Comparison of Behavioral Traits and Invasion Success Between Two Global Freshwater Fish Invaders—Gambusia holbrooki and Gambusia affinis

Abstract

In the early 1900s, eastern mosquitofish (Gambusia holbrooki) and western mosquitofish (Gambusia affinis) were deliberately and globally introduced for the biological control of mosquito larvae. Subsequently, both species developed a reputation for causing impacts on native small-bodied fish, amphibian larvae, and other aquatic species. This led to both species being considered some of the world’s worst invasive species. Due to morphological similarities, organizations worldwide often consider these species jointly when discussing their introduction and impacts. Recent studies suggest these species differ in fundamental ways, which could affect invasion success. Our goal was to compare eastern and western mosquitofish behavior and invasion success. Replicate populations were collected from the U.S. states of Florida (eastern mosquitofish) and Louisiana (western mosquitofish) to assess variation in aggression, boldness, and sociability. Mesocosm trials were used to compare invasion success between species following introduction to an ecosystem occupied by another small-bodied poeciliid. Eastern mosquitofish caused more damage to similar-sized heterospecifics and western mosquitofish exhibited greater boldness. No differences were found in sociability between the two species. In mesocosms, impacts were observed for both mosquitofish species but were greatest for heterospecifics with eastern mosquitofish. This suggests that two invasive species, even with similar life history and morphology, can differ in traits related to invasion success and ecological impacts. It is important to correctly identify mosquitofish species when evaluating their invasion.

1. Introduction

Biological invasions are one of the dominant causes of species extinction [1], while also causing economic impacts on par with natural disasters [2]. These impacts can vary among species, even for closely related species that are difficult to distinguish [3]. Thus, despite minimal interspecific variation, and even intraspecific variation, seemingly minor differences can lead to altered ecological outcomes [4]. Behavioral trait variation may be one of the dominant ways these minor differences manifest into alternate ecological outcomes. In fishes, traits associated with invasion success include aggression [5,6], boldness [7,8], and sociability [9,10]. Thus, interspecific and intraspecific variation in these traits could potentially alter ecological outcomes following introduction, potentially altering invasion success. These behavioral traits are thought to be partially responsible for the invasion success of two global invasive species, eastern mosquitofish (Gambusia holbrooki, Girard, 1859) and western mosquitofish (Gambusia affinis (Baird and Girard, 1853); Figure 1), two closely related species that are difficult to distinguish.

Eastern and western mosquitofish have been introduced globally for the biological control of mosquito larvae and are now established on six continents [11]. Although they are often intentionally introduced, numerous studies indicate that mosquitofish cause a decline in small-bodied fish [12,13,14] and amphibian larvae [15,16,17]. Due to the potential impacts on native species and their global distribution, eastern and western mosquitofish are considered model organisms in invasion ecology [18]. However, the importance of mosquitofish as model organisms can hinge upon correct identifications for global introductions and identifying intra- and inter-specific variation in traits affecting invasion success. Several global agencies, including the International Union for Conservation of Nature (IUCN), jointly list eastern and western mosquitofish as invasive due to their similar morphology [19]. That said, eastern and western mosquitofish do differ in fin ray counts [20], sex determination [21], and the physical structure of the gonopodium in males [22]. Such differences, however slight, may mask significant variation in behavior, possibly affecting invasion success [23].

Several studies indicate there are differences between mosquitofish species in terms of behavior. For example, western mosquitofish exhibit greater dispersal tendencies compared to eastern mosquitofish and both species exhibit unique behavioral plasticity in responding to a predator [24,25]. The cause of this behavioral divergence is unknown, and it is uncertain if this variation is influenced more by interspecific or intraspecific differences. Invasion success in mosquitofish is potentially influenced by this behavioral variation, and differences in outcome have been observed between mosquitofish species when in the presence of another globally distributed invader, the guppy (Poecilia reticulata). In the U.S. state of Florida, the native eastern mosquitofish limits the survival of non-native female guppies [26], but in Okinawa, Japan, non-native male guppies are believed to have caused population declines of non-native western mosquitofish through reproductive interference [27]. This outcome variation may portend differences in behavior between the two mosquitofish species, perhaps suggesting eastern mosquitofish are more invasive than western mosquitofish.

The goal of this work was to investigate how behavioral traits associated with invasion success may differ between two global invaders. We tested the hypotheses that these two species would differ in (1) aggression, boldness, and sociability using aquariums trials and (2) invasion success using replicated outdoor mesocosms where mosquitofish “invade” mesocosms occupied by a similar small-bodied livebearer species. Based on the outcomes of previous research, we expected that eastern mosquitofish would display more aggression and exhibit greater impacts and invasion success. We also expected western mosquitofish to be bolder and less social compared to eastern mosquitofish due to their tendency to disperse at greater rates [24], and because both traits (boldness and sociability) have been associated with dispersal in the western mosquitofish [28,29].

2. Materials and Methods

2.1. Fish Collection, Maintenance, and Experimental Design

Eastern and western mosquitofish were collected using butterfly nets (66.0 cm depth; 33.0 cm ring) in their native range from Louisiana to Florida, USA (Figure 2). Both species were collected along approximately the same latitude to mitigate potential variation in behavioral traits associated with latitude [30]. No fish were collected in Alabama or Mississippi because coastal areas of these states are a hybrid zone [31].

Approximately 100 individuals from each of eight populations of eastern mosquitofish and eight populations of western mosquitofish were collected in July 2021. Three populations of eastern mosquitofish and three populations of western mosquitofish were recollected from six of these sites in August 2022 if the stock population of mosquitofish was no longer available.

Collected mosquitofish were transported to the University of Florida, Institute of Food and Agricultural Sciences, Tropical Aquaculture Laboratory (TAL) in Ruskin, FL, USA. Each population was separately maintained in individual 151 L tubs, which were continuously aerated and used a flow-through system at 25 °C of degassed, well water. Water quality parameters were measured weekly, including dissolved oxygen (DO), pH, alkalinity, and nitrogenous compounds. Rearing parameters fell within acceptable ranges for both species [11]. Fish were fed at approximately 3% of their body weight with Zeigler^TM Tropical Flake (Garners, PA, USA) per day and the uneaten portion was siphoned daily.

All 16 populations were used for aggression trials after one week of acclimation to laboratory conditions. Eight populations, four of each species, supplied enough fish and were also used for mesocosm trials, which ran for seven weeks from October through November 2021. Populations were re-collected in August 2022 to replenish the stock and reduce the effects of long-term captive rearing [32,33]. Mosquitofish from these six populations, three of each species, underwent boldness trials in November 2022 and sociability trials in January 2023. The aggression and mesocosms trials used guppies to compare behavior and invasion success between mosquitofish species. The guppy was chosen for these experiments because the species is similar in body size to mosquitofish [34] and because of earlier work [26,27]. Guppies were a feeder variety maintained in outdoor ponds at the TAL and were collected one week prior to the start of any experiment to acclimate to laboratory conditions.

Behavioral assays (aggression, boldness, and sociability) were conducted between 10:00 and 17:00 h and took place in 38 L (50.8 cm length × 20.3 cm width × 30.5 cm height) aquaria filled with 22.0 cm of degassed well water. Each aquarium was continuously aerated and used a flow-through system with water at a relatively constant 25 °C. Aquaria had opaque screens covering three sides with one uncovered side. This allowed observation of the fish’s interactions while minimizing outside visual cues that may alter behavior. Aggression trials were completed following previously defined methods [26,35].

2.2. Aggression Trials

Aggressive behavior towards guppies was noted in aquarium trials using 16 populations, eight of both species of mosquitofish (Figure 3). Each tank contained artificial vegetation placed in the center of the aquarium, which consisted of eight, 3 cm wide strips of black plastic attached to a 5 cm length PVC pipe. Two male guppies (mean standard length (SL) ± SD = 16.8 ± 4.2 mm) were introduced to each tank and allowed to acclimate for 24 h prior to the mosquitofish introduction. Male guppies were used in the trial because they are more likely to exhibit aggressive behavior than females [36]. Each tank had either eastern mosquitofish (one of each sex; males = 17.5 ± 2.5 mm SL; females = 21.1 ± 2.5 mm) or western mosquitofish (males = 17.6 ± 2.4 mm; females = 20.1 ± 2.6 mm) from the populations collected. Standard length did not differ between mosquitofish species for males (F1,14 = 0.01, p = 0.919) or females (F1,14 = 0.61, p = 0.448) in the aggression trials.

Mosquitofish were introduced to aquaria with a transparent barrier separating guppies from mosquitofish but allowed visual interactions. Mosquitofish were given 5 min to acclimate before the barrier was removed and a 20-min observation period began when the mosquitofish resumed normal swimming behavior (~15 s). Aggressive nips, attempted nips, and chases were noted during the observation period for all tanks and these behaviors were observed on all species of fish [37]. Following the observation period, trials ran for seven days. If a fish did not survive the full seven days of the trial, it was removed from the tank. Fish were fed approximately 3% of their body weight in flake per day and the survival of all fish was recorded daily. At the conclusion of the experiment all individuals were humanely euthanized with tricaine methanesulfonate (Syncaine^® MS-222; Syndel, WA, USA), whereupon both mosquitofish and guppies were measured for SL and scored for caudal fin damage. Caudal fin damage was scored on an ordinal scale consisting of three categories: none (no visible damage), moderate (≤50% of the fin missing), and severe (>50% of the fin missing; [38,39]).

2.3. Boldness Trials

For both mosquitofish species, boldness was measured as time to emerge from a refuge into an open compartment in a new environment (Figure 4) [28]. Shorter latency to emerge from a refuge indicates greater boldness and has been used extensively in behavioral studies on both species [29,40,41]. Tanks were divided into two parts using two opaque barriers. The first barrier had a small door (4 cm length × 5 cm height) cut into it located 3 cm from the bottom of the tank, allowing fish access to the open compartment following removal of the second barrier. The refuge area was 1/3 of the total length of the tank and included a 15 cm black nylon brush placed on the bottom to provide additional refuge for the fish. Female eastern mosquitofish (n = 28; mean SL ± SD = 28.3 ± 3.4 mm) and female western mosquitofish (n = 35; 26.6 ± 3.1 mm) were individually introduced into the aquarium from three different populations for both species.

Mosquitofish were placed individually into the refuge area and allowed to acclimate for 5 min prior to removal of the second barrier. The trial began when the focal individual resumed normal swimming behavior (~15 s) and the fish’s movement was recorded with a GoPro Hero 8TM (San Mateo, CA, USA). Time to emerge from the refuge into the open compartment was used as data points for each trial and shorter time indicated higher boldness [29]. Mosquitofish were given 45 min to emerge from the refuge area and were measured after being anesthetized by a low dose (5 mg∙L⁻¹) of MS-222. Mosquitofish would then be returned to their original enclosures and trials were replicated a second time for each individual fish after 72 h of rest per population. These individuals were used for a second trial to examine whether the behavior is repeatable for individuals and increase the power due to the relatively small sample size.

2.4. Sociability Trials

The tendency for mosquitofish to shoal, or sociability, was measured by recording an individual’s time spent in a social zone near a shoal of conspecifics for both species (Figure 5). Tanks were divided into three sections using transparent dividers, with a large center compartment (25.8 cm length) and two smaller compartments (12.5 cm length) on both sides. A shoal of either 14 eastern mosquitofish or 14 western mosquitofish (7 males, and 7 females) were placed into one of the smaller compartments while the other remained empty as a control. The individuals in the shoal had no previous encounters with the fish tested. Female eastern mosquitofish (n = 21; 32.2 ± 2.57 mm) and female western mosquitofish (n = 31; 29.8 ± 2.64 mm) were individually introduced into the aquarium by placing the fish in a removable, opaque, cylinder that was positioned in the center of the tank that exceeded the water level in height. Eastern mosquitofish were larger compared to western mosquitofish by ~7% on average for boldness and sociability trials (F1,219 = 15.6, p < 0.001). Mosquitofish acclimated inside the cylinder for 5 min before the cylinder was removed and the trial began when the focal individual resumed normal swimming behavior (~15 s). Fish remained in the aquarium for 10 min and their time was recorded when they were in the social zone, or when their position was less than or equal to 2 cm (represented by a small mark on the aquarium) from the divider, separating the shoal from the focal individual [35]. Upon conclusion of the trial, fish were anesthetized, measured, and returned to their original enclosure. Trials were then repeated a second time with each individual fish after 72 h of rest per population.

2.5. Mesocosms

Twenty-eight oval 151 L mesocosms (102 cm length × 73.0 cm width × 33.7 cm height) were placed outside in partial shade and filled to a depth of 20 cm (114 L) using water sourced from a fishless pond at the TAL (Figure 6). Four populations of eastern mosquitofish (males = 21.6 ± 3.0 mm SL; females= 31.4 ± 3.3 mm) and four populations of western mosquitofish (males = 21.2 ± 3.9; females = 29.1 ± 3.1) were haphazardly selected out of the original 16 populations collected. Four populations from each species were replicated three times (total n = 24). The remaining four mesocosms served as a control, and did not include mosquitofish. This experiment was used to compare invasion success, or mosquitofish’s ability to successfully displace guppies. Fish were not measured before the trial to avoid handling stress.

Standard length did not differ among populations (F6,85 = 2.11, p = 0.060) nor between species for male mosquitofish (F1,49 = 0.19, p = 0.664) but did with female mosquitofish (F1,40 = 5.70, p = 0.021) with female eastern mosquitofish being ~7% larger compared to female western mosquitofish. Each mesocosm had 25 g Spanish moss, 16 g muddy substrate, 2 L wet leaf litter, and one small clump (~10 stems) of Chara spp. added to each tank to mimic a ditch environment, a typical ecosystem for both species of mosquitofish [42]. Twenty-four hours after the mesocosms were established, 18 guppies were stocked into each mesocosm. Six mosquitofish were introduced into each mesocosm 24 h after the guppies were introduced. All fish were introduced in an even sex ratio and trials ran for seven weeks from 6 October 2021 until 24 November 2021. At the conclusion of the trial, all surviving fish were measured, sexed, and scored for caudal fin damage using the same ordinal scale as the aggression trials.

2.6. Data Analysis

All analyses were completed using R (V. 4.2.1; R Core Team 2022). Initial attacks and fish survival in aquaria were analyzed with generalized linear models (glm) using the MASS package [43] and recorded as counts. Model factors for analyzing initial attacks included species identity (eastern or western mosquitofish), sex, and a species × sex interaction as fixed factors. All data were checked for overdispersion and zero-inflation [44]. Initial attacks on guppies within the observation period displayed overdispersion; thus, the quasi-Poisson and negative binomial models were fit and evaluated for model suitability. Akaike’s information criterion (AIC) and the residual deviance indicated the negative binomial model with a log link function displayed the best fit.

For the analysis of Guppy survival in the aggression trials, just one model term was included, species identity. The data did not display overdispersion, nor zero-inflation; therefore, Guppy survival was analyzed using the default dispersion parameter (1) and fit to the Poisson distribution. All mosquitofish survived to the conclusion of the aggression trials.

Caudal fin damage was analyzed based on an ordinal scale using the cumulative link mixed model (clmm) function in the ordinal package [45]. Species identity was used as a fixed factor and population as a random factor. Level of caudal fin damage was broken into three categories: none (no visible damage), moderate (≤50% of the fin missing), and severe (>50% of the fin missing).

Time to emerge from a refuge was analyzed using a linear mixed-effects model (lmer) within the lme4 package [46]. Residuals were checked for normality (Shapiro–Wilk test, [47]) and homogeneity of variance (Fligner–Killeen test, [48]). Boldness data did not display normality and was thus log transformed. Models contained the fixed effect of species, SL, and trial as well as species × SL × trial interaction; population nested within species was used as a random factor for linear mixed models.

Time spent in a social zone of conspecifics was analyzed using the same model function and factors as the boldness trials. Sociability data displayed normality and homogeneity of variance; thus, no transformation was required.

Survival in mesocosms was recorded as count data and analyzed using a Generalized Linear Mixed-Effects Models (glmer) with species identity as a fixed factor and population nested within species as a random factor. Guppy survival in the mesocosms displayed overdispersion and a negative binomial with a log link function displayed the best fit. Mosquitofish survival in the mesocosms was analyzed using the default dispersion parameter (1) and a Poisson distribution. Caudal fin damage was analyzed using the same scale as the aggression trials. Model factors for analyzing caudal fin damage in mesocosms included species, sex, species × sex, and population of mosquitofish nested within species. Analysis of variance (ANOVA) was used to assess differences in SL between the two species of mosquitofish for all individual experiments and to assess overall population differences in survival and caudal fin damage within mesocosms.

3. Results

3.1. Aggression Trials

Eastern mosquitofish initiated the highest total number of attacks at 130 (trial mean ± SD = 8.1 ± 19.4) followed by western mosquitofish at 61 attacks (3.8 ± 7.8). Guppies performed the least number of total attacks at 59 (1.8 ± 2.1). However, there was no difference in the number of attacks between mosquitofish species (z_2,58 = −1.30, p = 0.386) or between mosquitofish and guppies (z_2,58 = 1.85, p = 0.065). Male mosquitofish performed more attacks than female mosquitofish regardless of species identity (z_1,58 = 4.13, p < 0.001). A significant interaction was found between species of mosquitofish and sex, with male eastern mosquitofish preforming more attacks than male western mosquitofish (z_2,58 = −3.49, p < 0.001). Most guppies survived the full length of the trial with both species of mosquitofish: 75% (12/16) survival with eastern mosquitofish and 87.5% (14/16) survival with western mosquitofish. Mosquitofish species did not affect guppy survival in the aggression trials (z_1,30 = 0.56, p = 0.577). Guppies with eastern mosquitofish exhibited greater caudal fin damage compared to guppies with western mosquitofish (z_1,11 = −2.45, p = 0.014). All guppies with eastern mosquitofish had caudal fin damage by the end of the trial whereas 12.5% (2/16) of guppies exhibited unharmed caudal fins with western mosquitofish.

3.2. Boldness Trials

Eastern mosquitofish took longer to emerge from the refuge compared to the western mosquitofish (t_1,108 = −2.68, p = 0.009; Figure 7). Four eastern mosquitofish and five western mosquitofish never emerged from the refuge area and were excluded from the analysis. No effects of SL (t_1,108 = 1.59, p = 0.114) nor trial number (t_1,108 = −0.192, p = 0.848) were found on time taken to emerge from the refuge. No interaction between species and SL (t_1,102 = −0.44, p = 0.660), species and trial number (t_1,102 = 0.11, p = 0.910), nor SL and trial number (t_1,102 = −1.14, p = 0.256) was found.

3.3. Sociability Trials

Eastern and western mosquitofish displayed similar shoaling tendencies and no differences were found in sociability between species (t_1,100 = −1.03, p = 0.306; Figure 8). Larger mosquitofish spent less time shoaling with conspecifics regardless of species identity (t_1,100 = −3.60, p < 0.001; Figure 9). Mosquitofish spent, on average, one minute longer in the social zone in the first trial (trial mean ± SD = 378 ± 123 s) than in the second (323 ± 136 s; t1,100 = −2.24, p = 0.027) which is likely due to familiarity of the experimental set up. No interactions were found between species and SL (t_1,94 = −0.92, p = 0.361), species and trial number (t_1,94 = −0.90, p = 0.373), or SL and trial number (t_1,94 = −0.30, p = 0.767).

3.4. Mesocosms

Compared to the control treatment, overall guppy survival was lower with eastern mosquitofish (z_2,22 = −4.80, p < 0.001) and western mosquitofish (z_2,22 = −2.63, p = 0.008). However, eastern mosquitofish had a greater impact reducing overall guppy survival compared to western mosquitofish (z_1,19 = 2.57, p = 0.010; Figure 10). Of all surviving guppies, control mesocosms maintained a similar percentage of male (54.7%) and female (45.3%) guppies while mesocosms with eastern mosquitofish and western mosquitofish had higher proportions of female survivors (81.4% and 69.5%) compared to male survivors (18.6% and 30.5%). Population (nested factor) had no effect on guppy survival (F_1,16 = 1.68, p = 0.189). There was no difference in mosquitofish survival among treatments (z_1,20 = 1.14, p = 0.255); eastern mosquitofish averaged 56.9% (SD ± 27.9%) survival while western mosquitofish averaged 72.2% (±20.5%) survival in mesocosms.

Compared to the absence of mosquitofish, guppies with eastern mosquitofish exhibited greater caudal fin damage (z_2,168 = 2.03, p = 0.042) and guppies with western mosquitofish exhibited no difference in caudal fin damage (z_2,168 = −0.17, p = 0.866). Guppies with eastern mosquitofish displayed the lowest percentage of intact caudal fins and the highest percentage in the moderate category; all three treatments had similar percentages of severe caudal fin damage.

Population nested within species had no effect on caudal fin damage exhibited by guppies (F_1,117 = 1.73, p = 0.120). Sex was not a significant factor in the severity of caudal fin damage on guppies (z_1,168 = −0.10, p = 0.919) and no interaction was found between sex and eastern mosquitofish (z_2,168 = −0.18, p = 0.861) or western mosquitofish (z_2,168 = 1.23, p = 0.219). Most mosquitofish exhibited no caudal fin damage (eastern = 87.8%; western = 84.6%). No difference was found in caudal fin damage between mosquitofish species (z_1,90 = 0.613, p = 0.540).

4. Discussion

Eastern and western mosquitofish share many similarities in morphology and life history, which has led the two species to be combined when discussing their biology, ecology, and invasion history [19,49]. Due to their global distribution and impacts on native aquatic fauna, these two species are considered model organisms for invasion ecology [18]. However, variation in behavioral traits within this study may suggest that combining these two species is inappropriate. At the species level, eastern and western mosquitofish exhibited variation in impacts, boldness, and overall invasion success during mesocosm trials, while no differences were found among mosquitofish populations. This suggests that there are behavioral differences between the two species and highlights the importance of correctly identifying mosquitofish within invaded areas. These findings enhance our overall understanding of how behavior can impact invasion success and how behavioral traits could vary between closely related species [3].

4.1. Aggression, Boldness, and Sociability

Mosquitofish are well known to fin-nip and injure other fish species [12,13]. The 20-min observations indicated both species of mosquitofish readily attack similar, small-bodied heterospecifics; however, species identity did not affect the number of attacks. Most guppies housed with either mosquitofish species survived the full length of the aggression trials but were more damaged with eastern mosquitofish. This high survival could be due to the number of mosquitofish within the tanks, as mosquitofish can increase agonistic interactions at higher densities [39,50]. Further, if the trials were allowed to run longer, guppy survival could have been reduced as the sublethal effects of fin-nipping become lethal. The purpose of these aggression trials was to quantify individual mosquitofish aggression based on species and sex; hence, each tank contained only two mosquitofish of the same species and of each sex. Tuckett et al. [26] aggression trials resulted in much lower guppy survival (28%) when housed with four eastern mosquitofish, two males and two females. Due to the higher density of mosquitofish within that study, aggression may have been heightened, resulting in lower survival for guppies. This enhanced aggression at higher densities may be an indirect consequence of the mating system in mosquitofish, as males must compete to reproduce with females, which can increase agonistic interactions within a system [36]. This assumption is supported by these results, as male mosquitofish were more aggressive than females, regardless of species identity.

Bolder species and individuals may be more likely to spread and cause impacts [4]. Due to this, boldness has been routinely evaluated on both mosquitofish species [24,29,35]. Boldness can also influence reproductive fitness in eastern mosquitofish [40] and impact mate choice in western mosquitofish [41], both of which may influence the probability of establishing in a novel environment. Previous research has suggested that western mosquitofish have a greater tendency to disperse than eastern mosquitofish, but they exhibit similar boldness [24]. However, that research measured boldness as the average proportion of fish that dispersed out of a refuge within a population, rather than on an individual level. In this study, western mosquitofish were quicker to emerge from a refuge when introduced as individuals and were considerably bolder than eastern mosquitofish. These contrasting results are expected because mosquitofish shoal behavior can affect the boldness of an individual [29].

Increased sociability, or the amount of time spent near or in a shoal of conspecifics [51], can be beneficial for fish invaders becoming established in new habitats [3,23]. Shoaling in small fishes can enhance predator detection [52], have hydrodynamic advantages that can reduce energetic costs associated with movement [53], and increase foraging efficiency [54]. In this work, larger mosquitofish exhibited more asocial tendencies, regardless of species identity. This is likely because smaller individuals have greater shoaling tendencies to recognize and ultimately evade predators [55]. Additionally, sociability is an indicator of dispersal distance in western mosquitofish, with asocial individuals dispersing further distances [28]. Due to this, western mosquitofish were hypothesized to be more asocial compared to eastern mosquitofish [24]. Contrary to expectations, no differences in sociability were found between species. These results could be due to intraspecific variation among mosquitofish populations regarding social behavior. Variation in sociability within a population of invaders increases the speed of the invasion, with asocial individuals dispersing faster and facilitating the growth of social individuals that stay more localized [56]. Mosquitofish could display high polymorphism in sociability between and among populations, which may explain these results; however, more research needs to be conducted before this conclusion can be made.

4.2. Invasion Success

Outdoor mesocosms were used to assess the ability of eastern and western mosquitofish to invade and displace a population of heterospecifics in more natural conditions. Compared to the control, both mosquitofish species reduced guppies; however, guppy survival in the western mosquitofish treatment was almost doubled, on average, compared to guppy survival in the eastern mosquitofish treatment. The caudal fins of guppies housed with western mosquitofish were less damaged compared to guppies housed with eastern mosquitofish. We examined both species’ identity and population of origin on invasion success within experimental mesocosms and found differences only between species. However, increasing the number of collected populations and replicates could show population variation, as statistical power was relatively low, particularly if fish were collected across environmental gradients [30,57].

Mosquitofish were collected from multiple source populations to better represent the natural diversity of these two species. Mosquitofish were captured across a range of abiotic and biotic conditions, including lentic and lotic habitats, with variation in habitat complexity, presumably differences in predator regimes and fish communities, and water chemistry. Habitat complexity and predator cues can influence the behavior of eastern mosquitofish [39,58] and western mosquitofish [50,59], which could be important for understanding the extent of behavioral variation between species. All mosquitofish collected for these experiments were gathered from their native range and may exhibit behavioral variation from populations in non-native regions. This is supported by several studies, as population differences have been found for a variety of traits in mosquitofish, including social preferences [28,30,60] and antipredator responses [61,62]. This work supports the conclusion that interspecific variation in behavior is prevalent between mosquitofish species; however, additional work needs to occur to determine how source habitats of these populations may affect these traits.

The results from this work suggest eastern mosquitofish are more impactful as an invasive species compared to western mosquitofish, which may be consistent with the literature. In Australia, eastern mosquitofish have been shown to cause widespread negative impacts on fish [14,63,64,65] and amphibians [15,66,67]. In Europe, eastern mosquitofish are well documented as harming other aquatic species in Spain [1,68,69,70] and Italy [17]. In New Zealand and China, western mosquitofish impacts have been documented on fish [12,37,71] and amphibians [16,72] but the literature is less plentiful. Ling [73] has even questioned whether the western mosquitofish is ‘misunderstood’, since the species seems to have occupied a vacant niche in New Zealand that entails little to no competition with native fishes. Results from the aggression trials and mesocosms suggests both mosquitofish species can cause impacts, but the question remains as to whether eastern mosquitofish are a more problematic invader. This research, as well as previous work from Tsurui-Sato et al. [27] and Tuckett et al. [26] supports the suggestion that eastern mosquitofish are more impactful. This has yet to be determined though, as impacts are often combined, and uncertainty exists over species identification in many invaded areas.

5. Conclusions

This study identifies key differences between eastern and western mosquitofish that may influence their potential for invasion and ecological impacts. Eastern mosquitofish caused more caudal fin damage on guppies, while western mosquitofish displayed greater boldness. Both species affected guppies in mesocosm trials, but the impacts were most pronounced with eastern mosquitofish. These findings demonstrate that closely related species with similar life histories can differ in important behavioral traits which alter ecological impact. Identification to species is thus important when evaluating mosquitofish invasions, and the continued practice of lumping these two species together may obscure understanding of the global mosquitofish invasion. Although this study detected behavioral variation, further research with larger sample sizes is needed to clarify the extent and consequences of these differences.