1 NATIVE MOSQUITOFISH AS BIOTIC RESISTANCE TO INVASION: PREDATION ON TWO NONINDIGENOUS POECILIIDS By KEVIN ALLEN THOMPSON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008
2 2008 Kevin Allen Thompson
3 To my Mom and Dad, for always encouraging my academic pursuits
4 ACKNOWLEDGMENTS I thank m y chair, Dr. Jeffrey Hill for i nvaluable guidance and support throughout this research as well as my supervisory committee, Dr. Charles Cichra and Dr. Leo Nico for their roles as mentors. I thank Craig Watson, Dan Bu ry, Jonathan Foster, Scott Graves, Kathleen Hartman, Carlos Martinez, Debbie Pouder, and Amy Wood of the Tropical Aquaculture Lab for research support as well as Colin Calway of Happy Trails Aquatics for providing research animals. Gary Meffe provided information in this study and statistics consulting was provided by James Colee of the IFAS statistics consulting unit. I thank the faculty and students of the Department of Fisheries and Sciences as we ll as my friends and family that provided encouragement throughout this process. This res earch was supported in part by the University of Florida Institute of Food and Agri cultural Sciences and a grant from the University of Florida School of Natural Resources and the Environment.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT.....................................................................................................................................9 CHAP TER 1 INTRODUCTION..................................................................................................................11 2 METHODS.............................................................................................................................19 Study Species..........................................................................................................................19 Mesocosm Overview.............................................................................................................. 21 Experiment 1: Adult Introduction........................................................................................... 22 Experiment 2: Stage-St ructured Population ...........................................................................24 Experiment 3: Effects of Structural Complexity.................................................................... 25 Experiment 4: Behavioral Measurements............................................................................... 26 Analysis of Differing Stem Densities (4A)..................................................................... 27 Analysis of Eastern Mosquitofish Presence (4B)............................................................ 27 Data Analysis..........................................................................................................................29 3 RESULTS...............................................................................................................................34 Experiment 1: Adult Introduction........................................................................................... 34 Experiment 2: Stage-St ructured Population ...........................................................................35 Experiment 3: Effects of Structural Complexity.................................................................... 36 Experiment 4: Behavioral Measurements............................................................................... 37 Analysis of Differing Stem Density (4A)........................................................................ 37 Day 1........................................................................................................................37 Day 3........................................................................................................................38 Analysis of Eastern Mosquitofish Presence (4B)............................................................ 39 Day 1........................................................................................................................39 Day 3........................................................................................................................39 4 DISCUSSION.........................................................................................................................50
6 APPENDIX A FRESHWATER FISH INTROD UCTIONS INTO FLORIDA .............................................. 61 B DATA SHOWN IN FIGURES FOR ALL EXPERIMENTS ................................................. 63 LIST OF REFERENCES...............................................................................................................66 BIOGRAPHICAL SKETCH.........................................................................................................73
7 LIST OF TABLES Table page 2-1 Overview of four different mesocosm experim ents testing th e effects of eastern mosquitofish (MF) predation on two nonindigenous poeciliids........................................ 32 3-1 Summary of caudal fin damage results for surviving adult variab le platyfish in all experim ents and treatments................................................................................................ 41 3-2 Summary of caudal fin damage results fo r surviving swordtails in all experim ents and treatments................................................................................................................. ...42 A-1 Status of known freshwater fish spec ies with m aximum TL < 15cm introduced into the state of Florida........................................................................................................... ..61 B-1 Survival data from all treatments of E xperim ents 1-3 for adults and juveniles of variable platyfish and swordtails....................................................................................... 63 B-2 Total attack data from all treatments of Experim ents 4a and 4b for adults of variable platyfish and swordtails.....................................................................................................64 B-3 Habitat use data from all treatments of E xperim ents 4a and 4b for adults of variable platyfish and swordtails.....................................................................................................65
8 LIST OF FIGURES Figure page 2-1 Pictures of experime ntal tank and vegetation.. .................................................................. 33 3-1 Results from Experiment 1................................................................................................43 3-2 Results from Experiment 2................................................................................................44 3-3 Results from Experiment 3................................................................................................45 3-4 Total attack results from Experiment 4a............................................................................ 46 3-5 Habitat use results from Experiment 4a............................................................................. 47 3-6 Total attack results from Experiment 4b............................................................................ 48 3-7 Habitat use results from Experiment 4b............................................................................. 49
9 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science NATIVE MOSQUITOFISH AS BIOTIC RESISTANCE TO INVASION: PREDATION ON TWO NONINDIGENOUS POECEILIIDS By Kevin Allen Thompson August 2008 Chair: Jeffrey E. Hill Major: Fisheries a nd Aquatic Sciences Widespread introduction of nonindigenous a quatic species by anthropogenic means has resulted in homogenization of worldwide fa unas and some negative ecological and economic impacts. When a species is in troduced outside of its natural ra nge, various abiotic and biotic factors influence the ability of a nonindigenous species to successfully establish. Native predators can contribute to biotic resistance, the reduction of invasion success by the native biotic community. There have been numerous freshwater fish intr oductions into Florida, but few small-bodied freshwater species have maintained reproducing populations. Eastern mosquitofish Gambusia holbrooki is a widespread and abunda nt native fish throughout the freshwater systems of Florida and is known to prey on small fish, a nd recent data have indicated that mosquitofish may consume fish similar to their own size. To test the hypothesis that eas tern mosquitofish predation cr eates a biotic resistance to invasions of small-bodied freshw ater fish, a series of mesoco sm experiments were conducted using two nonindigenous speci es, variable platyfish Xiphophorus variatus and green swordtail X. hellerii Both species are very common livebearers in the aquarium trade and aquacultured in Florida. Four experiments with both Xiphophorus species were conducted to test (1) the effect of eastern mosquitofish density on survival of noni ndigenous adult fish foll owing introduction, (2)
10 the effect of eastern mosquitofish density on a stage-structured population of nonindigenous fishes, (3) the effect of hab itat structural complexity (ste m density) on predation by eastern mosquitofish, and (4) the effect of habitat structural comple xity and eastern mosquitofish presence on the attack rates and habitat use pa tterns of adults of the nonindigenous species. Gape size did not limit predation in these experiments, with eastern mosquitofish frequently killing adults of the significantly larger nonindigenous species. The level of predation from eastern mosquitofish was density dependent with decreasing survival of the nonindigenous species with increasing eastern mosquitofish dens ity. Predation effects were more pronounced in swordtails compared to variable platyfish. Strong predation e ffects from eastern mosquitofish were observed on the juveniles of both species. Eastern mosquito fish presence, not density, was the significant factor in juvenile survival. St em density experiments did not show a pattern in survival for variable platyfish, but swordtails had decreased survival in high stem densities. Behavior data suggested that increases in interaction time, not shifts in eastern mosquitofish predation efficiency, were responsible for this pa ttern in survival. The results of the mesocosm experiments support the hypothesis th at eastern mosquitofish act as biotic resistance to the invasion of these nonindigenous species. Pred ation on adults is likely not limiting variable platyfish; however predation on adult swordtails may limit their invasion success, dependent on the densities of eastern mosquito fish and habitat complexity at the introduction site. Predation on juveniles is potentially a str ong factor in limiting invasions of both species. Furthermore, swordtails were found to be excluded from vege tation by mosquitofish in these experiments, which may facilitate other predators and increase the overa ll invasion resistance of the community.
11 CHAPTER 1 INTRODUCTION In recent decades, increasing attention has been given to anthr opogenic invasions where hum an activities homogenize floras and fauna s across the globe (McKinney and Lockwood 1999; Rosenzweig 2001; Rahel 2002; Olden et al. 2004). Despite this attention, fundamental questions remain concerning the in vasion process, particularly regarding the success or failure of particular types of invaders. Some research has focused on identifyi ng characteristics that predispose species to be successful invaders (Kolar and Lodge 2002; Marchetti et al. 2004; Ruesink 2005); however, the receiving environment cannot be ignored and invasion success may be determined by complex intera ctions between the biological attributes of the introduced species and the abiotic and biotic characteristic s of the invaded community (Kolar and Lodge 2002; Ruesink 2005). Hypothetically, biotic proce sses of a community, such as competition or predation, may reduce the success of invaders or prevent their establishment (Maron and Villa 2001; deRivera et al. 2005). Biotic resistance is one of several hypotheses proposed to explain the establishment of nonindigenous species (Laprieur et al. 2008). The concept of biotic resistance has a long history, dating back at least to Elton (1958) who hypothesized that more diverse communities would be more resistant to invasion. Numer ous studies have been conducted to test Eltons hypothesis, but results are varied and often c ontradictory, seemingly dependent on the types of organisms, communities, or geographic scales examined (Espi nosa-Garca et al. 2004; Fridley et al. 2007). Some investigations on plants a nd sessile invertebrates have provided support for the concept of biotic resistance, showing negative relationships between invasi on success and native species richness, observations ultimately considered to be the result of competitive processes. For example, in some terrestrial plant communities, higher native species richness has been shown to
12 reduce invasion success of nonindige nous species at small spatial scales (Fridley et al. 2007). Stachowicz et al. (2002) correla ted increasing native species richness of marine, sessile invertebrates to lower invasion success of similar nonindigenous species. More generally, Case (1990) showed that diverse native communities reduce invasion success, but this effect was dependent on the competitive intera ctions between i ndividual species. As an alternative to competition, predati on by native species may be a factor limiting success of mobile invaders in aquatic communities. Resident predators consume invading individuals or their offspring a nd may (1) directly eliminate intr oduced populations, (2) facilitate extirpation by other mechanisms, or (3) redu ce the distribution and a bundance of established invaders (Maron and Vila 2001; de Rivera et al. 2005; Ruesink 2007). For example, field and laboratory experiments i ndicate that blue crab Callinectes sapidus predation, possibly in combination with other factor s, limits the abundance and di stribution of the nonindigenous European crab Carcinus maenas along the U.S. Atlantic coast (deRivera et al. 2005). Similarly, laboratory feeding experiments provided evidence that blue crab predation also may control nonindigenous rapa whelk Rapana venosa populations in the Chesapeake Bay and associated estuaries (Harding 2003). Additional support co mes from studies on native and nonindigenous fishes. Baltz and Moyle (1993) investigated assemblages of California stream fishes and reported that native fishes showed the ability to resist invasion of in troduced fishes. They concluded that invasion resistan ce was due to both environmental and biotic factors, and suggested that predation is impor tant biotic factor. Subsequent ly, Harvey et al. (2004) conducted experiments in artificial laboratory streams a nd provided evidence of complex predator-prey interactions among native and nonindigenous Califor nia freshwater fish assemblages. They observed biotic resistance through multiple predator effects where predation risk from two native
13 fishes, prickly sculpin Cottus asper and coastrange sculpin C. aleuticus, combined with predation by the introduced Sacramento pikeminnow Ptychocheilus grandis caused greater mortality of nonindigenous speckled dace Rhinichthys osculus than predicted based on their separate effects. These results suggested that multiple predator effects were a factor in reducing the distribution of speckled dace in the drainages studied. Florida is one of most heavily invaded states in the United States and, similar to trends found throughout the world, the number of introduced fish species in the state has increased substantially over recent decades (Fuller et al. 1999; Nico and Fuller 1999). Of particular relevance to the concept of biol ogical resistance are the number a nd types of introduced fish taxa introduced into the state versus th e number of fish species that ha ve actually become established. In a recent summary report, the Florida Fish and Wildlife Conservation Commission (FWC) listed 34 reproducing exotic (= forei gn) freshwater fish species in the state, but determined that only 23 species to be established (FWC 2007). Esta blished, as defined by FWC and used in this study is a nonindigenous species unl ikely to be eliminated by ma n or natural causes and from which individuals can be regularly collected (FWC 2007). Other researchers have concluded that over 30 nonindigenous fishes ar e established or possibly establ ished in Florida; however the difference in numbers is based la rgely on how the term established is defined (Fuller et al. 1999; Nico and Fuller 1999). In addition to those nonindigenous fish sp ecies known to be reproducing, many other nonindigenous fish have been obs erved or collected in Florida waters, but without evidence of reproduction (Fuller et al. 1999; Ni co and Fuller 1999). Certain phys ical traits are apparently more common among those that form permanen t populations. For example, established nonindigenous fishes in Florida tend to have medium or large adult body size (Nico and Fuller
14 1999; Nico 2005; J. E. Hill, University of Flor ida, personal communication). Few small-bodied nonindigenous fish have reproduc ing populations in Florida op en waters and only one has become widely established. Of 43 small species (<15 cm maximum total length) reported by the USGS as introduced into the st ate (USGS 2007), only five of th ese have reproducing populations (USGS 2007; FWC 2007). Of the five reproducing species, only one species, the African jewelfish Hemichromis letourneuxi has spread beyond a localized range (Fuller et al. 1999; USGS 2007; FWC 2007). The low incidence of nonindigenous small-bodied fishes becoming established in Florida is interesting due to a combina tion of factors. For one, in a global meta-analysis, Ruesink (2005) reported that establishment success was higher for fishes in families with small body size, especially in poeciliids where > 90% of introductions were successful. However, most of the reported success is based on a few widely introdu ced species. Specifically in Florida, the ornamental aquaculture industry has been active since the 1930s, commonly raising small fishes in outdoor ponds. Hundreds of small fish specie s are traded in Florid a in the aquarium and ornamental aquaculture industrie s (Hill and Yanong 2002; C. V. Martinez, University of Florida, personal communication), which ar e major pathways for introduced fishes (Courtenay et al. 1974; Fuller et al. 1999; Semmens et al. 2004). Popul arity in these trades has been considered a proxy of propagule pressure (Duggan et al. 2006; Gertzen et al. 2008), the number of introductions and the number of individuals intr oduced, an important fa ctor in invasion success (Von Holle and Simberloff 2005; Hollebone and Hay 2007). Given the popularity of many of these organisms in Florida in both the aquari um trade and in aquaculture, the number of introductions of small-bodied, ornamental sp ecies is probably high (e.g., Fuller et al. 1999; USGS 2007). Moreover, the number of species and individuals introduced is likely under-
15 represented in the literature and nonindigenous speci es databases. In addition to the high number and diversity of small ornamental fishes potentially available, a nother factor favoring establishment of many small ornamental fish is the fact that much of Florida has a relatively mild, subtropical climate and a broad diversity of aquatic habitats. Consequently, it appears unlikely that abiotic factors such as temper ature, water chemistry, and habitat prevent establishment of all potential small-bodied fish in vaders in peninsular Fl orida, particularly in southern Florida, especially c onsidering the success of mediumand large-bodied fish invaders from the same source regions. Given that introductions of nonindigenous sma ll-bodied freshwater fishes are likely common and that these fish are introduced in to suitable habitats, why does Florida lack widespread or permanent populations of in troduced small-bodied freshwater fishes? Hypothetically, predation may lim it success of these invasions. Predation by large-bodied piscivorous species such as largemouth bass Micropterus salmoides or other centrarchids may be limiting invasion success of small-bodied fishes (J. E. Hill, University of Florida, personal communication). However, large-bodied predat ory fishes cannot access extremely shallow water habitats, and, when present, tend to be inefficien t predators in shallow or densely-vegetated areas where small-bodied species may take refuge (Savino and Stein 1982; Anderson 1984; DeVries 1990; Hayse and Wissing 1996). On the other hand, eastern mosquitofish Gambusia holbrooki Girard, a small-bodied native predator (maxim um adult size is 53 mm; Boschung and Mayden 2004), is nearly ubiquitous in aquatic systems in Florida and is often common in the same habitats apparently most suitable for small-bodied invaders. The western mosquitofish G. affinis (Baird and Girard) and th e eastern mosquitofish are common poeciliid fishes native to much of the United States (Page and Burr 1991). In addition
16 to having broad native distribut ions, both species have been introduced widely within and outside North America (Courtenay and Meffe 1989). The two taxa are morphologically and ecologically similar and both were treat ed as a single species under the name G. affinis and then later as subspecies. It was not until the late 1980s that genetic and morphological data indicated they were separate species (Wooten et al. 1988; Raushenberger 1989). Because the two were long grouped together as a single species, many references to G. affinis in the early literature actually pertain to G. holbrooki (Fuller et al. 1999). Because of widespread introduction, the precise natural geographic ranges of these two species are unclear (Rauchenbe rger 1989). The western mosquitofish is generally thought to be native to the southern and central United States from Alabama west to Texas and north into Illi nois whereas the eastern mo squitofish is native along much of the Atlantic slop e drainages from Florida to Ne w Jersey (Meffe and Snelson 1989; Rauchenberger 1989; Page and Burr 1991; Boschung and Mayden 2004). These species typically occur in small stream, wetland, and littoral zone habitat and have wide temperature and salinity tolerances (Meffe and Sn elson 1989; Pyke 2005). Mosquitofi sh are nearly ubiquitous in aquatic systems in Florida (Hoyer and Canfield 1994). In some littoral habitats, they are the most abundant fish species (Taylor et al. 2001; Baber and Babbitt 2004). Three lines of evidence suggest that mosquitofi sh may contribute to the biotic resistance of freshwater communities(1) effects on fishes wh ere mosquitofish are introduced, (2) effects on naturally co-occurring fishes, and (3) effects of mosquitofish as pests of aquaculture. Mosquitofish have been introduced for mos quito control throughout much of the world (Courtenay and Meffe 1989; Pyke 2005). In in troduced ranges, mosquitofish have been implicated in many negative ecological impacts. For example, in the American Southwest, transplanted populations ha ve been associated with declines and local extirpations of endemic,
17 small-bodied fishes (Meffe et al. 1983; Meffe 1985; Galat and Robertson 1992). Similar impacts have been cited in other regions including Eu rope and Australia (Courtenay and Meffe 1989; Howe et al. 1997; Caiola and de Sostoa 2005). Negative impacts resulting from mosquitofish introductions on small-bodied native fishes ma y be due to competition (Rincon et al. 2002); however, mesocosm studies have shown the stronge st factor is predation by adult mosquitofish on heterospecific juveniles (Meffe 1985; Mills et al. 2004; Laha and Mattingly 2007). In addition to research on the effects and in teractions of introduced mosquitofish, there have been studies investigating the interactions between native mosquitofish and co-occurring small native fishes. Experiments conducted in mesocosms demonstrated that mosquitofish preyed heavily on the ju veniles of other species. (Belk a nd Lydeard 1994; Schaefer et al. 1994; Taylor et al. 2001). Based on experiments with fish from the Florida Everglades, Taylor et al. (2001) concluded that predation by adult mosquito fish on juvenile fish, including conspecifics, was the strongest effect observed among a wide range of intraand interspecific age-structured interactions examined. Lastly, mosquitofish are pests in aquaculture ponds, reducing the production of small-bodied ornamental fishes (J. E. Hill and C. A. Watson, University of Florida, personal communication; see also Hill and Watson 2007). The effectiveness of mosquitofish as pisciv ores is attributed to both behavioral and morphological characteristics. Behaviorally, they tend to be more aggressive than most smallbodied freshwater fishes, including other poeciliids (Meffe and Snelson 1989). Morphologically, they have strong conical teeth and short gut length, adaptations indicative of a predatory fish (Meffe et al.1983). Studies have shown th at gape size limits consumption of prey by mosquitofish (Taylor et al. 2001). However, they are known to harass, bite, and severely damage fins of fish of similar size to themselves (M effe 1985; Laha and Mattingly 2007). Furthermore,
18 there is evidence that mosquitofish are not st rictly gape-limited and can disable and consume prey slightly larger than themselves. This is achieved by first damaging the fins of the prey and immobilizing them before consumption (Baber and Babbitt 2004; J. E. Hill, University of Florida, personal communication). Given their aggressive nature and abundance in most stream, wetland, and littoral habitats, mosquitofish predati on might be a factor limiting some freshwater fish invasions. The objective of this study wa s to test the hypothesis that predation by native resident eastern mosquitofish is contributing to the bi otic resistance to i nvasions of small-bodied freshwater fishes in Florida. A series of meso cosm experiments were conducted to determine the effect of mosquitofish predati on on nonindigenous, small-bodied freshwater fishes of the genus Xiphophorus These nonindigenous fishes are similar in general body shape, but with potentially different anti-predator behavior al characteristics. Four me socosm experiments with two nonindigenous poeciliids were conducted to test (1 ) the effect of mosqu itofish density on adult mortality of nonindigenous fish following an in troduction event, (2) effect of mosquitofish density on a stage-structured populat ion, (3) effect of habitat stru ctural complexity on predation effects of mosquitofish, and (4) beha vioral shifts in habitat use patter ns as a result of the threat of mosquitofish predation
19 CHAPTER 2 METHODS Study Species Two nonindigenous poeciliids of sim ilar m orphology, variable platyfish Xiphophorus variatus (Meek) and a variant of the green swordtail X. hellerii Heckel, were tested separately in all mesocosm studies. Both are common in orna mental aquaculture production in Florida (Hill and Yanong 2002) and in the aquarium trade. The va riable platyfish is native to drainages along portions of the Atlantic slope of eastern Mexico (Rosen and Bailey 1963; Rosen 1979; Page and Burr 1991). The green swordtail has a larger nati ve distribution, ranging from Veracruz, Mexico south into Honduras (Rosen and Kallman 1969; Page and Burr 1991; Field and Thomerson 1997). The two species typically inhabit springs streams, canals, and small ponds (Page and Burr 1991). Variable platyfish and green swordtails are gene ralized omnivores, feeding on a variety of small invertebrates, plant material, and detritus (M ills and Vevers 1989). Maximum total length (TL) for variable platyfish is approximately 70 mm (Page and Burr 1991). Green swordtail maximum TL is approxim ately 80 mm with males reaching greater total lengths due to sexual dimorphism of the caudal fin (Page and Burr 1991). The variable platyfish has lo calized populations in Alachua County, Florida in two small streams (Fuller et al. 1999; J.E. Hill and C.E. Cichra, University of Florida, personal communication) and has been reported in other parts of the state (Nico 2007) It is one of two species on Floridas clean list, meaning that in troducing the species is not illegal (Hill 2006). This regulation is in effect because in the 1970s, variable platyfish was considered for use as bait for the black crappie Pomoxis nigromaculatus fishery (Ogilvie et al. 1974). Green swordtails and color varieties of this sp ecies have been collected in Florida and localized, ephemeral populations have been reported in various regions of the Florida Peninsula (Fuller et al. 1999).
20 In addition, numerous unreporte d introduction events of these two species are likely. No permanent, wide-spread established populations of either species occur in Florida. Mosquitofish used in all experiments were native Gambusia holbrooki These came from two sources: individuals collect ed from a detention pond of th e UF/IFAS Tropical Aquaculture Laboratory (TAL), Ruskin, Florida and others obtained from a commerc ial producer in Venus, Florida. To confirm correct species identification, the gonopodial stru cture of several male specimens from both sources were examined (Rauchenberger 1989).Variable platyfish were obtained from a commercial produc er and held in a pond at TAL until needed for experiments. Individuals used were the sunset variety, a common commercial morph with a golden color with red caudal peduncle and caudal fin. It is the same variety loca lly established in Gainesville, Florida (J.E. Hill and C.E. Cichra, University of Florida, personal communication). There is little sexual dimorphism in this species. Green swordtails used in expe riments were the velvet wag variety (red body with black caudal fin) an d collected from populations maintained in aquaculture ponds at TAL. This variety was possibly created from hybridization with other species of Xiphophorus (Angus 1989) and to reduce possible confusion the term swordtail will be used in the remainder of this paper instea d of the accepted common name, green swordtail. This species is sexually dimorphic with males having a sword which is an extended portion of the lower part of the caudal fin. The sex ratio used in the experiments was not manipulated and was representative of the cultu re ponds with the majority of individuals being females. Subsamples yielded a population of 32 3.5% males. Predation effects by mosquitofish on swordtails were not analyzed separately by sex. For all experiments the adults of the nonindigeno us species were size selected so that all individuals were approximately 35-50 mm total length (TL) for variable platyfish and 45-60 mm
21 TL for swordtail. Fish in the experimental units were not directly measured and weighed to reduce handling that might result in mortality. Prio r to start of each experi ment, individuals from the captive populations (n = 50 eastern mosquito fish and 30 of each nonindigenous species) were removed and their length and weights measured to provide estimates for the experimental populations. Based on those measurements, eastern mosquitofish mean sizes were: TL = 28 5 mm and weight (W) = 0.27 0.16 g; variable plat yfish mean values were: TL = 40 4 mm and W = 1.20 g 0.35; and swordtail measurements we re: TL (not including sword on males) = 50 3 mm and W = 1.73 0.41 g. The two nonindigenous sp ecies were significan tly longer than the eastern mosquitofish (two-sample t-test; variab le platyfish t0.05, 238 = -20.30, p < 0.0001, swordtails t0.05, 238 = -40.41, p < 0.0001) and heavie r (variable platyfish t0.05, 238 = -25.67, p < 0.0001, swordtails t0.05, 238 = -37.83, p < 0.0001) across all experiments. Mesocosm Overview A series of m esocosm experiments was conducte d to investigate possi ble predatory effects of mosquitofish on the study species (Table 21). All experiments were carried out in a greenhouse located at TAL. Experimental units we re black, oval polyethylen e tanks with an area of approximately 1.2 m2 at the base and with a water surface area of 1.4 m2. Water depth was maintained at 0.23 m, which was chosen to be sim ilar to that in small streams and shallow littoral zones commonly inhabited by mosquitofish. Th e tanks were set up on a flow-through system using aerated well water (2 L/min). Each tank was aerated by a single air stone and had a single, 44-cm standpipe that protruded 7 cm from the si de of the tank and was 5 cm from the bottom (Figure 2-1A). Structural complexity was provided in each tank by artificial vege tation consisting of multiple strips of black plastic sheeting cut to be approximately 2 cm x 40 cm each and glued to a rectangular plastic lighting grate that was 91.4 x 61.0 cm and covered 49% of the tank bottom
22 (cf. Savino and Stein 1982) (Figure 2-1A). Artif icial vegetation density was standardized for the first two experiments (216 stems/m2) and varied in the third and fourth. Water quality was checked for six tanks (two of each treatment) randomly chosen at the start of every experiment and once each week during the experiment. For experiments that were less than one week in duration, water was tested at the start of the expe riment and then on the third day. Parameters varied lit tle during all of the experiments. Water chemistry parameters were measured using a Hach fish farmers water quality kit (Hach Company, Loveland, CO) and were (mean SD): hardness 485.7 47.1 ppm, alkalinity 177.8 16.4 ppm, nitrites 0.13 0.23 ppm, and unionized ammonia nitrogen undetec table. A YSI handheld meter (YSI Inc., Yellow Springs, OH) was used to measure temp erature (27.0 0.23 C), pH (8.0 0), dissolved oxygen (6.68 0.57 ppm), and salinity (0.4 0 ppt). Experiment 1: Adult Introduction Experim ent 1 was designed to determine effects of predation by eastern mosquitofish on introduced adult variable platyfish and swordtai ls at three different eastern mosquitofish densities (Table 2-1). This experiment mimick ed an introduction event that might occur in nature whereby marketable-sized ornamental fi sh are introduced into an environment already containing a population of eastern mosquitofish. Mesocosm studies conducted by others ha ve shown that the number of eastern mosquitofish present is closely related to the st rength of this species pr edatory effects (Belk and Lydeard 1994; Taylor et al. 2001; Mills et al. 2004). In th e present study, eastern mosquitofish were added to mesocosms to create three di fferent density treatment s: 21 (= low), 43 (= medium), and 86 (= high) eastern mosquitofish/m2 water surface area (numbers of fish in the treatments were 30, 60 or 120 per tank) (Table 2-1) The numbers of eastern mosquitofish used in the different treatments are within the range of observed densities with in natural systems in
23 Florida (Trexler et al. 2005: suppl emental material). Five replicates of each treatment were run for variable platyfish and then fi ve were run using swordtails. In most natural Florida systems, eastern mos quitofish would already be present in a water body receiving released or escaped fish. Therefore, to more closely imitate natural field conditions, eastern mosquitofish were added to tanks 4 days prior to the other poeciliids and allowed to acclimate to mesocosm conditions (T able 2-1). During the acclimation period and throughout the experiment, fish were fed 5% body weight per day of a commercial fish feed (Purina 33% Tropical Fish Chow, Purina Mills, St Louis, MO) and feed weights were adjusted upward upon the addition of the nonindigenous species. This feeding protocol was repeated in all subsequent experiments. Any dead eastern mosquitofish found in the tanks during the acclimation period were removed and repl aced to maintain target density. After the 4-day acclimation period, 10 adult variable platyfish or swordtails were introduced into the mesocosms (density = about 7 fish/m2 surface area). Fish were observed and mortalities removed daily. The trials with variab le platyfish lasted 11 da ys, whereas swordtails trials lasted only 5 days due to higher morta lity rates (Table 2-1). On the final day, the vegetation was removed and all fish were removed from each tank and counted. Fin damage indices have been used in other experimental studies to describe sublethal damage caused by mosquitofish (Meffe 1985; Galat and Robertson 1992). Therefore, all surviving adult variable platyfish and swordtails were insp ected for damage to the caudal fin. Individuals with damage were eu thanized with an overdose of tricaine methanesulfonate (MS222) and then preserved in a 10% formalin soluti on for later analysis. Indi vidual fish were given a score based on amount of damage to the caudal fin with the values of: 0 for no damage, 1 for
24 moderate damage (less than 50% of fin area), and 2 for severe da mage (greater than 50% of the fin). Experiment 2: Stage-Structured Population This experim ent was designed to test the effect of mosquitofish densities on survival rates in stage-structured populations of nonindigenous fish. Adult and j uvenile variable platyfish and swordtails were stocked into mesocosms to repr esent a population of adults that survived the introduction event and are subsequently produ cing offspring in the presence of eastern mosquitofish. Selective predation by mosquitofish on juvenile fish is frequently cited as a major impact of mosquitofish on other fish populati ons, possibly preventing recruitment (Meffe 1985; Belk and Lydeard 1994; Schaefer et al. 1994; Taylor et al. 2001; Mills et al. 2004; Laha and Mattingly 2007). Eastern mosquitofish and the nonindigenous species were added to the experimental units at the same time (Table 2-1). This was done to mimic co-existing populations of eastern mosquitofish and the two introduced species, as well as prevent an artificially high predation effect on the introduced juven iles when initially added to the tanks. Three eastern mosquitofish densities of 0, 21, and 86 eastern mosquitofish/m2 were conducted to represent a control, low, and high density (Table 2-1). Five replicates of each treatment were run with both vari able platyfish and then swordta ils. A treatment without eastern mosquitofish was used to control for possible canni balism of the juveniles by conspecific adults. Adults of the nonindigenous species were selected with the same minimum sizes as with all other experiments. The juveniles used were the smallest fish that could be acq uired to represent newly recruited juveniles. Ten adult a nd ten juveniles of either variable platyfish or swordtails were added to each experimental unit. Subsamples of 30 juvenile fish of each species were used to estimate lengths and weights of the donor popul ation. Mean variable platyfish juvenile
25 measurements were: TL = 10 2 mm and W = 0.02 0.01 g. Swordtail juvenile mean measurements were: TL = 14 1 mm and W = 0.07 0.01 g. As in Experiment 1, fish were observed a nd all mortalities were removed daily. The experiment was run for 4 days for both variable pl atyfish and swordtails (T able 2-1). At the end of the 4 days, all fish were removed and counted. Surviving adults were inspected for caudal fin damage. Although this was a short-term experiment, evidence of reproducti on and appearance of offspring were noted and analy zed as a response to mosquitofi sh treatment. Neonates were noticeably smaller and could be readily distingu ished from the stocked juveniles. Female poeciliids are capable of bear ing live young nearly year round (Meffe and Snelson 1989) and gravid females were commonly included as part of the 10 introduced adult fish in all experiments. However, due to the short-term nature of this experiment and difficulty in differentiating reproductive stage, no attempt wa s made to select females of equal reproductive potential (e.g., embryo number and developmental stag e). Consequently, the analysis of recruits should be considered preliminary. Experiment 3: Effects of Structural Complexity Aquatic sys tems vary in habitat complexity and this variation may have important implications relative to pred ator-prey dynamics (Savino a nd Stein 1989; Hayse and Wissing 1996). Multiple experimental studies have shown decreased predation efficiency in fishes with increased habitat complexity (Savino and Stein 1982; Anderson 1984; Eklov and Hamrin 1989; Savino and Stein 1989; Hayse and Wissing 1996; Eklov and VanKooten 2001; Hickey and Kohler 2004). However, few studi es have investigated if this relationship also holds for smallbodied predatory fishes (but s ee Baber and Babbitt 2004). Experiment 3 was designed to test the influence of habitat structural complexity on predation rates of eastern mosquitofish on
26 introduced poeciliids. This experiment was simila r to Experiment 1 with adults being introduced into an already acclimated eastern mosquitofish population. Eastern mosquitofish density was 86 fish/m2 (the high density treatment level in previous experiments) and the vegetation densities were varied across three trea tment densities of 72, 216, and 645 stems/m2 across the area of the grate (number of stems was either 40, 120, or 360) at the bottom of the tanks (Table 2-1; Figure 2-1B). Thes e densities were chosen to be within the range of native plant densities and othe r experimental work on habitat complexity and fish predation (e.g., Savino and Stein 1982; Anderson 1984; Hayse and Wissing 1996). Five replicates of each stem density were conducted with vari able platyfish and then swordtails. Eastern mosquitofish were acclimated for 4 days prior to initiation of experiment. During the period of acclimation any dead eastern mosquitofish discovered in mesocosms were replaced to achieve target densities when the experiment started. The e xperiment began when either 10 variable platyfish or sw ordtails were added to mesocosm ta nk containing eastern mosquitofish. Run times for this experiment were set to equal those of Experiment 1: 11 days for variable platyfish and 5 days for swordtail (Table 2-1) Fishes in mesocosms were observed daily and mortalities removed. At the end of the experime ntal periods, all fish were removed and counted and any remaining nonindigenous fish were inspected for caudal fin damage. Experiment 4: Behavioral Measurements Predators can alter the behavior of prey in a variety of ways. Experim ental work with fishes has shown that apparent predation threat results in prey species changing their behavior (Savino and Stein 1982; Holker et al. 2007) and habitat use (W erner et al. 1983; Savino and Stein 1989; Hayse and Wissing 1996; Holker et al. 2007). During the preliminary stages of Experiments 1, 2 and 3 it was observed that eastern mosquitofish were closely associated with artificial vegetation, rarely movi ng outside areas without stem cove r, and that variable platyfish
27 and swordtails typically were found only outside the artificial c over, apparently being excluded from cover due to eastern mosquitofish aggressio n. Furthermore, the two species also appeared to respond differently (behavior and habitat use pa tterns) to the presence or absence of eastern mosquitofish as well as to stem density variations Experiment 4 was designed to test behavioral changes more quantitatively based on those preliminary observations. Two main hypotheses were tested in this experiment: (1) introduced poe ciliids would encounter di fferential attack rates and utilize habitat in the tank differently across va ried stem densities in the presence of eastern mosquitofish; and (2) at a constant stem densit y, habitat use patterns of the introduced poeciliids would differ depending on the presence versus absence of eastern mos quitofish. Analysis of this experiment was separated into two groups of treatments. Analysis of Differing Stem Densities (4A) This portion of Experiment 4 tested the effect of habitat complexity on eastern mosquitofish attack rates and habitat use in the two introduced species. Th e treatm ents in this experiment were 72, 216, and 645 stems/m2 w ith the high density (86/m2) of eastern mosquitofish (Table 2-1). Analysis of Eastern Mosquitofish Presence (4B) This second set of treatments in Experim ent 4 tested the effect of eastern mosquitofish presence on attack rate and ha bitat use across a constant stem density. The treatment groups included the medium stem density treatment (216 stems/m2) with 86 mosquitofish/m2, which was also analyzed in Experiment 4a (treatment referre d to as MF in the analysis of Experiment 4b), and two different control treatments at the same stem density. One control had no mosquitofish or any other fish stocked previous to the introduction (treatment referred to as NoMF throughout) and the second control had zebra danios Danio rerio (Hamilton) stocked at 86 fish/m2 (treatment referred to as Danios throughout) (Table 2-1). Th e Danios control was used to control for any
28 density effects on behavior by the mosquitofish not linked to aggres sion. The zebra danio, a small cyprinid fish from Asia that is a common ornamental fish and research animal, was chosen to serve as a control for the experiment because it is similar in size, appearance, and activity level to eastern mosquitofish (personal observa tions). However, in contrast to eastern mosquitofish, the zebra danio is generally not aggressive towards other small fish (Mills and Vevers 1989). Except for differences described above, the re maining protocols for both experiments were the same. Eastern mosquitofish and zebra danios were given a 4 day acclimation period in their respective mesocosms prior to in troduction of the study species. The experiment began when 10 adult variable platyfish or sw ordtails were introduced into the mesocosm tanks, including the NoMF treatment which had no other fish previously stocked. There were th ree replicates of each of the treatments, for a total of 15 tanks used with each study species (Table 2-1). Behavior was measured in two ways. The fi rst was a focal animal sampling protocol (Martin and Bateson 1993) whereby individuals of the introduced species were observed over time to obtain a quantitative sample of the number a ttacks an introduced fish sustains at different mosquitofish or stem density treatments (Table 2-1). An attack was defined as in Laha and Mattingly (2007) as the sum of all aggressive movements, including chases and nips (bites to the caudal fin area). Conspecific att acks of the introduced species we re also recorded and, because there were few attacks (maximum of 2 conspeci fic attacks per observation period), these were included as part of the total atta cks observed. Observations were made on an arbitrarily chosen individual that was followed for a maximum of 5 minutes or until lost after which another fish was chosen. The total focal animal sa mpling period was 30 minutes for each tank.
29 The second behavioral measure was achieved with scan sampling (Martin and Bateson 1993). This protocol was used to quantify the hab itat use patterns of the introduced species as a response to mosquitofish presence or stem dens ity. In this procedure, the position of all 10 introduced fish was described at regular intervals throughout a sa mpling timeframe. In this instance the recorded positions were either in the vegetation, describing fish in or above the simulated vegetation mat, or outside the vegetati on, to describe those fish outside of the area defined (Table 1). If a fish was on the border, the position was assigned to the habitat containing > 50% of the body length based on visual estimation. Both types of behavior observ ations were made twice duri ng this experiment. A focal animal sampling period of 30 minutes began immediately when the 10 study fish were introduced. This was followed up with a scan sa mpling period about 4 hr later on the same day. During the scan sampling period, the positions of the fish were noted every 5 minutes for 30 minutes giving 7 total position observations made per sampling period. The focal animal sampling and scan sampling periods were repeat ed on Day 3 to allow for observation after a period of acclimation for the nonindigenous stud y species, where habitat use patterns may be more indicative of preference and anti-predato ry behavior may be mo re apparent. In all behavioral observations, recordi ng began after allowing 5 minutes fo r the fish to acclimate to the observer. After the Day 3 observations, all fish were removed, counted, and scored for extent of caudal fin damage. Data Analysis Percent su rvival of the introduced species in each experiment was analyzed using one-way analysis of variance (ANOVA; Proc GLM; SAS Institute, Cary, NC). A type-I error rate of = 0.05 was used for all ANOVAs. Following a si gnificant ANOVA, Tukeys post-hoc pairwise tests were used to determine which treatments differed. Data were arcsine, square root-
30 transformed prior to analysis. In all of the experiments, the assumptions of ANOVA were checked using Levenes test for homogeneity (Pro c GLM) as well as the Shapiro-Wilk test for normality on the residuals of th e ANOVA model (Proc UNIVARIATE). The independent variable in Experiments 1 and 2 was eastern mosquitofish density. For Experiment 2, the juvenile and adult percent survivals were analyzed independently. The independent variable in Experiment 3 was stem density. In Experiment 2, the additional recruit count data were analyzed using a one-way ANOVA model (Proc GLM). Count data were square-root transformed prior to analysis. The Experiment 4 dependent variable in the focal animal observations was total attacks (eastern mosquitofish + conspecific) versus e ither stem density (Experiment 4a) or eastern mosquitofish treatment (MF, NoMF, or Danios; E xperiment 4b). Total attack count data were analysed using the non-parametric Kruskal-Wallis (K-W) rank-sum test (Proc NPAR1WAY) because the data (both raw and square-root tran sformed) violated the assumptions of normality and homogeneity of variance required in th e ANOVA model. Dunns non-parametric multiple comparison procedure (Hollander and Wolfe 19 73) was used post-hoc to determine which treatment groups differed significantly. Because of the conservative nature of Dunns procedure, a type-I error rate of = 0.15 was used (Hollander and Wolfe 1973). For Experiment 4, the dependent variable in the scan sampling data was the proportion of introduced fish in the artificial vegetation. In Experiment 4a, th e independent variable was stem density. The Experiment 4b independent variable was treatment (MF, NoMF, or Danios). The proportion data were arcsine, square-root transformed and an alyzed using an ANOVA model (Proc GLM). The least-squares means procedur e (Proc GLM) was used post-hoc to determine which treatments differed.
31 Fin damage scores were analyzed using the non-parametric K-W rank-sum test (Proc NPAR1WAY). The independent vari able in all experiments was th e fin damage scores assigned to the surviving adults of th e study species. Experiment 1 and 2 dependent variables were mosquitofish density, Experiment 3 and 4a de pendent was stem density, and treatment group (MF, NoMF, or Danios) was the dependent in Experiment 4b. Dunns non-parametric multiple comparison procedure (Hollander and Wolfe 19 73) was used post-hoc to determine which treatment groups differed significantly.
32Table 2-1. Overview of four different meso cosm experiments testing the effects of eas tern mosquitofish (MF) predation on two nonindigenous poeciliids. Expt.a Independent variable Treatment levels Replicates Introduced population MFb added Variable platyfish trial duration (days) Swordtail trial duration (days) Response variables 1 Eastern mosquitofish density 21, 43, 86 MF/m2 5 10 Adults 4 Days prior 11 5 Adult survival 2 Eastern mosquitofish density 0, 21, 86 MF/m2 5 10 Adults, 10 Juveniles Same time 4 4 Adult survival, juvenile survival, recruitment 3 Stem density 72, 216, 645 Stems/m2 5 10 Adults 4 Days prior 11 5 Adult survival 4a Stem density 72, 216, 645 Stems/m2 3 10 Adults 4 Days prior 3 3 Number of attacks, proportion of fish in vegetation 4b Eastern mosquitofish presence 0, 86 MF/m2, or 86 zebra danios/m2 3 10 Adults 4 Days prior 3 3 Number of attacks, proportion of fish in vegetation a Expt. = Experiment b MF = eastern mosquitofish
33 A B Figure 2-1. Pictures of experime ntal tank and vegeta tion. A) Tank shown with mesocosm set up with a vegetation grate at th e medium density (216 stems/m2). B) The three stem density grates with vegetation shown as used in Experiments 3 and 4a.
34 CHAPTER 3 RESULTS Experiment 1: Adult Introduction Eastern m osquitofish harassed and nipped th e fins of the adults of both nonindigenous species in all tanks, causing mortalities in all trea tments except for variable platyfish in the low density treatment. Survival of the nonindigenous species decreased with increasing mosquitofish density (variable platyfish ANOVA, F2, 12 = 13.96; p = 0.0007; swordtail ANOVA, F2, 12 = 6.77; p = 0.018) (Figure 3-1). The high density treatmen t reduced variable plat yfish survival by about 28% (SE = 4.9). However, there was little mortal ity in the low and medium density treatments with survival near 100% (Figure 3-1). Mean survival of swordtails in the high eastern mosquitofish density treatment (30%, SE = 6.3) wa s only about half of th at observed in the low density treatment (64%, SE = 4.0) (Figure 3-1). Damage to the caudal fin was noted in surviv ing individuals of va riable platyfish and swordtails. However, no statisti cal differences were found in the extent of fin damage among the different treatments for either species. Most surviving variable platyfish ( 96%) had no damage, but a few individuals had moderate damage (Table 3-1). Similarly, most swordtails (> 90%) had no caudal fin damage, but some individu als were observed with moderate to severe damage (Table 3-2). In contra st, all study fish found dead in mesocosm tanks and presumably killed by eastern mosquitofish exhibited substantial damage to the caudal fin and caudal peduncle. Similar to what was observed in each of the other experiments, the fin damage of dead fish was severe and would have been scored as 2 (i.e., > 50% of caudal fi n missing) if they had survived to end of experiment.
35 Experiment 2: Stage-Structured Population Results in dicated that eastern mosquitofish exerted variable mort ality effects on adult variable platyfish and sw ordtails, but had strong ne gative effects on juvenile survival of the two introduced species. Eastern mosquitofish density had little effect on adult survival of variable platyfish (ANOVA; F2, 12 = 1.00; p = 0.3966), with mean survival near 100% across all treatments (Figure 3-2). Conve rsely, eastern mosquitofish de nsity had a significant effect on adult swordtails (ANOVA; F2, 12 = 32.87; p < 0.0001) with survival reduced to 86% (SE = 2.4) in the low density and 58% (SE = 5.8) in the high dens ity treatments relative to the control (Figure 3-2). In the absence of mosqu itofish, no variable platyfish died and swordtail survival was 98% (SE = 2). Eastern mosquitofish density had a signifi cant negative effect on juvenile survival (variable platyfish ANOVA; F2, 12 = 71.92; p < 0.0001; swordtail ANOVA; F2, 12 = 34.08; p < 0.0001) (Figure 3-2). Indeed, mosquitofish presence was the critical factor with juvenile survival in the controls being 3 to 7 times higher than in the low density treatment. Juveniles were eliminated in each replicate of the high dens ity treatment, with one exception (i.e., a single surviving juvenile swordtail in one tank). All juvenile variable pl atyfish survived in the control treatment, evidence that adult variable platyfish did not canni balize their young (Figure 3-2). Swordtail adults exhibited a small cannibalistic effect with juve nile survival of 96% (SE = 4.0) in the controls. In contrast to the interactions observed between eastern mosquitofish and adult nonindigenous species, the bodies of dead juveni les were not found and it was concluded that eastern mosquitofish had killed and completely consumed all juvenile mortalities in this experiment. Variable platyfish recruitment in this e xperiment varied signi ficantly with eastern mosquitofish density (ANOVA; F2, 12 = 6.45; p = 0.0125). Additional recruits were noted in four
36 of five control replicates for the variable platyfish experiment averaging an additional 6.8 fish (SE = 4.2) per mesocosm. In contrast, neonate va riable platyfish were not observed in the two mosquitofish treatments. There was no swordtail recruitment observed in the control or eastern mosquitofish treatments and no st atistical analysis was performed. Surviving adults of both study species were found to sustain caudal fin damage in this experiment. Variable platyfish recovered from the high density eastern mosquitofish treatment had greater amounts of damage to the caudal fin than in either the control or low density treatments (Table 3-1). Similarly, swordtail ad ults were shown to have significantly higher fin damage in the high density than in the control treatment (Table 3-2). Experiment 3: Effects of Structural Complexity Adult m ortalities of both study species were observed across the varying stem density treatments. Stem density had no significant e ffect on variable platyf ish survival (ANOVA; F2, 12 = 1.79; p = 0.208) across the range of stem density investigated in this study (Figure 3-3). Nevertheless, analysis of the va riable platyfish data indicated a low power (30%; Proc POWER) for this ANOVA and the test was only able to detect differences between means of 53%. Conversely, swordtail survival in the low and medium stem dens ity treatments was about two times higher than in the high st em density treatment (ANOVA; F2, 12 = 5.86; p = 0.0168) (Figure 3-3). None of the surviving variable platyfish e xhibited fin damage in this experiment, but swordtails did have some caudal fin damage (Tab les 3-1, 3-2). However, there was no significant difference among the different stem densities in le vel of fin damage for the adult swordtails in this experiment (Table 3-2).
37 Experiment 4: Behavioral Measurements Analysis of Differing Stem Density (4A) Day 1 As observed qualitatively in Experim ents 13, eastern mosquitofish commonly attacked individuals of both introduced species. For variable platyf ish, there were no statistical differences in number of attacks among stem densities on the day of introduction (K-W; 2 = 0.2667, p = 0.8725) (Figure 3-4). Mean numbers of attacks were between 12 and 14 for the 30 minutes of observation, but the data were highly variable (ranging from 2 to 35). Indeed, the standard error was greater in magnitude than th e mean in the medium density treatment (Figure 3-4). Similarly, total attack results were not significantly different among stem densities for swordtails (K-W; 2 = 1.681, p = 0.4316). Mean number of a ttacks were 12 to 25 and ranged from 2 to 45. These results also exhibited high variability across the treatment range (Figure 34). Although differences were not detected in th e mean number of attacks, the proportion of nonindigenous fish in artificial vegetation in the presence of a high density of eastern mosquitofish differed significantly by stem density for both variable platyfish and swordtails on the day of introduction (var iable platyfish ANOVA; F2, 6 = 5.74; p = 0.0404; swordtail ANOVA; F2, 6 = 5.84; p = 0.0391) (Figure 3-5). In general, the highest propor tion of fish (about 0.5) in vegetated areas versus outside vegetated areas was found at the highest stem density. For variable platyfish, data indicated an increasing number of fish in the vegetation with increasing stem density with about 1.7 times the number of fish occupying the vegetation in the highest compared to the lowest stem density treatment (F igure 3-5). The proportion of swordtails in the vegetation at the high stem density was significantly greater (by a factor of 3 to 4.5) than in the low and medium density treatments.
38 Day 3 After two days of exposure to eastern m osquitofish, attacks on the study species continued. The only mortality during the experiment was a single dead variable platyfish found in the medium stem density treatment. Mean attack numbers ranged from 15 to 17 across the three stem densities for variable platyf ish (the range of attacks was fr om 8 to 19 across all replicates) and mean attacks were from 6 to 11 for swordtails (across all replicates the range was from 1 to 13 attacks) (Figure 3-4). Sim ilarly to Day 1, there were no significant differences across stem densities for either species (variable platyfish K-W; 2 = 2.529, p = 0.2823; swordtail K-W; 2 = 2.056, p = 0.3576). Unlike Day 1 results, there was no significant difference in use of vegetation by variable platyfish across stem densities (ANOVA; F2, 6 = 1.04; p = 0.4080) (Figure 3-5). However, analysis showed that the vari able platyfish model had low power (29%) and a detectable difference of 0.31 in proportions of fish in the vegetation among treatments. On the other hand, swordtail results were similar to Day 1, with the proportion of swordtails in the vegetation at the high stem density significantly greater than in the low and medium stem densities (ANOVA; F2, 6 = 6.75; p = 0.0292) (Figure 3-5). Fin damage in both study species was noted from all three stem density treatments. Most variable platyfish recovered show ed no damage (> 76%), but some individuals had moderate to severe damage. There was no significant pattern in fin damage by stem density for this species (Table 3-1). Similarly, swordtails with moderate and severe damage were found in all treatments, but a small percentage in the high stem density treatment was scored as severe (score = 2). However, there was no significant difference in sw ordtail fin damage among the three treatments (Table 3-2)
39 Analysis of Eastern Mosquitofish Presence (4B) Day 1 Mosquitofish were again observed attacki ng both introduced species. However, there were no con specific attacks observed in either the Danios or NoMF treatments or any attacks observed by the zebra danios on either study specie s. The only attacks noted were in the MF treatment and the analysis showed that the num ber of attacks was significantly different among treatments for both species (variable platyfish K-W; 2= 7.623, p = 0.0.0221; swordtail K-W; 2 = 7.624, p = 0.0.0220). Mean total attacks on variable platyfish observed in the MF treatment was 11 (SE = 18), with 25 (SE = 20) in the swordtail trials (Figure 3-6). In the MF treatments, variability was high, with standard error grea ter than or near the value of the mean. The scan sampling data showed that hab itat use patterns differed by treatment with variable platyfish, with more i ndividuals using vegetation habita t in the presence of eastern mosquitofish (ANOVA; F2, 6 = 6.98; p = 0.0271). The mean proportion of variable platyfish in the vegetation in the NoMF and Danios treatments were two to five times less than in the MF treatment (Figure 3-7). On the other hand, swordt ails showed no significant difference in habitat use among treatments (ANOVA; F2, 6 = 1.78; p = 0.2474). Day 3 Attacks were again noted on Day 3 of observation am ong these treatments for both species. The single variable platyfish mortality in the medium density treatment was reported in the results for Experiment 4a (this treatment was shared by 4a and 4b) there were no other mortalities in the other two treatments for either species. The pattern of attacks by treatment was the same as in Day 1, with all observed attacks occurring in the MF treatment. Attacks were therefore significantly different by treatment (variable platyfish K-W; 2 = 7.714, p = 0.0211; swordtail K-W 2 = 7.624, p = 0.0220) (Figure 3-6). The mean number of attacks on both
40 variable platyfish and sw ordtails was 11 (variable platyfish SE = 5; swordtail SE = 2) in the MF treatment. Again, attacks were not observed in the NoMF and Danios treatments (Figure 3-6). The results of the scan sampling data were si milar to Day 1 results, in that one species showed a response to treatment in habitat use whereas the other did no t. On this day of observation, variable platyfish did not show a sign ificant difference (ANOVA; F2, 6 = 1.04; p = 0.4080) whereas the proportion of swordtails in the vegetation was significantly different by treatment (ANOVA; F2, 6 = 6.99; p = 0.0271) (Figure 3-7). Th e proportion of swordtails using the vegetated area in the tanks was three to five times higher in the Danios treatment compared to the NoMF and MF treatments (Figure 3-7). Variable platyfish and swordtails showed simila r patterns of fin damage in this experiment with significantly greater amounts of fin damage in the MF treatment than either the NoMF or Danios treatments (Table 3-1, 3-2). The only surv iving variable platyfish recovered with either moderate or severe fin damage were in the MF treatment (Table 3-1). In addition to fin damage in the MF treatments, between 3 and 10% of surviving swordtails had moderate fin damage in the NoMF and Danios treatments. None of the su rviving swordtails had severe damage in any of the treatments (Table 3-2).
41 Table 3-1. Summary of caudal fin damage results for surviving adult variable platyfish in all experiments and treatments. Experiment Treatmenta Scoreb 0 (%) Score 1 (%)Score 2 (%) K-W ( 2 )c P-value / Dunns groupingd 1 2.25 0.325 21 100.0 0.0 0.0 A 43 96.0 4.0 0.0 A 86 100.0 0.0 0.0 A 2 17.64 0.001* 0 100.0 0.0 0.0 A 21 100.0 0.0 0.0 A 86 91.4 4.2 4.4 B 3 1.133 0.568 72 100.0 0.0 0.0 A 216 100.0 0.0 0.0 A 645 100.0 0.0 0.0 A 4a 0.640 0.726 72 83.3 13.3 3.4 A 216 76.3 17.0 6.7 A 645 76.7 16.7 6.6 A 4b 15.52 0.004* MF 76.3 17.0 6.7 A Danios 100.0 0.0 0.0 B NoMF 100.0 0.0 0.0 B a Treatment levels for Experiments 1 and 2 are eastern mosquitofish/m2 whereas the levels for Experiments 3 and 4a are stems/m2. Danios and no eastern mosquitofish (= NoMF) treatments in Experiment 4b are controls. b Percent scores are the mean per centages of fish with the score among surviving adult fish. Scores used are as follows: 0 = no damage (caudal fin whole and intact); 1 = moderate damage (< 50% of caudal fin missing), and 2 = severe damage (> 50% of caudal fin missing). c The values calculated by Kruskal-Wallis (K-W) rank-sum test. d Significant differences are indica ted by p-values shown with an and the subsequent letters indicate which groups were significantly different by Dunns post hoc analysis.
42 Table 3-2. Summary of caudal fin damage results fo r surviving swordtails in all experiments and treatments. Experiment Treatmenta Scoreb 0 (%) Score 1 (%)Score 2 (%) K-W ( 2 )c P-value / Dunns groupingd 1 2.02 0.364 21 94.2 2.9 2.9 A 43 90.2 2.9 6.9 A 86 100.0 0.0 0.0 A 2 8.51 0.014* 0 100.0 0.0 0.0 A 21 90.6 9.4 0.0 AB 86 64.3 21.2 14.5 B 3 0.00 1.00 72 93.3 4.4 2.3 A 216 93.3 6.7 0.0 A 645 100.0 0.0 0.0 A 4a 0.48 0.783 72 80.0 20.0 0.0 A 216 83.3 16.7 0.0 A 645 76.7 20.0 3.3 A 4b 5.82 0.054* MF 83.3 16.7 0.0 A Danios 90.0 10.0 0.0 B NoMF 96.7 3.3 0.0 B a Treatment levels for Experiments 1 and 2 are eastern mosquitofish/m2 whereas the levels for Experiments 3 and 4a are stems/m2. Danios and no eastern mosquitofish (= NoMF) treatments in Experiment 4b are controls. b Percent scores are the mean per centages of fish with the score among surviving adult fish. Scores used are as follows: 0 = no damage (caudal fin whole and intact); 1 = moderate damage (< 50% of caudal fin missing), and 2 = severe damage (> 50% of caudal fin missing). c The values calculated by Kruskal-Wallis (K-W) rank-sum test. d Significant differences are indica ted by p-values shown with an and the subsequent letters indicate which groups were significantly different by Dunns post hoc analysis.
43 variable platyfish0 10 20 30 40 50 60 70 80 90 100 214386 A A B swordtails0 10 20 30 40 50 60 70 80 90 100 214386 A AB B Percent survival Eastern mosquitofish / m2 Figure 3-1. Results from Experiment 1. Average percent survival ( SE ) of adult poeciliids introduced into previously-acclimated popul ation of eastern mosquitofish across a range of mosquitofish density. Five repl icates of each treatment (10 individual nonindigenous fish per tank) were run for va riable platyfish and five were run using swordtails. Letters indicate si gnificantly different groupings.
44 variable platyfish adults0 10 20 30 40 50 60 70 80 90 100 02186 A A A swordtail adults0 10 20 30 40 50 60 70 80 90 100 02186 C B A variable platyfish juveniles0 10 20 30 40 50 60 70 80 90 100 02186 A B B swordtail juveniles0 10 20 30 40 50 60 70 80 90 100 02186 A B B Percent survival Eastern mosquitofish / m2 Figure 3-2. Results from Experiment 2. Average pe rcent survival ( SE) of adult and juvenile variable platyfish and swordtails when sy mpatric with three different densities of eastern mosquitofish. Letters indicat e significantly different groupings.
45 variable platyfish0 10 20 30 40 50 60 70 80 90 100 72216645 A A A swordtails0 10 20 30 40 50 60 70 80 90 100 72216645 A B A Percent survival Stems / m2 Figure 3-3. Results from Experiment 3. Average percent survival ( SE ) of adult poeciliids introduced into a previously-e stablished population of mosqu itofish across a range of artificial vegetation stem de nsity with constant eastern mosquitofish density. Letters indicate significantl y different groupings.
46 variable platyfish Day 10 5 10 15 20 25 30 35 40 45 50 72216645 A A A variable platyfish Day 30 5 10 15 20 25 30 35 40 45 50 72216645 A A A swordtails Day 10 5 10 15 20 25 30 35 40 45 50 72216645 A A A swordtails Day 30 5 10 15 20 25 30 35 40 45 50 72 216645 A A A Total attacks Stems / m2 Figure 3-4. Total attack results from Experiment 4a. Mean observed attacks ( SE) on adult poeciliids introduced into a previously-est ablished population of mosquitofish across a range of stem density. Letters indi cate significantly different groupings.
47 variable platyfish Day 10 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 72216645 AB A B variable platyfish Day 30 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 72216645 A A A swordtails Day 10 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 72216645 A A B swordtails Day 30 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 72216645 A A B Proportion of fish in vegetation Stems / m2 Figure 3-5. Habitat use results from Experime nt 4a. Mean proportion of adult introduced poeciliids observed in the vegetated area of the experimental mesocosm ( SE) when introduced into a previously-e stablished population of mosqu itofish across a range of artificial vegetation stem density. Letters indicate si gnificantly different groupings.
48 variable platyfish Day 10 5 10 15 20 25 30 35 40 45 50 DaniosNoMFMF A A B variable platyfish Day 30 5 10 15 20 25 30 35 40 45 50 DaniosNoMFMF A A B swordtails Day 10 5 10 15 20 25 30 35 40 45 50 DaniosNoMFMF A A B swordtails Day 30 5 10 15 20 25 30 35 40 45 50 DaniosNoMFMF A A B Total attacks Treatment Figure 3-6.Total attack results from Experiment 4b. Mean observed attacks ( SE) on adult poeciliids introduced into a previously-est ablished population of mosquitofish across a range of stem density. Letters indi cate significantly different groupings.
49 variable platyfish Day 10 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 DaniosNoMFMF A A B variable platyfish Day 30 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 DaniosNoMFMF A A A swordtails Day 10 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 DaniosNoMFMF A A A swordtails Day 30 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 DaniosNoMFMF A B B Proportion of fish in vegetation Treatment Figure 3-7. Habitat use results from Experime nt 4b. Mean proportion of adult introduced poeciliids observed in the vegetated area of the experimental mesocosm ( SE) when introduced into a previously-e stablished population of mosqu itofish across a range of artificial vegetation stem density. Letters indicate si gnificantly different groupings.
50 CHAPTER 4 DISCUSSION This series of experim ents demonstrated th at eastern mosquitofish have strong negative effects through predation on the survival of variable platyfish and swor dtails, two nonindigenous poeciliids. Effects were appare nt at both adult and juvenile li fe stages of these species, indicating a possible influence on various steps of the invasion process (Maron and Vila 2001). Results also provide evidence to support the hypothesis that eastern mosquitofish may be limiting the invasions of small-bodied fishes in Fl oridas freshwater syst ems, functioning as an agent of biotic resistance. Although previously suggested as gape-limite d (Taylor et al. 2001), these experiments showed that eastern mosquitofish are capable of immobilizing and killing fishes larger than themselves. Eastern mosquitofish in this study typically attacked their prey by biting the caudal fin prior to killing them. This type of attack behavior has been described previously with mosquitofish (see Meffe 1985; Baber and Ba bbitt 2004; Laha and Mattingly 2007; for a description of mosquitofish atta cks on organisms of similar size). In the present study, all adult variable platyfish and swordta il mortalities showed considerab le damage to the caudal region, often with substantial portions of the caudal fin missing as well as loss of scales from the caudal peduncle. Even in instances where the attacked fish survive d, it is likely that any damaged sustained would cause the fish to be more susceptible to further attacks or die from wound infection (Noga 1996; Wildgoose 2001). Variable platyfish and swor dtails have been shown to exhibit conspecific aggression (Earley 2006). Howe ver, in Experiment 4b no conspecific attacks were observed in treatments lacking eastern mos quitofish and numbers of conspecific attacks were low ( 2/observation period) in the mosquitofish treatments. The lack of conspecific
51 attacks confirms that mosquitofish attacks and aggression were the source of mortality in these species. Meffe (1985) found that mosquitofish were cap able of harassing and killing fish similar size, but the present stud y is the first experimental demonstr ation of mosquito fish predation on fish substantially larger than themselves. In th ese mesocosms, eastern mo squitofish killed adult variable platyfish and swordtails that were between one and a half to two times longer (i.e., total length) than the eastern mosquito fish. In addition, variable plat yfish in this study were heavier than eastern mosquitofish by a factor of about four and swordtails were about six times heavier. The substantial differences in size and weight of eastern mosquitofish and the prey used in this experiment is evidence that this small predator may have the ability to im pact a larger size range of fish species than previously shown (e.g., Me ffe 1985; Taylor et al. 2001; Laha and Mattingly 2007). Furthermore, the data suggest that swordta ils, the larger of the tw o prey species in this study, were killed at a higher rate in many tr eatments. Based on these results, morphology and size may not be the only factor in determining th e effect of eastern mosquitofish on a species. Increasing eastern mosquitofish density was shown to have a negative effect on survival of adults of the two nonindigenous study species. This was expected as predator density increases can increase predation effects overall (Relyea 2003). Moreover, mosquitofish predation effects have been demonstrated as density dependent in mesocosms (Belk and Lydeard 1994; Taylor et al. 2001; Mills et al. 2004). Experiment 1 show ed declining survival with increasing eastern mosquitofish density in both species and this effect was again noted with swordtails in Experiment 2. On the other hand, swordtail surv ival was higher overall in Experiment 2 and density effects were not observed with variable platyfish. This was likely due to trial duration. Run time of Experiment 2 was 4 days, less than half of the run time for the variable platyfish and
52 a day shorter than with the swordtails in E xperiment 1. Furthermor e, it is possible that mosquitofish are slower to exhibit aggressive or predatory behavior when concurrently introduced into the tanks with the nonindigenous study species, ra ther than havi ng established prior residency during the acclimation pe riod (Chellappa et al. 1999; Earley 2006). Although predation by eastern mosquitofish on a dult variable platyfis h and swordtails was documented in the present study, results from the current experiments indicate that the strongest effect on the two species was from predation on j uveniles. These results agree with previous published information describing predation by mosqu itofish on juvenile and larval fishes (Meffe 1985; Belk and Lydeard 1994; Schaefer et al. 1994; Taylor et al. 2001; Mi lls et al. 2004; Laha and Mattingly 2007). In the current study, preda tion on juvenile fish by eastern mosquitofish was pronounced even in treatments where eastern mosquitofish density was low. The strong effect of eastern mosquitofish on the size st ructure and recruitment of prey was further demonstrated by observations on control treatments from Experiment 2 (eastern mosquitofish were absent) in which variable platyfish pr oduced additional young. Meffe (1985) demonstrated that mosquitofish presence did not reduce fecundity in other fish species, but reduced recruitment by continual predation on the juveniles of other species. This predator may effect introduced poeciliids in a similar fashion. In Experiments 1 and 3 of this study, where adult variable platyfish were not introduced into mesocosm tanks lacking eastern mosquitofish, no juvenile variable platyfish were found. It is likely that eastern mosqui tofish did not prevent reproduction, but no juveniles were observed because eastern mosquitofish preyed so heavily on poeciliid neonates (Meffe 1985; Belk and Lydeard 1994; Schaef er et al. 1994; Taylor et al. 2001; Mills et al. 2004; Laha and Mattingly 2007; Experiment 2). Swordta ils did not produce additional recruits in these experiments either, but this might not have been due solely to eastern
53 mosquitofish predation. Failure of swordtails to recruit was potenti ally environmentally determined, because these fish produce few juven iles in the hottest periods of the summer in west-central Florida (C. A. Wa tson, University of Florida, personal communication). Indeed, swordtail recruitment was low in TAL aquacultur e tanks and production ponds at the time of the experiments (C. A. Watson, University of Florida, persona l communication). The apparently higher survival of swordtail juveniles compar ed to the variable platyfish (Figure 3-2) is most likely a result of the size ranges of these two species used. Variable platyfish were more similar in size to neonate s whereas swordtail juveniles were larger. As previously discussed, swordtail recruitment during the timeframe of the study was low due to temperature, limiting the availability of small juveniles. The averag e length of juvenile swordtails was about 4 mm greater than the variable platyfish and these fish represent individuals that were about 2 4 weeks old (R. P. E. Yanong, University of Florida, personal communication). Although different in size range, given the lengths of eastern mosquitofish used, the juveniles of both nonindige nous species were larger than th e estimated gape sizes of the eastern mosquitofish (Taylor et al. 2001). Therefor e, these fish were likely subdued and killed in a similar manner to the adults. Although significant mortality was observed for both adult and juveniles of these species, some mortality in these fishes may be from mosq uitofish harassment, not directed predation. In all experiments, all killed adult fish were recove red with substantial damage to the caudal region, but few had further tissue damage to the rest of the body. Indeed, there was little subsequent scavenging of the dead fish by the mosquitofish. However, a lack of scavenging may be due to the daily removal of dead individuals wh ich may have allowed insufficient time for consumption. Across all experiments and replic ates, one adult variable platyfish and three
54 swordtails were recovered with substantial am ounts of tissue missing from the abdomen and gut cavity. These were the only insi stences of substantial consump tion. However, consumption of an entire fish may not be necessary to define pred ation in eastern mosquitofish. It is likely that mosquitofish gained nutrition from the consumed fi n tissue, scales, and associated tissues of the caudal peduncle. For example, scales and fin tis sue have been demonstrated as nutritionally valuable, even dominating the diet of some neotropical fishes (N ico and DeMorales 1994). The results of fin damage analysis support the hypothesis that a dult mortalities were caused by eastern mosquitofish harassment. Most treatment groups, where mortality was significantly different, had no statistical differences in the amount of damage to surviving fish. Furthermore, the only fish observed with differen ces in fin damage were from Experiments 2 and 4, where the experimental time was short (Table 3-1; 3-2). It is likely that in these experiments, had they continued, those fish with damage would have died as a result. The data on mortalities and fin damage as well as obser vations suggest that mosquitofish focus aggression on a few fish early on, tend to attack previously damaged fish until they are killed, and, after killing these individuals, direct attacks on new individuals. This pattern of focused attacks on already damaged fish would not necessarily be expected if scale and fin tissue was a targeted food source in these experiments unless previous damage rende rs fish easier to successfully attack to obtain more fin tissue or mucus. Attacks on juveni les were likely directed predation by eastern mosquitofish. Although exceeding gape size, juveniles were completely consumed in all cases and this result is more indicativ e of predation rather than harassment. Harassment deaths with adults and predation on juveniles, although th ey may represent different behaviors in mosquitofish, have the same ecological implications for biotic resistance to these species.
55 The results from these mesocosm studies show the potential for mosquitofish to act as a biotic resistance to invasion of variable platyf ish and swordtails. By reducing adult survival, mosquitofish may directly eliminate an in troduced population of these fishes, preventing establishment, or limit their range and expans ion (Maron and Vila 2001; Kolar and Lodge 2002). If eastern mosquitofish are limiting invasion suc cess of introduced populations of these species, then based on the results of th is study and previous experiment s (Belk and Lydeard 1994; Taylor et al. 2001; Mills et al. 2004), the strength of the effect will be depe ndent on the density of mosquitofish in the invaded habitat. Increa sing predator abundance has shown to be an important factor in limiting the ra nge of some aquatic invaders (deRivera et al. 2005); however even a high density of predators can be overw helmed by propagule pressure (Von Holle and Simberloff 2005; Hollebone and Ha y 2007), a factor not studied in these experiments. Biotic resistance through predation on juve niles is likely in natural systems given the strength of this interaction observed in mesocosms. Reducing or preventing recruitment can limit these species ability to successfully establish and can eliminate introduced populations over time. Previous studies have found that introduced populations of mosquitofish have reduced or extirpated endemic populations of certain small-bodied fishes through recruitmen t failure (Meffe 1985; Belk and Lydeard 1994). As short-lived speci es with low population storage capacity (Secor 2007), brief periods of recruitment failure coul d potentially eliminate a local population of variable platyfish or sw ordtails. While there have been no studies conducted to monitor wild populations of introduced poeciliid s in Florida (Fuller et al. 1999; L. G. Nico, personal communication), previously published observations seem to indicate that introduced populations of variable platyfish and swordtails in Flor ida are typically uncomm on, highly localized, and rarely persist for more than a few years (F uller et al. 1999; L. G. Nico, USGS, personal
56 communication). A possible factor is consum ption of neonates by eastern mosquitofish preventing recruitment. Eastern mosquitofish predation on adults and j uveniles represent a poten tial biotic factor in determining invasion success of introduced poeciliids and other small-bodied, nonindigenous fishes. However, abiotic factors can also be an important contributing factor in determining success of freshwater fish inva sions (Kolar and Lodge 2002). Th is study investigated how an environmental variable, habitat complexity at an introduction site, mi ght influence invasion success of variable platyfish a nd swordtails in the presence of eastern mosquitofish. The declining survival of swordtails at high stem density in Experiment 3 was an unexpected result. Increasing stem densities have been show to redu ce predation efficiency with large, gape-limited, active piscivorous fish such as largemouth bass and northern pike Exox lucius in multiple field and mesocosm settings (Savino and Stein 1982; Anderson 1984; Eklov and Hamrin 1989; Savino and Stein 1989; Hayse and Wissing 1996; Eklo v and VanKooten 2001; Hickey and Kohler 2004). Although mosquitofish predation is from re peated attacks, which is different from the large piscivores studied in relation to habitat complexity, the same trend of reduced predation effects in increasingly complex ha bitat was noted with mosquitofish predation on tadpole species by Baber and Babbitt (2004). It is important to note that Experiment 3 did not measure the effect of stem density on predation efficiency of mos quitofish per se, but the survival of the two introduced poeciliid populations in the presence of mosquitofish, wh ich is a different measure. While the overall survival of the study sp ecies may be impacted by changing predation efficiency of mosquitofish, it is al so affected by the prey species be havior and ability to refuge in habitats of different complexity. The results from Experiment 4a indi cate that changes in behavior may be an important fact or in declining swordtail surviv al with increasing stem density.
57 Although no pattern was apparent in attacks in relation to stem density, the highest proportion of swordtails in the vegetation was found in the high stem density treatments for both Days 1 and 3 in this study. Qualitative observa tions suggested mosquitofish heavily associated with vegetation in these tanks across all stem densities studied. Therefore with more swordtails present in the vegetation at high stem density, these fish would be in contact with mosquitofish more than in the other stem density treatments. While mosquitofish efficiency may decrease with increasing habitat complexity (Baber a nd Babbitt 2004), encounter rates may increase at high stem densities due to greater use of ve getated habitat by the st udy species under this condition. Encounter ra te is the fundamental component in determining predation rates (Osenberg and Mittlebach 1989) and thus may be the cause of the lower swordtail survival at high stem density in this experiment. Prey species frequently use complex habitat as a predatory refuge (Savino and Stein 1982; Anderson 1984; Eklov and Hamrin 1989; Savi no and Stein 1989; Hayse and Wissing 1996; Eklov and VanKooten 2001; Hickey and Kohler 2004) however different species vary in ability to alter behavior and habitat use patterns as an anti-predato r response (Savino and Stein 1989; Bean and Winfield 1995). Holker et al. (2007) demonstrated th e inability of rudd Scardinius erythophthalmus to shift behavior in the presence of a novel predator, thereby suffering much higher mortality than morphologically similar speci es that exhibited anti -predatory responses. Baber and Babbitt (2004) demonstrated this eff ect specifically with mosquitofish predation, showing the inability of one species of tadpole to use habitat as effectively as a similar species as an anti-predatory response. Differing anti-p redatory responses may explain the seemingly differential predation rates betw een variable platyfish and swor dtails by mosquitofish. The increased use of the vegetated area, by the swordt ails at high stem density despite increasing
58 encounter rates and subsequent higher mortality rates suggests an inability for swordtails to appropriately respond to predation threats. Although variable platyfish were found more commonly in the vegetation at high stem density during Day 1 as well, fish moved out of the vegetation by Day 3 and mortality was not signif icantly higher at high stem density for this species. Furthermore the data suggest higher mortal ity rates, over shorter tim e, of the swordtails compared to the variable platyfish in Experiment s 1 and 3. The data also imply that variable platyfish, while similar morphologically to swor dtails, respond more effectively to eastern mosquitofish predation, and that a differing antipredatory response between the two species may explain the observed differing su rvival patterns. Interestingly, variable platyfish is locally established in Florida and swordtail is not. Acts of predation and aggression causing reduced survival of adults and juveniles is a direct impact of mosquitofish on these two studied poeciliids. However, the data on habitat use preferences of these fish in the presence and abse nce of mosquitofish in Experiment 4b indicate the potential for mosquitofish-mediated indirect eff ects. The Day 3 data on swordtails indicate that mosquitofish are excluding swordtails fr om the vegetation compared to the Danios treatment. The Danios treatment in this experime nt is likely a better m easure for the behavioral preferences of these species th an the NoMF control because zeb ra danios provided interaction and a density of fish that would be more like th e native habitat of these poeciliids. Without zebra danios, in the NoMF treatments and control groups of Experiment 2, qualitative observations were that the nonindigenous poeciliids moved little and remained schooled. This behavior can indicate a problem of scale, where the tank is too much space for such a small group of fish and their behavior is similar to that under threat of predation (Loiselle 1985) Although zebra danios are not a part of the Central American icht hyofauna and therefore va riable platyfish and
59 swordtails would not normally encounter them, they served as dither fish which are frequently used in aquaria and captive situations to elucidat e a more natural behavioral response from other species, by indicating that there is no predation threat (Loiselle 1985). There are potential implications for biotic resistance if mosquitofish aggression and predation threat excludes swordtai ls from structurally complex ha bitat. In these experimental systems with one predator, the sw ordtails chose to separate them selves and decrease interaction with the mosquitofish by avoiding the complex habitat (except at high levels of structural complexity). However, in natural systems, this mode of behavior may not be available to a small-bodied introduced species. Florida has abunda nt larger piscivorous species that will feed readily on fish of this size. Pond studies have sh own that utilizing shallo w or vegetated areas is often the main way for small fishes to avoid la rger predators (DeVries 1990; Persson and Eklov 1995; Holker et al. 2007). Mosquitofish exclusion could therefore facilita te predation by other, larger piscivores and increase overall predat ory effects on small-bodied, introduced fishes. Multiple predator effects where predators are facilitated by another species through habitat exclusion have been shown in freshwater fish es (Eklov and VanKooten 2001). Harvey et al. (2003) specifically demonstrated multiple pred ators limiting invasion success where predators excluded a nonindigenous fish from refuge habitat and thereby facilitated th e predation effect of an open water predator. If mos quitofish are able to facilitate larger predators in a similar manner, then this may reduce invasion success of th ese species. For example, while the variable platyfish showed high survival in the presence of only eastern mos quitofish, it is possible that in a natural system their survival would be lower because of complex, multi-species interactions. Swordtail mortality could increase as well, as these fish were shown to be excluded from vegetation by eastern mosquitofish, which is an important refuge in a natural system. The
60 strength of a facilitative effect by eastern mosquitofish on other predators will depend on the behavior and morphology of co-exis ting predators, the behavior of the invading species, and the nature of the habitat. Multiple predator effects may represent a fruitful area of future biotic resistance research. Although the present study focused on impact s of eastern mosquitofish on variable platyfish and swordtail, there are many other small-bodied fish species in the ornamental and aquarium trade in Florida (e.g., Hill and Yanong 2002). Some have been collected or observed in the wild, but few are known to have reproducing populations and, w ith one exception, none are widely established (Fuller et al. 1999; Nico and Fuller 1999) Mosquitofish, based on the results of this study and prev ious research, are likely to prey upon these species and may augment the ability of aquatic communities to resi st their establishment. Investigating resident predator and introduced prey intera ctions is necessary to increase prediction and assess risks of introduced species. While there is a growing body of literature describing characteristics of successful fish invasions based on the introdu ced species biology and native range, and the abiotic characteristics of the introduction site (Kolar and L odge 2002; Marchetti et al. 2004; Ruesink 2005), few if any, include measures of biotic resistance su ch as occurrence and abundance of predators. More studies of predat or effects are needed to understand biological invasions and the factors that contribute to the failure or success of introductions.
61 APPENDIX A FRESHWATER FISH INTROD UC TIONS INTO FLORIDA Table A-1. Status of known fres hwater fish species with maximum TL < 15cm introduced into the state of Florida. Common name Scientific name Status1 Siamese fighting fish Betta splendens Formerly reproducing Banded gourami Colisa fasciata Reported Thicklip gourami Colisa labiosa Reported Dwarf gourami Colisa lalia Reported Paradise fish Macropodus opercularis Reported Pearl gourami Trichogaster leerii Reported Blue gourami Trichogaster trichopterus sumatranus Reported Croaking gourami Trichopsis vittata Formerly reproducing Corydoras Corydoras spp. Reported Bloodfin tetra Aphyocharax anisitsi Reported Black tetra Gymnocorymbus ternetzi Reported Serpae tetra Hyphessobrycon serape Reported Silver dollar Metynnis hypsauchen Localized Redeye tetra Moenkhausia sanctaefilomenae Reported Blue acara Aequidens pulcher Reported Convict cichlids Cichlasoma nigrofasciatum Formerly reproducing Eastern happy Haplochromis callipterus Possibly established African jewlfish Hemichromis letourneuxi Established Scrapermouth cichlid Labeotropheus spp. Reported Freshwater angelfish Pterophyllum scalare Reported Freshwater angelfish Pterophyllum spp. Reported Lake Tanganyka dwarf cichlid Telmatochromis bifrenatus Reported Coolie loach Pangio kuhlii Reported Malabar danio Danio malabaricus Reported Zebra danio Danio rerio Reported Fathead minnow Pimephales promelas Reported Rosy barb Puntius conchonius Reported Dwarf barb Puntius gelius Reported Tiger barb Puntius tetrazona Reported Bristlenosed catfish Ancistrus spp. Possibly established Western mosquitofish Gambusia affinis Reported Lyretail black molly Poecilia hybrid Reported Black m olly hybrid Poecilia latipinna x velifera Reported
62 Table A-1. Continued Common Name Scientific Name Status1 Broadspotted molly Poecilia latipunctata Reported Penten molly Poecilia petenensis Reported Guppy Poecilia reticulate Formerly reproducing Mexican molly Poecilia sphenops Reported Green swordtail Xiphophorus hellerii Formerly reproducing Red swordtail Xiphophorus hellerii x maculatus Reported Platyfish/swordtail Xiphophorus hellerii x variatus Reported Southern platyfish Xiphophorus maculates Formerly reproducing Platyfish hybrid Xiphophorus maculates x variatus Reported Variable platyfish Xiphophorus variatus Localized Note: Status from FWC (2007) and reports from USGS (2008). 1 Established = permanent populations, possibly established = species believed to be repr oducing, localized = a confined reproducing populations, and form erly reproducing = at one point showed evidence of reproduction but has since been eliminated.
63 APPENDIX B DATA SHOWN IN FIGURES FOR ALL EXPERIMENTS Table B-1. S urvival data from all treatments of Experiments 1-3 for adults and juveniles of variable platyfis h and swordtails. Experiment Species1 Life stage Treatment2 Mean survival (%) Standard error 1 Platy Adult 21 100 0 Platy Adult 43 96 2.4 Platy Adult 86 82 4.9 Sword Adult 21 64 4 Sword Adult 43 48 8 Sword Adult 86 30 6.3 2 Platy Adult 0 100 0 Platy Adult 21 100 0 Platy Adult 86 98 2 Platy Juvenile 0 100 0 Platy Juvenile 21 14 11.7 Platy Juvenile 86 0 0 Sword Adult 0 98 2 Sword Adult 21 86 2.5 Sword Adult 86 58 5.8 Sword Juvenile 0 96 4 Sword Juvenile 21 28 11.7 Sword Juvenile 86 2 2 3 Platy Adult 72 32 4.9 Platy Adult 216 74 10.3 Platy Adult 645 48 20.3 Sword Adult 72 82 7.3 Sword Adult 216 76 7.5 Sword Adult 645 40 8.9 1 Platy=variable platyfis h; Sword=swordtail. 2 Treatment levels for Experiments 1 and 2 are eastern mosquitofish/m2 whereas the levels for Experiments 3 are stems/m2.
64 Table B-2. Total attack data from all treatments of Experiments 4a and 4b for adults of variable platyfish and swordtails. Experiment Species1 Day Treatment2 Mean attacks observed Standard error 4a Platy 1 72 12.3 9.6 Platy 1 216 14 18.4 Platy 1 645 15 8.9 Sword 1 72 25 17.6 Sword 1 216 25 20 Sword 1 645 12.3 9.3 Platy 3 72 17.3 1.5 Platy 3 216 11 5.2 Platy 3 645 15 3.6 Sword 3 72 10.3 4.7 Sword 3 216 11 2.6 Sword 3 645 6.3 4.7 4b Platy 1 Danios 0 0 Platy 1 NoMF 0 0 Platy 1 MF 14 18.3 Sword 1 Danios 0 0 Sword 1 NoMF 0 0 Sword 1 MF 25 20 Platy 3 Danios 0 0 Platy 3 NoMF 0 0 Platy 3 MF 11 5.2 Sword 3 Danios 0 0 Sword 3 NoMF 0 0 Sword 3 MF 11 2.6 1 Platy=variable platyfis h; Sword=swordtail. 2 Treatment levels for Experiment 4a are stems/m2 whereas 4b treatments are: MF (high mosquitofish density) and two controls 1) with zebra danios at high density (Danios), and 2) no east ern mosquitofish (NoMF) treatments.
65 Table B-3. Habitat use data from all treatments of Experiments 4a and 4b for adults of variable platyfish and swordtails. Experiment Species1 Day Treatment2 Mean proportion fish in vegetation Standard error 4a Platy 1 72 0.27 0.00 Platy 1 216 0.43 0.03 Platy 1 645 0.48 0.07 Sword 1 72 0.15 0.02 Sword 1 216 0.10 0.03 Sword 1 645 0.44 0.13 Platy 3 72 0.20 0.18 Platy 3 216 0.25 0.11 Platy 3 645 0.38 0.09 Sword 3 72 0.08 0.01 Sword 3 216 0.11 0.05 Sword 3 645 0.38 0.10 4b Platy 1 Danios 0.08 0.05 Platy 1 NoMF 0.17 0.07 Platy 1 MF 0.43 0.03 Sword 1 Danios 0.48 0.12 Sword 1 NoMF 0.39 0.19 Sword 1 MF 0.10 0.03 Platy 3 Danios 0.10 0.04 Platy 3 NoMF 0.11 0.04 Platy 3 MF 0.25 0.11 Sword 3 Danios 0.61 0.04 Sword 3 NoMF 0.24 0.15 Sword 3 MF 0.11 0.05 1 Platy=variable platyfis h; Sword=swordtail. 2 Treatment levels for Experiment 4a are stems/m2 whereas 4b treatments are: MF (high mosquitofish density) and two controls 1) with zebra danios at high density (Danios), and 2) no east ern mosquitofish (NoMF) treatments.
66 LIST OF REFERENCES Anderson, O. 1984. Optim al foraging by largemouth bass in structured environments. Ecology 65:851-861. Angus, R.A. 1989. A genetic overview of Poeciliid fishes. Pages 51-68 in G. K. Meffe and F. F. Snelson, Jr., editors. Ecology and evolution of livebearing fishes (Poeciliidae). Prentice Hall, Englewood Cliffs, NJ. Baber, M.J., and K.J. Babbitt. 2004. Influence of habitat complexity on predator-prey interactions between the fish (Gambusia holbrooki ) and tadpoles of Hyla squirella and Gastrophryne carolinensis Copeia 2004:173-177. Baltz, D.M., and P.B. Moyle. 1993. Invasion re sistance to introduced species by a native assemblage of California stream fish es. Ecological Applications 3:246-255. Bean, C.W., and I.J. Winfield. 1995. Hab itat use and activity patterns of roach ( Rutilus rutilus (L.)), rudd ( Scardinius erythrophthalmus (L.)), perch ( Perca fluviatilis L.) and pike ( Esox lucius L.) in the laboratory: the role of pred ation threat and stru ctural complexity. Ecology of Freshwater Fish 4:37-46. Belk, M.C., and C. Lydeard. 1994. Effect of Gambusia holbrooki on a similar-sized, syntopic poeciliid, Heterandria formosa : competitor or predator? Copeia 1994:296-302. Boschung, H.T., and R.L. Mayden. 2004. Fishes of Alabama. Smithsonian Books, Washington, D.C. Caiola, N., and A. de Sostoa. 2005. Possible reasons for the dec line of two native toothcarps in the Iberian peninsula: evidence of competition w ith the introduced eastern mosquitofish. Journal of Applied Ichthyology 21:358-363. Case, T.J. 1990. Invasion resistance arises in strongly interacting species-rich model competition. Proceedings of the National Academy of Sciences of the United States of America 87: 9610-9614. Chellappa, S., M.E. Yamanoto, M.S.R.F. C acho, and F.A. Huntingford. 1999. Prior residence, body size and the dynamics of territorial disputes between male freshwater angelfish. Journal of Fish Biology 55:1163-1170. Courtenay, W.R., Jr., and G.K. Meffe. 1989. Small fishes in strange pl aces: a review of introduced poeciliids. Pages 319-331 in G. K. Meffe and F. F. Snelson, Jr., editors. Ecology and evolution of livebearing fishes (Poeciliidae). Prentice Hall, Englewood Cliffs, NJ. Courtenay, W.R., Jr., H.F. Sahl man, W.W. Miley II, and D.J. Herrema. 1974. Exotic fishes in fresh and brackish waters of Florid a. Biological Cons ervation 6:292-302.
67 deRivera, C.E., G.M. Ruiz, A.H. Hines, and P.Jivoff. 2005. Biotic resistance to invasion: native predator limits abundance and distributi on of an introduced crab. Ecology 86:3364-3376. DeVries, D.R. 1990. Habitat use by bluegill in la boratory pools: where is the refuge when macrophytes are sparse and alternative prey are present? Environmental Biology of Fishes 29:27-34. Duggan, I.C., C.A.M. Rixon, and H.J. M acIsaac. 2006. Popularity and propagule pressure: determinants of introduction and establishment of aquarium fish. Biological Invasions 8:377-382. Earley, R.L. 2006. Xiphophorus : carving a niche towards a broader understanding of aggression and dominance. Zebrafish 3:287-298. Eklov, P., and S.F. Hamrin. 1989. Predatory efficiency and prey se lection: interactions between Pike Esox lucius, perch Peca fluviatilis and rudd Scardinus erythrophthalmus Oikos 56:149-156. Eklov, P., and T. VanKooten. 2001. Facilitation am ong piscivorous predators: effects of prey habitat use. Ecology 82:2486-2494. Elton, C. 1958. The ecology of invasions by animals and plants. The University of Chicago Press, Chicago, IL. Espinosa-Garca, F. J., J. L. Villaseor, and H. Vibrans. 2004. The rich generally get richer, but there are exceptions: correlations between sp ecies richness of native plant species and alien weeds in Mexico. Divers ity and Distributions 10:399-407. Field, D.W., and J.E. Thomerson. 1997. Fishes of the continental waters of Belize. University Press of Florida, Gainesville, FL. Fridley, J.D., J.J. Stachowicz, S. Naeem, D. F. Sax, E.W. Seabloom, M.D. Smith, T.J. Stohlgren, D. Tilman, and B. Von Holle 2007. The invasion paradox: reconciling pattern and process in species invasions. Ecology 88:3-17. Fuller, P.L., L.G. Nico, and J.D. Williams. 1999. Nonindigenous fishes introduced into inland waters of the United States. American Fisheries Society, Special Publication 27, Bethesda, MD. FWC (Florida Fish and Wildlife Conserva tion Commission). 2007. List of exotic freshwater fishes collected from Florida fresh waters. Available: http://www.floridaconservation.org/f ishing/pdf/Exotic%20List.pdf (May 2008). Galat, D. L., and B. Robertson. 1992. Response of endangered Poeciliopsis occidentalis sonoriensis in the Rio Yaqui drainage, Arizona, to introduced Gambusia affinis Environmental Biology of Fishes 33:249-264.
68 Gertzen, F., O. Familiar, and B. Leung. 2008. Quantifying invasion pathways: fish introductions from the aquarium trade. Ca nadian Journal of Aquatic and Fisheries Sciences 65:1265-1273. Harding, J.M. 2003. Predation by blue crabs, Callinectes sapidus on rapa whelks, Rapana venosa : possible natural controls for an invasive species? Journal of Experimental Marine Biology and Ecology 297:161-177. Harvey, B.C., J.L White, and R.J. Nakamoto. 200 4. An emergent multiple predator effect may enhance biotic resistance in a stream fish assemblage. Ecology 85:127-133. Hayse, J.W. and T. Wissing. 1996. Effects of stem density of artificial vegetation on abundance and growth of age-0 bluegills and predation by largemouth bass. Transactions of the American Fisheries Society 125:422-433. Hickey, C.W., and C.C. Kohler. 2004. Comparis on of bluegill consumption rates by largemouth bass and sunshine bass in structured a nd nonstructured artificial environments. Transactions of the American Fisheries Society 133:1524-1528. Hill, J.E. 2006. Regulations pertaining to non-nativ e fish in Florida aquaculture. University of Florida IFAS Extension Ci rcular FA121. Available: http://edis.ifas.ufl.edu/FA121 (January 2008). Hill, J.E., an d C. A. Watson. 2007. Diet of th e nonindigenous Asian swamp eel in tropical ornamental aquaculture ponds in west centr al Florida. North America Journal of Aquaculture 69:139-146. Hill, J.E., and R.P.E. Yanong. 2002. Ornamental fish commonly cultured in Florida. University of Florida IFAS Extension Circular FA054. Available : http://edis.ifas.ufl.edu/FA504 (January 2008). Holker, F.,H. Dorner, T. Schulze, S.S. Haerte l-Borer, S.D. Peacor, and T. Mehner. 2007. Species-specific responses of planktivorous fish to the introduction of a new piscivore: implications for prey f itness. Freshwater Biology 52:1793-1806. Hollander, M. and D.A. W olfe. 1973. Nonparametric statistical methods. John Wiley and Sons, New York, NY. Hollebone, A.L., and M.E. Hay. 2007. Propagule pressure of an invasive crab overwhelms native biotic resistance. Ma rine Ecology Progress Series 42:191-196. Howe, E. C., R. Lim, and M. Burchett. 1997. Impact of the introduced poeciliid Gambusia holbrooki (Girard, 1859) on the growth and reproduction of Pseudomugil signifer (Kner, 1865) in Australia. Mari ne and Freshwater Research 48:425-434.
69 Hoyer, M.V., and D.E. Canfield, Jr. 1994. Ha ndbook of common freshwater fishes in Florida lakes. University of Florida, Institute of Food and Agricultural Sciences, Gainesville, FL. Kolar, C.S., and D.M. Lodge. 2002. Progress in invasion biology: predicting invaders. Trends in Ecology and Evolution 16:199-204. Laha, M., and H.T. Mattingly. 2007. Ex situ evaluation of impacts of invasive mosquitofish on the imperilled Barrens tompminnow. Environmental Biology of Fish 78:1-11. Leprieur, F., O. Beauchard, S. Blanchet, T. Oberdorff, and S. Brosse. 2008. Fish invasions in the world's river systems: when natural pro cesses are blurred by hu man activities. PLoS Biology 6:404-410. Loiselle, P.V. 1985. The cichlid aquarium Tetra Press, Morris Plains, NJ. Marchetti, M.P., P.B. Moyle, and R. Levine. 2004. Alien fishes in California watersheds: characteristics of successful and failed invaders. Ecological A pplications 14: 587596. Maron, J.L., and M. Vila. 2001. When do herbivor es affect plant inva sion? Evidence for the natural enemies and biotic re sistance hypothesis. Oikos 95: 361-373. Martin, P., and P. Bateson. 1993. Measuring be haviour: an introductory guide. Second Edition. Cambridge University Pres s, Cambridge, Great Britain. McKinney, M.L., and J.L. Lockwood. 1999. Bi otic homogenization: a few winners replacing many losers in the next ma ss extinction. Trends in Ecology and Evolution 14:450-453. Meffe, G. K. 1985. Predation and species replacemen t in American southwestern fishes: a case study. Southwest Naturalist 30:173-187. Meffe, G. K., D. A. Hendrickson, W. L. Minc kley, and J. N. Rinne. 1983. Factors resulting in decline of the endange red Sonoran topminnow Poeciliopsis occidentalis (Atheriniformes: Poeciliidae) in the United Stat es. Biological Conservation 25:135-159. Meffe, G.K., and F.J. Snelson. 1989. An ecologi cal overview of Poeciliid fishes. Pages 13-31 in G. K. Meffe and F. F. Snelson, Jr., editors. Ecology and evolution of livebearing fishes (Poeciliidae). Prentice Hall, Englewood Cliffs, NJ. Mills, M.D., R.B. Rader, and M.C. Belk. 2004. Complex interactions between native and invasive fish: the simultaneous effects of multiple negative interactions. Oecologia 141:713-721.
70 Mills, D., and G. Vevers. 1989. The tetra encyclop edia of freshwater tropical aquarium fishes. Tetra Press, Morris Plains, NJ. Nico, L. G. 2005. Changes in the fish fauna of the Kissimmee River Basin, peninsular Florida: Non-Native additions. Pages 523-556 in J. N. Rinne, R. M. Hughes, and B. Calamusso, editors. Historical changes in large river fi sh assemblages of the Americas. American Fisheries Society Symposium 45. Bethesda, MD. Nico, L. 2007. Xiphophorus variatus USGS Nonidigenous Aquatic Species Database, Gainesville, FL. Available: http://nas.er.usgs.gov/queries/ FactSheet.asp?speciesID=874 (May 2008). Nico, L.G., and M. DeMorales. 1994. Nutrient c ontent of piranha (Char acidae, Serrasalm inae) prey items. Copeia 2:524-528. Nico, L. G., and P. L. Fuller. 1999. Spatial and temporal patterns of nonindigenous fish introductions in the United States. Fisheries 24:16-27. Noga, E.J. 1996. Fish Disease, diagnosis and treatment. Mosby-Yearbook, St. Louis, MO. Ogilvie, V.E., J.W. Davis, R. Seaton, and C.Gibson. 1974. Florida game and freshwater commission annual report 1973-1974. Non-na tive research project, Boca Raton FL. Olden, J.D., N.L. Poff, M.R. Douglas, M.E. Douglas, and K.D. Fausch. 2004. Ecological and evolutionary consequences of biotic homogenization. Tren ds in Ecology and Evolution 19:18-24. Osenberg, C., and G.G. Mittlebach. 1989. Effect s of body size on the pred ator-prey interaction between pumpkinseed sunfish and gastr opods. Ecological Monographs 59:405-432. Page, L. M., and B. M. Burr. 1991. A field guide to freshwater fishes of North America north of Mexico. The Peterson Field Guide Seri es, volume 42. Houghton Mifflin Company, Boston, MA. Persson, L., and P. Eklov. 1995. Prey refuges affecting interactions between piscivorous perch and juvenile perch and roach. Ecology 76:70-81. Pyke, G.H. 2005. A review of the biology of Gambusia affinis and G. holbrooki Reviews in Fish Biology and Fisheries 15:339-365. Rahel, F.J. 2002. Homogenization of freshwat er faunas. Annual Review of Ecology and Systematics 33:291-315.
71 Rauchenberger, M. 1989. Systematics and biogeography of the genus Gambusia (Cyprinodontiformes: Poecilidae). Amer ican Museum Novitates No. 2951:1-74. Relyea, R.A. 2003. How prey respond to combined predators: a review and an empirical test. Ecology 84:1827-1839. Rincon, P.A., A.M. Correas, F. Morcillo, P. Risueno, and J. Labon-Cervia. Interaction between the introduced eastern mosquito fish and two autochthonus Spanish toothcarps. Journal of Fish Biology 61:1560-1585. Rosen, D.E. 1979. Fishes from the uplands and intermontane basins of Guatemala: Revisionary studies and comparative geogra phy. Bulletin of the American Museum of Natural History 162:267-376. Rosen, D. E., and R. M. Bailey. 1963. The po eciliid fishes (Cyprinodoniformes), their structure, zoogeography, and systematics. Bulletin of the Amer ican Museum of Natural History 126:1-176. Rosen, D. E., and K. D. Kallma n. 1969. A new fish of the genus Xiphophorus from Guatemala, with remarks on the taxonomy of endemic forms. American Museum Novitates 2379:1-29. Rosenzweig, M.L. 2001. The four questions: what does the introduction of exotic species do to diversity? Evolutionary Ecology Research 3:361-367. Ruesink, J.L. 2005. Global analysis of factors a ffecting the outcome of freshwater fish introductions. Conservation Biology 19:1883-1893. Ruesink, J.L. 2007. Biotic resistance and fac ilitation of a non-native oyster on rocky shores. Marine Ecology Pr ogress Series 331:1-9. Savino, J. F., and R. A. Stein. 1982. Predator -prey interaction between largemouth bass and bluegills as influenced by simulated, submersed vegetation. Transactions of the American Fisheries Society 111:255-266. Savino, J.F., and R.A. Stein. 1989. Behavioral intera ctions between fish predators and their prey: effects of plant density. An imal Behavior 37:311-321. Schaefer, J.F., S.T. Heulett, a nd T.M. Farrell. 1994. Interactions between two poeciliid fishes ( Gambusia holbrooki and Heterandria formosa ) and their prey in a Florida marsh. Copeia 1994:516-520. Secor, D.H. 2007. The year-class phenomenon and the storage effect in marine fishes. Journal of Sea Research 57:91-103.
72 Semmens, B.X., E.R. Buhle, A.K. Salomon, and C.V. Pattengill-Semmens. 2004. A hotspot of non-native marine fishes: evidence for the aquarium trade as an invasion pathway. Marine Ecology Progress Series 266:239-244. Stachowicz, J.J., H. Fried, R.W. Osman, and R.B. Whitla tch. 2002. Biodiversity, invasion resistance, and marine ecosystem functio n: reconciling pattern and process. Ecology 83:2575-2590. Taylor, R.C., J.C. Trexler, and W.F. Loftus 2001. Separating the eff ects of intraand interspecific age-structured interactions in an experimental fish assemblage. Oecologia 127:143-152. Trexler, J.C., W.F. Loftus, and S.Perry. 2005. Disturbance frequency and community structure in a twenty-five year in tervention study. O ecologia 145:140-152. USGS (U.S Geological Service) 2007. Nonindigenous Aquatic Spec ies Database, Gainesville, FL. Available: http://nas.er.usgs.gov/queries/SpeciesList.asp? Group=Fishes (May 2008). Von Holle, B., and D. Simberloff. 2005. Ecological resistance to biological invasion overwhelmed by propagule pressure. Ecology 86:3212-3218. Werner, E.E., J.F. Gilliam, D.J. Hall, and G. G. Mittelbach. 1983. An experimental test of the effects of predation risk on habitat use in fish. Ecology 64:1540-1548. Wildgoose, W.H. 2001. BSAVA manua l of ornamental fish, 2nd edition. British Small Animal Veterinary Asso ciation, Gloucester, England. Wooten, M.C., K.T. Scribner, and M.H. Smit h. 1988. Genetic variability and systematics of Gambusia in the southeastern United States. Copeia (1988):283-289.
BIOGRAPHICAL SKETCH Kevin Allen Thom pson was born and raised in S outh Florida. Through constant pursuit of outdoor and aquatic activities, espe cially fishing and surfing, as well as early academic interests in science, he became determined to study the aq uatic environment. He went on to attend The College of William and Mary where he earned a B.S. in biology. While there, he was fortunate enough to work as a lab assistant on a project studying the population biol ogy of shark species in and around Chesapeake Bay. Upon graduation, Kevin worked as a field biologist for the Virginia Institute of Marine Science assessing pop ulations of juvenile s portfish in Virginias coastal rivers. Soon after, in the Fall of 2006, he began graduate work at the University of Florida, Department of Fisheries and Aquatic Sc iences. Upon graduation from UF, Kevin will be attending Oregon State University, pursuing a Ph.D. in the Department of Fisheries and Wildlife.