949. The Muskrat in the Louisiana coastal marshes. Louisiana Department of Wildlife and Fisheries, New Orleans, LA. 152 pp. Platt, S.G., C.G. Brantley, and R.W. Hastings. 1990. Food habits of juvenile American Alligators in the Upper Lake Pontchartrain Estuary. Northeast Gulf Science 11:123–130. Scarborough, J., and E. Mouton. 2007. Nutria harvest distribution 2006–2007 and a survey of Nutria herbivory damage in coastal Louisiana in 2007. 40 pp. Available online at http://www.Nutria.com/uploads/0607Finalreport.pdf. Accessed 9 January 2008. Taylor, D. 1986. Fall foods of adult alligators from cypress lake habitats. Proceedings of the Southeastern Association of Fish and Wildlife Agencies 40:338–341. Valentine, J.M., J.R. Walther, K.M. McCartney, and L.M. Ivy. 1972. Alligator diets on the Sabine National Wildlife Refuge, Louisiana. Journal of Wildlife Management 36:809–815. Wolfe, J.L., D.K. Bradshaw, and R.H. Chabreck. 1987. Alligator feeding habits: New data and a review. Northeast Gulf Science 9:1–8.