2009. Proceedings of the Indiana Academy of Science 118(1):88-95

MESOCOSM EVALUATION OF WESTERN IMPACTS ON NORTHERN STARHEAD TOPMINNOWS

Trent M. Sutton1 and Rebecca A. Zeiber2: Department of Forestry and Natural Resources, Purdue University, 195 Marsteller Street, West Lafayette, Indiana 47907 USA Brant E. Fisher: Indiana Department of Natural Resources, Atterbury and Wildlife Area, 7970 S. Rowe St., P.O. Box 3000, Edinburgh, Indiana 46124 USA

ABSTRACT. Western mosquitofish, stocked frequently throughout Indiana, are highly aggressive and often negatively impact native that occupy a similar ecological niche. Because northern starhead topminnows are the most vulnerable of the Indiana topminnows to western mosquitofish aggression, we conducted a 110-day outdoor mesocosm experiment from June through October 2006 to examine changes in abundance, biomass, and wet weight per individual fish at different topminnow and mosquitofish densities. Control mesocosms (three replicates per treatment) contained either 30 northern starhead topminnows or 30 western mosquitofish, while experimental treatments (three replicates per treatment) contained both species at one of the following density levels: (1) 20 topminnows and 10 mosquitofish; (2) 15 individuals of each species; and (3) 10 topminnows and 20 mosquitofish. During the experiment, western mosquitofish exhibited successful reproduction while northern starhead topminnows declined in mean abundance, biomass, and wet weight per individual, indicating that the presence of western mosquitofish had a negative impact on topminnows. However, northern starhead topminnows did not have any negative impact on western mosquitofish. Based on these results, we do not recommend stocking western mosquitofish into Indiana waterways. Keywords: Western mosquitofish, northern starhead topminnow, abundance, biomass, mesocosm

Non-native species have become established species. In laboratory coexistence studies, in aquatic systems worldwide (Li & Moyle Sonoran topminnows (Poeciliopsis occidentalis) 1999; McKinney 2002; Casal 2006). Although did not successfully produce offspring and past attitudes regarding the stocking of non­ exhibited caudal fin damage in the presence of native fishes have been largely favorable, recent mosquitofish (Meffe 1985). Early life stages of research has demonstrated the potential for the Barrens topminnow ( julisia) were negative impacts (Schoenherr 1981; Fausch & preyed upon by western mosquitofish ( Gambu­ White 1986; Moyle & Light 1996; Li & Moyle sia affinis) in the laboratory, while adult 1999; Gale et al. 2004; Rieman et al. 2006). As a topminnows suffered physical impairment due consequence of these introductions, non-native to fin nips (Laha & Mattingly 2007). In Indiana fishes can lead to large-scale changes in aquatic waters, western mosquitofish and native top­ systems and population declines of native minnows (genus Fundulus) have been observed species (Moyle 1986; Everett & Sherfy 2001). to consume similar prey items, indicating high Competition for limited resources and ago­ diet overlap and potential trophic competition nistic behaviors or predation can result in (Clem & Whitaker 1995; Zeiber 2007). Western negative impacts by introduced fishes on native mosquitofish also chased and attacked topmin­ nows in laboratory microcosm studies which, in Correspondence: Trent M. Sutton, Department of some cases, resulted in the mortality of Forestry and Natural Resources, 195 Marsteller individual topminnows (Zeiber 2007). Street, Purdue University, West Lafayette, IN 47907 Behavioral responses of native fishes to 1 Current address: School of Fisheries and Ocean introduced species may also lead to a suite of Sciences, Fisheries Division, University of Alaska indirect negative effects from a subsequent shift Fairbanks, 905 Koyokuk Drive, Fairbanks, Alaska 99775 in niche use (Savino & Stein 1989). Smaller fish 2 Current address: New Hampshire Sea Grant, King­ are often restricted to habitats that offer man Farm, University of New Hampshire, Durham, protection from predators, but those habitats New Hampshire 03824 may have fewer resources and could lead to

88 SUTTON ET AL-MOSQUITOFISH AND TOPMINNOWS 89 increased intraspecific competition (Mittelbach dip net (diameter= 40 mm; Memphis Net and 1986). For example, small bluegills (Lepomis Twine, Memphis, Tennessee). These nets were macrochirus) are often restricted to locations in effective during collections because most fish natural lakes which contain aquatic vegetation were located at water depths less than 15 cm. A due to the risk of open-water predation by range of fish sizes was collected, including largemouth bass (Micropterus salmoides) (Wer­ males and females of both species; however, ner et al. 1983), and a decrease in foraging rates pregnant female mosquitofish were not includ­ may occur when fish are restricted to particular ed in the study. Following collections, fish were habitats. In contrast, bluegills utilized open­ placed into 8.93 1 ( ~ 8 gallons) buckets of water habitats and consumed less-preferred, water containing aquatic vegetation and trans­ smaller prey items when a larger, more­ ported to the Purdue University Cunningham aggressive species such as green sunfish (Lepo­ Forest complex (Lafayette, Indiana). mis cyanellus) were present (Werner & Hall Mesocosm experiment.-A 110-day outdoor 1977). However, the impacts on survivorship mesocosm study was used to examine changes and reproduction from a shift in habitat and/or in abundance and body size of western mos­ prey resource use for most fishes are often quitofish and northern starhead topminnows. unknown (Marchetti 1999). Fifteen 416 1 plastic mesocosms were placed in In terms of non-native fish introductions, a 10 X 10 m area in a location that allowed limited coordinated or inconsistent manage­ partial shade during the day. Mesocosms were ment actions coupled, with a general lack of filled with 379 1 of groundwater, inoculated recognition of the problem, can result in with 38 1 of pond water containing duckweed inadequate measures to protect ecosystem (Lemna spp., to provide overhead visual cover), balance (Koehn & MacKenzie 2004). For phytoplankton, zooplankton, and other aquat­ example, the Indiana Department of Natural ic macroinvertebrates to simulate natural pond Resources (IDNR) currently does not allow the conditions, and allowed to colonize for four stocking of mosquitofish within state public weeks prior to the introduction of the fish. waters. However, mosquitofish can be stocked Each mesocosm also contained four clusters of in private waters and, consequently, have artificial vegetation: two clusters were com­ escaped to public aquatic systems. Although posed of sixteen 45 cm long sections of green all topminnow species in Indiana are potential­ plastic tarp (to simulate submerged vegetation), ly vulnerable to western mosquitofish introduc­ while the other two clusters were composed of tions, Zeiber (2007) found that the northern sixteen 90 cm long sections of tarp (to simulate starhead topminnow (Fundulus dispar) was the floating vegetation and provide shade at the most vulnerable killifish species to western surface). For each vegetation cluster, plastic mosquitofish aggression in laboratory micro­ strips were attached to a 3.18 cm hex nut which cosm experiments. As a result, we conducted an held the cluster secure on the bottom. These outdoor mesocosm study to examine changes in vegetation clusters simulated the dense aquatic abundance, biomass, and wet weight per vegetation that both species prefer as feeding and individual of northern starhead topminnow at nursery areas, and also provided a substrate for different mosquitofish densities. The results of female northern starhead topminnows to deposit this research will fill a critical need to under­ their eggs (Balon 1981; Becker 1983). stand the impacts of western mosquitofish on Control mesocosms contained either 30 native topminnows in Indiana waters. northern starhead topminnows or 30 western mosquitofish, while the mesocosms for the METHODS experimental treatments contained both fish Field collections.-Adult western mosquito­ species at one of the following three density fish were collected from Martell Forest Pond combinations: (I) 20 northern starhead top­ (Tippecanoe County, Indiana), while northern minnows and 10 western mosquitofish; (2) 15 starhead topminnows were captured from Pine individuals of each species; and (3) 10 northern Lake and Upper Fish Lake (Laporte County, starhead topminnows and 20 western mosqui­ Indiana). Both species were collected during tofish. Thus, there were two control groups and May 2006, and sampling was conducted using a three density treatments, each with three 3.18 mm knotless mesh seine (length = 3.05 m; replicates. For both species, the sex ratio in depth = 1.22 m) and 3.18 mm knotless mesh each mesocosm was 50:50 males:females. At the 90 PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE

Table !.-Fish densities for control and treatment groups used during the mesocosm experiments, including northern starhead topminnows (NST) and western mosquitofish (WMF), mean number of fish at the end of the experiment and mean biomass (g) and mean wet weight (g) of individual fish before and after the experiment.

Mean number Mean biomass Mean biomass Wet weight Wet weight Number of fish after (g) before (g) after (g) before (g) after Experiment of fish experiment experiment experiment experiment experiment Control 30 NST 4 36.94 2.57 1.23 0.62 (3 - 5) (29.46 - 48.30) (2.32 - 2. 73) (0.98 - 1.61) (0.53 - 0.77) 30WMF 27 12.30 3.68 0.41 0.14 (17 - 38) (10.13 - 15.41) (2.83 - 4. 77) (0.34 - 0.51) (0.11 - 0.20) Treatment 20 NST

300 -.--~~~~~~~~~~~~~~~~~~~~~~~~~~~-----, Numbers C=:::J NST 250 -WMF 200 150 100 50

0 ;----,--.------,------.~ -50 c -100 Q) E -150~~~~~~~~~~~~~~~~~~~~~~~~~ ·;:: Q) $ 50 Biomass E rn 8 0 -+------.-~----- 0rn Q) ~ Ol -50 c ·;:: ::i 0 Q) -100 Ol c: ct! ..c: (.) -150 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

cQ) ~ 50 ~------, ~ Wet Weight I Individual

0 +------.--.,...-----

-50

-100

-150 ~~~~.--~~~~--.-~~~~---,-~~~~~.--~~~~-,---~~~ 30 NST 30WMF 20 NST I 15 NST I 10 NST I 10WMF 15WMF 20WMF Figure 1.-Mean percent changes in the number of individuals. total mesocosm biomass. and wet weight per individual northern starhead topminnow (NST) and western mosquitofish (WMF) during the mesocosm experiment for each experimental treatment.

as large a change from the initial density (t = cosm declined during the experiments, with the 1.0, p = 0.211). exception of western mosquitofish biomass at The biomass of northern starhead topmin­ the lowest density (Table 1; Fig. 1). Northern nows and western mosquitofish in each meso- starhead topminnow biomass in control meso- 92 PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE

cosms declined from a mean of 36.9 g to 2.57 g a decline in the mean abundance, biomass, and (-93% ), and biomass of western mosquitofish wet weight per individual for topminnows in all in control mesocosms also declined from a treatments, the declines were more pronounced mean of 12.29 g to 3.68 g (-70%). In meso­ in mesocosms also containing mosquitofish. cosms with an initial density of 20 northern Topminnows did survive in control mesocosms starhead topminnows and 10 western mosqui­ throughout the experiment, so the presence of tofish, the mean biomass for northern starhead western mosquitofish had an impact on this topminnows declined from 20.8 g to 0.12 g species. Rogowski & Stockwell (2006) showed (-9% ), but the change in western mosquito fish that when western mosquitofish were kept in biomass from 3.83 g to 4.13 g represented an aquaria with White Sands pupfish ( Cyprinodon increase in biomass (+8%). At equal densities tularosa), growth rates and biomass were lower (15 fish of each species), the biomass of than when the latter species was maintained northern starhead topminnows declined from alone. Eastern mosquitofish predation on 20.75 g to 0.00 g (-100%), but the decline in smaller least killifish (Heterandria formosa) western mosquitofish biomass from 10.68 g to caused a shift in size structure of the latter 5.90 g was not nearly as large (-45%; t = 2.2, species in mesocosms, leaving only larger P = 0.082). In mesocosms that contained 10 female topminnows (Schaefer et al. 1994). In northern starhead topminnows and 20 western field coexistence studies with western mosqui­ mosquitofish, there was a decline in biomass of tofish and Sonoran topminnows, all topmin­ northern starhead topminnows from 11.31 g to nows exhibited fin or body damage within two 0.00 g ( -100%) and western mosquitofish from weeks, and had died and were consumed by 12.85 g to 4.65 g (-64%). mosquitofish within three weeks (Meffe 1985). The wet weight per individual fish of Western mosquitofish were not impacted by northern starhead topminnows and western northern starhead topminnows during the mosquitofish in each mesocosm declined during mesocosm experiment. For most mesocosms, the experiment (Table l; Fig. 1). Northern there was typically a higher mosquitofish starhead topminnow mean weight per individ­ density at the end of the experiment than were ual in control mesocosms declined from 1.23 g initially stocked. Although both mean biomass to 0.62 g (-50% ). However, the mean weight and wet weight per individual fish in meso­ per individual for western mosquitofish in cosms for mosquitofish declined during the control mesocosms declined from 0.41 g to experiment, most of the fish at the end of the 0.14 g ( -66% ). In mesocosms with 20 northern experiment were juveniles, indicating that there starhead topminnows and 10 western mosqui­ was successful reproduction. This occurred in tofish initially, the mean weight per individual both control and experimental mesocosms, so declined for northern starhead topminnows the presence of topminnows did not influence from 1.19 g to 0.12 g (-90%) and western mosquitofish reproduction. In previous studies, mosquitofish from 0.38 g to 0.16 g ( - 58% ). At eastern mosquitofish exhibited a 50% increase equal densities (15 fish of each species), the in overall biomass in outdoor mesocosms, mean weight per individual declined for north­ which was due to an increase in density from ern starhead topminnows from 1.38 g to 0.00 g reproduction rather than increased fish size (-100%) and western mosquitofish from 0. 71 g (Sepulveda et al. 2005). Schaefer et al. (1984) to 0.19 g ( - 73% ). In mesocosms that initially showed that the density, biomass, and mean contained 10 northern starhead topminnows length per individual of eastern mosquitofish in and 20 western mosquitofish, the mean weight outdoor microcosms in a Florida marsh were per individual of topminnows declined from not influenced by least killifish. Similarly, 1.13 g to 0.00 g ( -100% ), but the decline in western mosquitofish exhibited increases in western mosquitofish weight per individual mean biomass and wet weight per individual, from 0.64 g to 0.25 g was not as large (-61%). regardless of the presence or absence of threespine stickleback ( Gasterosteus aculeatus) DISCUSSION (Offill & Walton 1999). The potential impacts of stocking western Although few northern starhead topminnows mosquitofish into systems containing northern survived in mesocosms that also contained starhead topminnows were simulated in a 11 O­ mosquitofish, there was also a large decline in day mesocosm experiment. Although there was the number of northern starhead topminnows SUTION ET AL-MOSQUITOFISH AND TOPMINNOWS 93 in control mesocosms. As a result, the observed Northern starhead topminnows and, to a declines in mean abundance, biomass, and wet lesser extent, western mosquitofish may also weight per individual may not have been caused have experienced a natural die-off over the solely by western mosquitofish. The upper summer in the control and experimental thermal tolerance of northern starhead top­ mesocosms. The mean longevity for northern minnows is unknown, but water temperatures starhead topminnows is two years, with some in mesocosms ranged from 10-27°C, which is individuals surviving to age 3 (Becker 1983; within the known spawning temperature range Taylor & Burr 1997). In contrast, few western for this species (l 7°-30°C; Taylor & Burr 1997). mosquitofish survive past one year, and the These mesocosm temperatures were also within maximum lifespan for females is 1.5 years the upper thermal tolerance limit (38°C) and (Daniels & Felley 1992; Haynes & Cashner spawning temperature range for mosquitofish 1995). Therefore, the fish used in the mesocosm (Otto 1974; Becker 1983). Therefore, water study may have been near the end of their life temperature was probably not a factor that span and, consequently, were unable to survive contributed to the differential success of the entire experimental period. While Zeiber mosquitofish relative to topminnows. (2007) successfully maintained both species in Both western mosquitofish and northern star­ the laboratory at similar environmental condi­ head topminnows spawn during spring and tions, those fish were almost strictly juveniles or summer months (Becker 1983; Taylor & Burr young adults and not senescent. Regardless of 1997). Female western mosquitofish have a the reason(s) for the topminnow declines, it is protracted spawning period, produce two to six likely that the negative effects that occurred in broods throughout the summer, and the number the experimental mesocosms were exacerbated oflive offspring produced per brood ranges from by the presence of western mosquitofish. 14--218 (Krumholz 1948; Haynes & Cashner Due to the impacts of western mosquitofish 1995). In contrast, northern starhead topmin­ on northern starhead topminnows observed nows only spawn from mid-April to mid-July, during this study, we recommend that the former species not be stocked into any aquatic produce only two egg clutches during this period, systems in Indiana. Unfortunately, western and the number of eggs per clutch ranges from 7- mosquitofish may already inhabit systems that 30 (Taylor & Burr 1997). Consequently, these support northern starhead topminnows. For differences may have contributed to the differ­ example, western mosquitofish were found for ential reproduction success of northern starhead the first time in 2006 in Loomis Lake, Indiana, topminnows (no reproduction) and western a system that has historically supported a large mosquitofish (successful reproduction). population of northern starhead topminnows. The declines in the number of northern Therefore, additional research is required to starhead topminnows in experimental meso­ determine the potential long-term impacts of cosms may also be due to food limitations. western mosquitofish on northern starhead Although the initial pond-water inoculation topminnows in systems where they co-occur. contained zooplankton and other aquatic macroinvertebrates, the density of these prey ACKNOWLEDGMENTS resources may not have been sufficient to We would like to thank L. Edenfield, N. support both topminnows and mosquitofish Richardson, J. Hoffmeister, and M. Sepulveda in the same mesocosm. Zeiber (2007) found for their assistance in field collections and the that these two species in Indiana had nearly mesocosm experiment design. The experimental identical diets that consisted of zooplankton, procedures used in this research were approved culicid (mosquito) larvae, other aquatic mac­ by the Purdue University Care and Use roinvertebrates, and terrestrial insects. Compe­ Committee as protocol 01-058. Permits for the tition for limited prey resources may have collection of western mosquitofish and north­ allowed only one species to receive adequate ern starhead topminnows were provided by the nutrition, and due to the aggressive nature of Indiana Department of Natural Resources (#s mosquitofish, this species was most likely able 06-0005 and 06-0022). This project was funded to outcompete topminnows. However, this does through the Indiana Nongame Fund and State not explain why topminnow numbers declined Wildlife Grants T-7-R-l through the Indiana in control mesocosms. Department of Natural Resources. Additional 94 PROCEEDINGS OF THE INDIANA ACADEMY OF SCIENCE funding for this research was provided by the Marchetti, M.P. 1999. An experimental study of American Fisheries Society Hutton Junior competition between the native Sacramento perch Fishery Biologist program. (Archoplites interruptus) and introduced bluegill (Lepomis macrochirus). Biological Invasions LITERATURE CITED 1:55-65. McKinney, M.L. 2002. Do human activities raise Balon, E.K. 1981. Additions and amendments to the species richness? Contrasting patterns in United classification of reproductive styles in fishes. States plants and fishes. Global Ecology and Environmental Biology of Fishes 6:377-389. Biogeography 11 :343-348. Becker, G.C. 1983. The Fishes of Wisconsin. The Meffe, G.K. 1985. Predation and species replacement University of Wisconsin Press, Madison. in American southwest fishes: A case study. The Casal, C.M.V. 2006. Global documentation of fish Southwestern Naturalist 30:173-187. introductions: The growing crisis and recommen­ Mittelbach, G. 1986. Predator-mediated habitat use: dations for action. Biological Invasions 8:3-11. Some consequences for species interactions. Envi­ & Clem, P.D. J.O. Whitaker, Jr. 1995. Distribution ronmental Biology of Fishes 16:159-169. of the mosquitofish, Gambusia affinis (Baird & Moyle, P.B. 1986. Fish introductions into North Girard), in Indiana, with comments on resource America: Patterns and ecological impact. Pp. 27-43. competition. Proceedings of the Indiana Academy In Ecology of Biological Invasions of North 104:249-258. of Sciences America and Hawaii (H.A. Mooney & J.A. Drake, Daniels, G.L. & J.D. Felley. 1992. Life history and eds.). Springer-Verlag, New York, New York. foods of Gambusia affinis in two waterways of Moyle, P.B. & T. Light. 1996. Fish invasions in southwestern Louisiana. Southwesern Naturalist California: Do abiotic factors determine success? 37:157-165. Ecology 77:1666-1670. Everett, R.A. & M.H. Sherfy. 2001. The Chesapeake Offill, Y.A. & W.E. Walton. 1999. Comparative Bay: A model for regional approaches to the efficacy of the threespine stickleback ( Gasterosteus prevention and control of aquatic non-indigenous aculeatus) and the mosquitofish (Gambusia affinis) species. Transactions of the North American for mosquito control. Journal of the American Wildlife and Natural Resources Conference Mosquito Control Association 15:380-390. 66:611-624. Otto, R.G. 1974. The effects of acclimation to cyclic Fausch, K.D. & R.J. White. 1986. Competition thermal regimes on heat tolerance of the western among juveniles of coho salmon, brook trout, and mosquitofish. Transactions of the American Fish­ brown trout in a laboratory stream, and implica­ eries Society 103:331-335. tions for Great Lakes tributaries. Transactions of Rieman, B.E., J.T. Peterson & D.L. Myers. 2006. the American Fisheries Society 115:363-381. Have brook trout (Salvelinus fontinalis) displaced Gale, W.L., M.S. Hill & G.B. Zydlewski. 2004. bull trout (Salvelinus confluentus) along longitu­ Physiological and behavioral differences of hatch­ dinal gradients in central Idaho streams? Canadi­ ery and wild-reared steelhead Oncorhynchus my­ an Journal of Fisheries and Aquatic Sciences kiss smolts of the same origin. Journal of Fish 63:63-78. Biology 65:328-329. Rogowski, D.L. & C.A. Stockwell. 2006. Assessment Haynes, J.L. & R.C. Cashner. 1995. Life history and of population impacts of exotic species on population dynamics of the western mosquitofish: populations of a threatened species, White Sands A comparison of natural and introduced popula­ pupfish, Cyprinodon tularosa. Biological Invasions tions. Journal of Fish Biology 46:1026-1041. 18:79-87. Koehn, J.D. & R.F. MacKenzie. 2004. Priority Savino, J.F. & R.A. Stein. 1989. Behavioral inter­ management actions for alien freshwater fish actions between fish predators and their prey: species in Australia. New Zealand Journal of Effects of plant density. Animal Behavior 37: Marine and Freshwater Research 38:457-472. 311-321. Krumholz, L.A. 1948. Reproduction in the western Schaefer, J.F., S.T. Heulett & T.M. Farrell. 1994. mosquitofish (Gambusia affinis) and its use in Interactions between two poeciliid fishes (Gambu­ mosquito control. Ecological Monographs 18:1-47. sia holbrooki and Heterandia formosa) and their Laha, M. & H.T. Mattingly. 2007. Ex situ evaluation prey in a Florida marsh. Copeia 2:516-520. of impacts of invasive mosquitofish on the Schoenherr, A.A. 1981. The role of competition in imperiled Barrens topminnow. Environmental the replacement of native fishes by introduced Biology of Fishes 78:1-11. species. Pp. 173-203. In Fishes in North American Li, H.W. & P.B. Moyle. 1999. Management of Deserts (R.J. Naiman & D.L. Soltz, eds.). John introduced fishes. Pp. 345-374. In Inland Fisheries Wiley, New York. Management in North America, 2nd edition (C.C. Seplliveda, M.S., T.S. Gross, S.D. Ruessler & J.A. Kohler & W.A. Hubert, eds.). American Fisheries Grosso. 2005. Microcosm evaluation of the bioac­ Society, Bethesda, Maryland. cumulation of organochloride pesticides from soils SUTTON ET AL-MOSQUITOFISH AND TOPMINNOWS 95

in the North Shore Restoration Area at Lake Werner, E.E., J.F. Gillia, D.J. Hall & G.G. Apopka. The Department of Water Resources, St. Mittelbach. 1983. An experimental test of the Johns River Water Management District, Special effects of predation risk on habitat use in fish. Publication SJ2005-SP11, Palatka, Florida. Ecology 64:1540-1548. Taylor, C.A. & B.M. Burr. 1997. Reproductive Zeiber, R.A. 2007. Effects of western mosquito­ biology of the northern starhead topminnow, fish on Indiana aquatic communities. Master's Fundulus dispar (Osteichthyes: ), with thesis. Purdue University, West Lafayette, a review of data for freshwater members of Indiana. the genus. American Midland Naturalist 137: 151-164. Werner, E.E. & D.J. Hall. 1977. Competition and habitat shift in two sunfishes (Centrarchidae). Manuscript received 19 May 2008, revised 22 Ecology 58:869-876. September 2008.