We investigated the seasonal habitat selection of Mosquitofish to increase our understanding of the ecology and invasiveness of this species. Sampling was conducted during the reproductive and non-reproductive season of Mosquitofish in four habitat types: Alternanthera philoxeroides habitat, Typha angustifolia habitat, Paspalum distichum habitat, and no vegetation habitat. Mosquitofish catch per unit effort was significantly lower in Typha habitat than in the other three habitats during the reproductive season. T. angustifolia can exude allelopathic chemicals and have the potential to negatively influence western Mosquitofish abundance. In the non-reproductive season, catch per unit effort of Mosquitofish were significantly higher in Alternanthera habitat than in the other three habitats. These results suggest that mats of A. philoxeroides provided an insulating layer for Mosquitofish overwintering. Therefore, Mosquitofish could expand its range further north in China as A. philoxeroides spreads northward with climate change. We suggest that an effective way of controlling western Mosquitofish is through eradication of the invasive macrophyte (A. philoxeroides) stands while retaining and restoring more native emergent macrophyte (T. angustifolia) stands; this strategy could limit overwintering habitat for western Mosquitofish and may draw them into open water habitats where they can be more easily removed.
Biological invasions are recognized as a major threat to global biodiversity and ecosystem function (Everett, 2000). Freshwater ecosystems are particularly affected by non-native species (Gherardi, 2007; Gozlan et al., 2010). In aquatic systems, fish are the largest group of introduced aquatic animals in the world (Gozlan, 2008), and once they are established, they are considered important contributors to species extinction (Vitule et al., 2009; Cucherousset and Olden, 2011). Information about habitat use by non-native freshwater fish can help with predicting and controlling non-native species establishment and spread. Hence, field studies need to pay more attention to habitat selection by non-native freshwater fish (Pyke, 2008).
The western Mosquitofish (Gambusia affinis, Baird and Girard, 1853) is a small fish species originated from North America. Both the western Mosquitofish and its close relative (G. holbrooki) were introduced into fresh and saline aquatic systems on all continents except the Antarctic, as biological mosquito control agents (Pyke, 2008). Now, they are the most widely distributed freshwater fishes in the world (Pyke, 2005). Western Mosquitofish is recognized as a serious threat to aquatic biodiversity because of their broad tolerance to variable environmental factors, extraordinary predation ability and high level of competitiveness (Pyke, 2005; 2008; Karraker et al.,2010; Xu and Qiang, 2011). Previous studies revealed that the invasion of Mosquitofish not only threatens invertebrates (Margaritora, 1990; Leyse et al., 2004), but also some small native fish (Goren and Galil, 2005; Ayala et al., 2007), and even amphibians (Gamradt and Kats, 1996; Goodsell and Kats, 1999). Due to its strong invasiveness, wide distribution, and strong negative impacts, western Mosquitofish waslisted in the Global Invasive Species Database as one of the 100 world’s worst invasive species (Lowe et al., 2000).
Western Mosquitofish was firstly introduced to Chinese Taiwan in 1911, and then to the Chinese mainland (Shanghai) in 1924. From the 1950s, western Mosquitofish was introduced to almost all of China (the Pearl River basin, Yangtze River basin, and Yellow River basin) for mosquito control (East China Sea Fisheries Research Institute, 1990). Established populations of western Mosquitofish have been discovered in many waterbodies in the south of China (Liu et al., 2017; Cheng et al., 2018; Xiong et al., 2018a; b; c; 2019). Recently, an increasing number of investigations show that the western Mosquitofish has expanded further north and established feral populations in the middle reach of the Yangtze River (Cheng et al., 2018). The climate in the middle reach of the Yangtze River is not expected to support Mosquitofish because of the low winter temperatures. Diverse aquatic macrophyte species in this region, with distinct morphological characteristics and life strategies (Fang et al., 2006), may create enough suitable habitats to facilitate the invasion of western Mosquitofish. However, few studies have investigated habitat selection by western Mosquitofish in invaded ecosystems, which could be a key factor underlying their successful invasion (Pyke, 2005; 2008).
In this study, we investigated the habitat selection by western Mosquitofish in a typical wetland of the central Yangtze River during the reproductive and non-reproductive seasons. Two basic questions were proposed: 1) what habitats does the western Mosquitofish prefers to occupy? 2) Is there any seasonal variation of habitat selection by western Mosquitofish? Answering these two questions could contribute to western Mosquitofish management and eradication in this area, as well as prevent further spread.
Materials and methods
The study was conducted at the Xixianling wetland (30°28′29.2″N, 114°22′40.96″E), which is on the south bank of the middle reach of the Yangtze River, Hubei Province, China. The Xixianling wetland, which connected the Yangtze River all year, is a representative wetland in the middle of the Mosquitofish distribution in the Yangtze River. Some small fish species (such as Oryzias latipes, Rhinogobius giurinus, Rhodeus ocellatus, Pseudorasbora parva, Hemiculter leucisculus, Hypseleotris swinhonis) were found in area of the Xixianling wetland (Xiong et al., 2015b; Cheng et al., 2018). The total area of Xixianling wetland is about 0.7 square km, with a mean depth of 0.5 m (range0.2-0.8m). Based on the composition of the dominant macrophytes, the shoreline of the wetland was divided into four habitats: Alternanthera philoxeroides habitat (AP habitat), Typha angustifolia habitat (TA habitat), Paspalum distichum habitat (PD habitat) and no vegetation habitat (NV habitat). Each macrophyte species has a unique life strategy and architecture, representing a distinct habitat type. The four different vegetation types (AP, TA, PD, and NV) have divided the shoreline of Xixianling wetlands into small patches called plaques. The size of each plaque varied from several square meters to hundreds of square meters. A. philoxeroides is a non-native perennial herb, rooted along banks or shallow water and extends hollow stems over the water surface to form mats. In summer and autumn (reproductive season for western Mosquitofish), mats of A. philoxeroides continue to develop to a thicknesses of 10-30cm. Mats of A. philoxeroides were maintained throughout the whole winter (non-reproductive season for western Mosquitofish), but the thicknesses of mats decrease to 5-15cm and no living leaves or roots were observed on the stems. P. distichum is a native perennial grass that forms mats like A. philoxeroides in summer and autumn, but the mats disappear when the emergent part dies in winter. T. angustifolia is a rooted emergent evergreen plant, with simple structure.
Fish sampling was carried out during the reproductive season (September 2012) and non-reproductive season (January 2013) of western Mosquitofish. For each habitat type, a dip net (0.5m in diameter and a stretched mesh size of 1mm) was used to sample fish; we sweep the whole square meter unit with the half meter net and the unit of CPUE is fishes square meter−1. Distance between sample sites in each habitat type was at least 20m in different vegetation area to minimize the effects of disturbance on fishes. Hence, each trapping procedure was considered as an independent sample. A total of 96 samples were collected, distributed in three periods: morning (06:00-07:00h), noon (14:00-15:00h) and evening (22:00-23:00h), with four replicates at each period in each habitat type. Fish were preserved in 4% formalin solutions for transport to the Laboratory of Biological Invasion and Adaptive Evolution, Institute of Hydrobiology, Chinese Academy of Sciences, where they were counted and identified.
Measure of samples in laboratory
In the laboratory, all fish were identified to species level. The sex of western Mosquitofish was determined using the morphology of the anal fin and gonads examination. Mosquitofish were classified as “male” if they possessed a fully developed gonopodium, and as “female” if they were larger than the smallest male and had no gonopodium. If the sex of Mosquitofish could not be determined through gonopodium, it was processed through the direct observation of gonads. When it was not possible to discern the sex, Mosquitofish were classified as “juvenile”.
Native small fish (Oryzias latipes, Rhodeus ocellatus) occurred sparsely in this study (3.12% of samples and 0.2% of the specimens), while Mosquitofish was dominant in the Xixianling wetland (99.8%). Therefore, only the results for Mosquitofish were considered in this study. To test the effects of habitat type and season on Mosquitofish abundance (CPUE), we applied generalized linear models (GLMs; with poission distribution and logarithmic link). The CPUE of Female, Male, Juvenile, and Total fish was treated as the response variable, respectively. Habitat type and season were used as fixed effects, and the interaction habitat type × season was included. For each season, we conducted Kruskal-Wallis test to compare among habitat types. Where results from the Kruskal-Wallis test were significantly different, Mann-Whitney tests were used for multiple comparisons. All statistical analyses were conducted in SPSS 19.0 at significant level of 0.05.
A total of 1928 Mosquitofish (1257 females, 552 males, and 119 juveniles) were caught and measured. Mosquitofish (female, male, and total) CPUE differed significantly among habitats and seasons, and their interaction effect was significant (Table 1). Juvenile differed among habitats or seasons, but the interaction effect was not significant (Table 1).
Mosquitofish (female, male, juvenile, and total) CPUE differed significantly among habitats during the reproductive season (September) and the non-reproductive season (January) (GLMs, September: Female: X2=25.563, p < 0.0001; Male: X2=20.365, p < 0.0001; Juvenile: X2=8.104, p = 0.044; Total: X2=26.997, p < 0.0001. January: Female: X2=12.058, p = 0.007; Male: X2=17.726, =0.001; Juvenile: X2=8.228, p = 0.042; Total: X2=19.093, p < 0.0001). During the reproductive season, Mosquitofish (female, male, juvenile, total) CPUE were significantly higher in PD, AP, and NV habitats than TA habitat (Fig. 1). During the non-reproductive season, Mosquitofish (male and total) CPUE were significantly higher in the AP habitat than in the other three habitats (Fig. 2). Also, there were no significantly different characteristic variables found among different habitats within both reproductive and non-reproductive seasons (see variables listed in Table 2).
By investigating four different mono-specific macrophyte stands, we found that western Mosquitofish showed avoidance of TA habitat during the reproductive season (Fig. 1), and preference for AP habitat over other habitats during the non-reproductive season (Fig. 2). Although previous studies have explored habitat preference of Mosquitofish in experiments (Casterlin and Reynolds, 1977), no information was available about the habitat preference of western Mosquitofish in the field, and across different seasons. To our knowledge, this study is the first attempt to investigate the relationship between plant-specific habitats and western Mosquitofish distribution in the central Yangtze River area. In general, many previous studies revealed fish CPUE were greater in vegetated habitats than in non-vegetated habitats in North and South America (Pelicice et al., 2005; Strakosh et al., 2009; Collingsworth and Kohler, 2010). Similarly, studies also found that Mosquitofish prefer vegetated habitats rather than non-vegetated habitats (Moyle and Nichols, 1973).
Some variations in habitat selection were found for western Mosquitofish in our study area. During the reproductive season, our investigation showed that western Mosquitofish prefers to utilize PD, AP and NV habitats over TA habitat. The habitat preference for floating mat (PD and AP Habitats) and no vegetation cover (NV habitat) over emergent plant (TA habitat) by Mosquitofish is not fully consistent with previous studies. Casterlin and Reynolds (1977) reported that Mosquitofish prefer submersed plant habitat to no vegetation cover habitat, and prefer no vegetation cover habitat to floating plant habitat in aquarium experiment. Nevertheless, our results showed that Mosquitofish preferred floating plant habitats (AP and PD habitats) and no vegetation cover habitat (NV habitat) rather emergent vegetation habitat (TA habitat). This is probably due to the fact that TA habitat provides nesting habitats for many birds, which may exert great predatory pressure on fish in this habitat (Polak, 2007; Zhang et al., 2012), leading Mosquitofish to avoid it.
In the non-reproductive season, our results are consistent with some analogous studies that reported lower densities of Mosquitofish during the winter (Howell et al., 2013). The Xixianling wetland is located in subtropical monsoon climate, with a mean temperature of 28.7 °C in summer and 0.4 °C in winter. Mosquitofish is a small fish that prefers warm water, and mortality of Mosquitofish is higher during winter than in warmer seasons (Haynes, 1993). In the past, many studies considered that feral populations of western Mosquitofish were distributed only in southern China due to limiting cold temperatures further north (Dudgeon and Corlett, 2004; Cheng et al., 2018). In this study, CPUE of Mosquitofish in AP habitat was significantly higher than for the other three habitats (Fig. 2), probably because AP habitat provides cover for western Mosquitofish from predators during the winter. Specifically, western Mosquitofish have lower swimming performance and thus a lower ability to evade endothermic predators in the winter, such as birds and mammals. In this study, we observed many water birds in the Xixianling wetland, and the mats of A. philoxeroides could efficiently protect western Mosquitofish from predation by water birds. It is well known that A. philoxeroides is one of the worst invasive weeds in China, dramatically affecting biodiversity and causing great loss in agriculture and aquaculture (Pan et al., 2007). In the past, some researches considered that A. philoxeroides could only occur at latitudes below 31.5°N in China (Julien et al., 1995). However, the fact is that A. philoxeroides has expanded its range north to 36.8°N under the background of global climate change (Lu et al., 2013). Based on our investigation, we found that with the assistance of non-native aquatic plants (A. philoxeroides) providing habitats for western Mosquitofish, it is very likely for western Mosquitofish to break this limitation and gradually spread their distribution after A. philoxeroides spreads further north due to climate change.
In conclusion, we found higher CPUE of Mosquitofish in A. philoxeroides stands than other habitats in winter. These findings suggest that the invasive aquatic plant (A. philoxeroides) may facilitate invasive Mosquitofish (G. affinis) overwintering in higher latitudes, such that it could expand its range further north in China following the spread of A. philoxeroides. The facilitative interactions between non-native aquatic plants and fish in the central Yangtze River has the potential to cause an invasion meltdown in the ecosystem (Simberloff and Von Holle, 1999). Currently, China has the highest number of non-native freshwater fishes and aquatic plants in the world (Xiong et al., 2015a; 2017; Wang et al., 2016). More non-native aquatic species are likely to be introduced in the future. Therefore, further studies should focus on interactions between these non-native aquatic species. Meanwhile, we found that western Mosquitofish avoid T. angustifolia habitat in the reproductive season, but prefer A. philoxeroides habitat to other habitats during the non-reproductive season. These results provide insights on the role of different habitats during western Mosquitofish invasion and the management of both western Mosquitofish and A. philoxeroides. It is more effective to consider the two non-native species together, when we take actions, rather than separately. Based on our study results, we propose two suggestions here. Firstly, in the reproductive season of Mosquitofish (April-October), efforts should be focused on eradicating floating dense macrophytes (such as A. philoxeroides), and restoring T. angustifolia in wetlands, as this plant repels Mosquitofish to concentrate in open water areas. This should be followed by physical or biological control of Mosquitofish in open water areas. Secondly, it would be recommendable to eradicate A. philoxeroides to decrease suitable habitat for Mosquitofish in winter.
We gratefully acknowledge the help in the field and laboratory from Xiaoyun Sui, Yintao Jia, Juan Lei, Xi’ao Zhang, Dengcheng Zhang, and Chaojun Wei. We also thank Doctor Harmony Patricio for English help. This research was supported by the Second Tibetan Plateau Scientific Expedition and Research (STEP) program (Grant No. 2019QZKK0501), the National Natural Science Foundation of China (No 31472016).