Introduction

The pike-perch, Sander lucioperca, is a carnivorous fish species present in brackish-water and freshwater habitats in the temperate waters of western Eurasia. Among percids, this predator has become important for local fisheries (Popova and Sytina 1976; Balık and Geldiay 2002; Becer and İkiz 2007). However, its introduction has consequences for the native fishes. After its introduction into Turkish lakes, a general collapse in the native fish population was reported. During the course of its introduction, S. lucioperca were distributed to approximately 20 regions, including Antalya, Burdur, Big Menderes, Meriç-Ergene, Kızılırmak, Konya Closed, Marmara, Sakarya, and Seyhan stream basins (Tarkan et al. 2015). Researchers have investigated S. lucioperca both within their native and non-native areas (Campbell 1992; Nolan and Britton 2018). One of the outcomes of studies has been the reported loss of native fishes in Central Asian lakes, with the extirpations due, no doubt, to the high predation by pike-perch on zooplankton (Willemsen 1977; Peltonen et al. 1996; Lehtonen et al. 1996; Yılmaz and Ablak 2003; Specziár 2005; Balık et al. 2006; Kangur et al. 2007; Apaydın Yağcı et al. 2014). S. lucioperca were also reported to consume macroinvertebrates along with these prey items most frequently encountered in the diets of smaller individuals (Hansson et al. 1997; Argillier et al. 2012).

The euryhaline, big-scale sand smelt, Atherina boyeri, is a common fish species in freshwater ecosystems, coastal lagoons, and estuarine waters of the Mediterranean and Atlantic (Koutrakis et al. 2004; Leonardos and Sinis 2000; Doulka et al. 2013). A. boyeri were recorded in approximately 30 water bodies from Orontes River, Antalya, West Mediterranean, East Mediterranean, Big Menderes, Euphrates—Tigris, Gediz, Kızılırmak, Konya Closed, North Aegean, Marmara, Sakarya, Seyhan, and Yeşilırmak stream basins (Tarkan et al. 2015; İlhan and İlhan 2018; Çevik et al. 2018). Since its introduction into freshwater lakes, it has developed dense populations. With food preferences for small crustaceans that are both planktonic and benthonic, A. boyeri is an opportunistic carnivore (Vizzini and Mazzola 2005). It has become acclimatized and is expanding its range, so it is now occupying the ecological niches of other pelagic fish species in Trichonis Lake (Chrisafi et al. 2007).

P. parva was recorded in Romania in the early 1950s (Wildekamp et al. 1997; Kotusz and Witkowski 1998). Subsequently over the next 40 years, it spreads into the River Danube and other European countries (Banerescu 1999; Gozlan et al. 2010; Grabowska et al. 2010). Currently, P. parva has been recorded in approximately 70 water bodies including Meriç-Ergene, Marmara, Susurluk, North Aegean, Gediz, Big Menderes, West Mediterranean, Antalya, Sakarya, West Blacksea, Kızılırmak, Konya Closed, East Mediterranean, and Ceyhan stream basins in Turkey (Erk’akan 1984; Barlas and Dirican 2004; Yeğen et al. 2015; Özcan and Tarkan 2019; Ozulug et al. 2019). While P. parva mainly feeds on zooplankton such as Cladocera, Copepoda, and Rotifera, Bacillariophyta and Cyanobacteria were identified in high proportions in the foregut contents of this species by (Yalçin-Özdilek et al. 2013). After the introduction of pike-perch in 1955, irreversible damage arose in the fishery of Lake Eğirdir. Additional changes occurred when the silver crussian carp (in 1996), big-scale sand smelt (in 2003), and P. parva (in 2011 and 2014) were introduced into the lake (Küçük et al. 2009; Yerli et al. 2013; Yağcı et al. 2014; Yeğen et al. 2015).

To understand the impact of these three invasive fish species in Turkish inland waters, we studied their feeding behavior and prey selection in Lake Eğirdir. Thus, the specific objectives of the study were to compare the dietary composition of three invasive species and to document their feeding preferences.

Materials and methods

Study area and sampling

Lake Eğirdir is located in southern Anatolia and covers a total area of about 47.250 ha, with a maximum depth of 13 m at 918 m altitude (Yarar and Magnin 1997). Pike-perch and big-scale sand smelt were collected monthly from January 2010 to December 2010. At four stations of Lake Eğirdir, (Turkey) by Using gillnets of mesh sizes and purse-seine of 6, 10, 16, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 80, and 100 mm. Gillnet nets were set in the evening and retrieved from the water in the morning. Thus, we collected fish over a period of approximately 12–15 h. Also, we collected samples of topmouth gudgeon at the same four stations between March 2010 and June 2011 (Fig. 1). A total of 241 pike-perch, 612 big-scale sand smelt, and 88 P. parva individuals were collected monthly at four stations (Fig. 1). The stomachs of the specimens were immediately preserved in a plasticbucket containing 4% formalin (Buijse and Houthuijzen 1992), and their contents were analyzed in a laboratory. The fork length (FL) of species was measured to the nearest millimeter and weighed (W) to the nearest gram. Percentage number and frequency of occurrence were used to estimate the dietary selection of each prey category. Frequency of occurrence (Fpi): (N1i/Np), the percentage of the IRI (Pinkas et al. 1971) was used to express prey selection: IRIi = (Ni % + Wi%)* Oi%. To estimate prey preference of fish species, the prey selection index V proposed by Pearre (1982) was calculated. The index was determined as follows:

$$Va = \frac{(ad*be) - (ae*bd)}{{\sqrt {(a*b*d*e)} }}$$

where Va is Pearre’s index for three invasive species selection of species a, ad is relative abundance of species a in the diet, be is the relative abundance of all other species in the environment, ae is the relative abundance of species a in the environment, and bd is the relative abundance of all other species in the diet. a = ad + ae, b = bd + be, d = ad + bd, e = ae + be. The selection index (Va) was statistically tested using the Chi-squared test. (χ2 = n * V2). Where, n = ad + ae + bd + be.

Fig. 1
figure 1

Study area and sampling sections

Full size image

Results

The size and weight composition

We examined 612 stomachs of big-scale sand smelt. Of these, the stomachs of 563 (91.99%) individuals were full and 49 (8.01%) were empty. A total of 241 stomachs of pike-perch were examined, of which 162 (67.2%) and 79 (32.8%) were full and empty, respectively. A total of 88 stomachs of topmouth gudgeon were examined of which 65 (73.8%) and 23 (26.2%) were full and empty, respectively (Table 1).

Table 1 Number of invasive fishes caught in Lake Eğirdir, 2010: fork length (FL), weight (W)
Full size table

Diet composition of invasive fishes

In this study, we compared the stomach contents composition of three invasive fish species in Lake Eğirdir using relative importance index (IRI %) and frequency of occurrence (O %). According to the relative importance index, Pike-perch preferred A. boyeri (67.2%) with about 19.6% of their diet being the remains of other organisms. P. parva, preferred mainly Chironomus sp., (60.8%), N. hibernica (17.4%), and C. curvispinum (14.8). According to the frequency of occurrence, sand smelt predominantly preferred remains of other organisms (61.1%), C. curvispinum (42.8%), B. longirostris (38.5%), N. hibernica (30.2%), remains of Arthropoda (30.02%), M. leuckarti (20.4%), remains of insects (18.5%), and A. quadrangularis (10.1%) (Table 2). The diet of pike-perch in Lake Eğirdir included fish species, insects, and other organisms. Fish were found in the stomachs of 97 pike-perch; these included Aphanius iconii, Knipowitschia caucasica, Pseudophoxinus egridiri, Pseudorasbora parva, Seminemacheilus ispartensis, A. boyeri, and Carassius gibelio. The index (IRI) indicated that fishes ad prey had a greater importance (80.30%) than the other prey categories, i.e., remains of organisms (19.60%) and insects (0.09%). Phytoplankton, Arachnida, fishes, Annelida, and unidentified organisms showed a minor role in the A. boyeri diet. The diet of P. parva comprised phytoplankton, zooplankton, Insecta, Arthropoda, Annelida, and unidentified eggs. IRI found that Insecta as prey (64.71%) had more importance than the zooplankton (20.05%) and Chelicorophium curvispinum (14.78%). Chironomus sp. had the highest index value (IRI = 60.80%) followed by N. hibernica (IRI = 17.40%). In general, Bosmina longirostris was the dominant prey in the diet of sand smelt in Lake Eğirdir. The frequency of occurrence of B. longirostris was the highest (38.54%), followed by N. hibernica (30.20%), and Mesocyclops leuckarti (20.43%). C. curvispinum were ingested by big-scale sand smelt (42.81%), while remains of insects (18.47%) and unidentified organisms also were present in the diet (61.10%). Phytoplankton, Arachnida, fishes, and unidentified organisms played a minor role in the A. boyeri diet. (Table 2).

Table 2 Diet composition of invasive fishes in Lake Eğirdir, Turkey (O): frequency of occurrence, and (IRI): relative importance index (main groups, titles and important numerical values are shown in bold)
Full size table

According to Pearre's (1982) prey selection index (V), Atherina boyeri individuals exhibited positive selection to B. longirostris, A. Quadrangularis, and N. hibernica, but V was negative for A. boyeri, A. anatoliae, Insecta, Keratella cochlearis, Trichocerca similis, and Copepod Nauplii larva (Apaydın Yağcı et al. 2018a). While the most abundant prey species in the lake (Chironomus sp.) was ingested by P. parva (Apaydın Yağcı et al. 2018b), this food item this was not a statistically significant component of its diet. Also, while Alona guttata, A. quadrangularis, C. curvispinum, Coronatella rectangula, Graptoleberis testudinaria, M. leuckarti, and nauplii were present in the ecosystems, these animals were not chosen by topmouth gudgeon. Likewise, Pediastrum sp. Additionally, N. hibernica was also not preferred in the diet of topmouth gudgeon despite their high abundance in the lake Eğirdir ecosystem (Apaydın Yağcı et al. 2018b). According to the prey selectivity index of pike-perch, the prey P. parva, S. ispartensis, A. boyeri, and Calopteryx splendens were positive, but their selection indices were not statistically significant. A. anatoliae and Chironomus sp. were a common prey in the lake, but negatively selected by the pike-perch. P. egridiri and C. gibelio were negatively selected; their selection indices were not statistically significant (Apaydın Yağcı et al. 2014).

Discussion

The diet of A. boyeri was comprised primarily of zooplankton, followed by Arthropods and other food items such as insects, phytoplankton, fishes, and arachnids. Compared to other studies (Bartulović et al. 2004a; Doulka et al. 2013) reporting that sand smelt is an opportunistic predator, our study indicated that sand smelt fed mainly on planktonic and benthic invertebrates, Bartulović et al. (2004a) showed copepods (45%), gammarids, and amphipods (34%) to be the dominant prey of A. boyeri in the Mala Neretva River. Similar to our study, Doulka et al. (2013) identified a positive selection for zooplankton in Lake Trichonis, whereas Aphanius anatoliae were negatively selected by sand smelt in Eğirdir Lake, although they were the most abundant prey fishes (Apaydın Yağcı et al. 2018a).

Approximately 67% of S. lucioperca individuals had full stomachs. Studies on the diet of pike-perch from different regions show that the prey are similar (Table 3).

Table 3 Comparison of food items in various pike-perch population (main groups and titles are shown in bold)
Full size table

P. parva had a diet comprising mainly of Chironomus sp., N. hibernica, C. sphaericus, and B. longirostris (Apaydin Yağcı et al. 2018b). We found that the diet of topmouth gudgeon in Lake Eğirdir was dominated by Chironomus sp. Results of our study are similar to Wolfram-Wais et al. (1999) who stated according to the IRI index that Chironomus spp are one of the most important food item in Neusiedler See (Austria). Hliwa et al. (2002) showed that the diet of P.parva in the Balaton Reservoir was composed of Bosmina sp., Chydorus sp., Copepoda, and Daphnia sp. Besides, P.parva from Turkey in Gelingüllü Reservoir fed mainly on Cyanobacteria, Insecta, and Cladocera (Yalçin-Özdilek et al. (2013). The role of the Oligochaeta, Ephemeroptera, Copepoda/Calanoida, Trichoptera, and Nematoda were reported as important food resources for P. parva by Nikolova et al. (2008), whereas eggs of native fishes and their larvae were reported by Gozlan et al. (2010) in China and Germany. Likewise, the impact of P. parva on zooplankton and components of the zoobenthic community was also reported by Musil et al. (2014). Didenko and Kruzhylina (2015) showed using Ivlev’s selectivity indices that P.parva positively selected zooplankters such as C. sphaericus, Alona affinis, Pleuroxus sp., and Cyclops sp., but avoided Bosmina sp. A. priodonta was also positively selected among rotifers. Prey selection indices showed that D. cucullata, B. longirostris, Annelida, Trichoptera larvae, and Gomphonema sp. were positively selected by P.parva in Lake Eğirdir. Also, their selection indices were statistically significant. Pearre’s selectivity indices indicated that A. guttata, C. rectangula, Chironomus sp., C. curvispinum, G. testunidaria, and M. leuckarti were neutrally selected (Apaydin Yağcı et al. 2018b).

The purpose of our study was to assess the impact of three invasive fish species on Lake Eğirdir as a proxy for other inland Turkish waters. We found that the dominant prey items of big-scale sand smelt and topmouth gudgeon were zooplankton, insects, and arthropods, and the diet of pike-perch often included fish species as prey. These invasive species pose a danger for native fishes and care should be taken to prevent their introduction to other Turkish waters.