DIET AND PREY SELECTION OF PIKEPERCH ( SANDER LUCIOPERCA LINNAEUS , 1758 ) POPULATION IN LAKE EĞİRDİR ( TURKEY )

The diet and prey selection of pikeperch (Sander lucioperca L., 1758) were assessed by determining the frequency of occurrence, numeric and weight percentages, and the index of relative importance (IRI%) between January 2010 and December 2010 in Lake Eğirdir, Turkey. Pearre’s index was used to estimate diet selectivity, while the Schoener Overlap Index was utilized to compare diets. The stomach contents of 241 S. lucioperca were analyzed. Pikeperch diet included prey fish, insects and other organisms. The diet was predominantly fish, consisting of Atherina boyeri, Knipowitschia caucasica, Aphanius anatoliae, Seminemacheilus ispartensis and Carassius gibelio. A. boyeri was the most abundant prey fish in the lake; it was a positively selected food item (V= 0.130, X2= 3.359, p> 0.05) and was not statistically significant. C. gibelio also inhabits the lake, but was not preferred by pikeperch (V=0.134, X2= 3.582, p> 0.05). In addition, A. anatoliae (V=-0.223, X2=9.977, p<0.01) and Chironomus sp., (V= -0.297, X2= 17.665, p<0.01) were negatively chosen by pikeperch despite their high abundance in the lake. Stomach fullness was highest in January, while feeding density was lowest in November and >50 cm in length pikeperch. Cannibalism was not evident during a decade, due to there being enough food for pikeperch in the lake.


INTRODUCTION
The pikeperch, Sander lucioperca, also known as Zander, is a species of fish from freshwater and brackish-water habitats in western Eurasia.Sander lucioperca inhabit Turkish inland waters; the species has a wide distribution that includes Lake Mermere, Lake Beyşehir, Lake Eğirdir and Hirfanlı Dam Lake in Turkey (Balık and Geldiay, 2002;Yılmaz and Ablak, 2003;Balık et al., 2006;Apaydın Yağcı et al., 2006).It can be described as a warm-water species with preferred temperature from 24 up to 29°C, and it is important both in recreational and commercial fisheries (Lappalainen, 2001;Kangur et. al., 2007).Among the percids, Sander lucioperca is an ecologically sig-nificant predator in the temperate waters of Eurasia and is of importance to fisheries (Popova and Sytina, 1976).Pikeperch has been managed and stocked in Europe in order to regulate forage fish stocks (Peltonen et al., 1996).It usually plays a crucial role in eutrophic inland water in the reduction of planktivorous and omnivorous fish abundance (Frankiewicz et al., 1999).Various studies about the feeding ecology of pikeperch are available in the world (Willemsen, 1977;Peltonen et al., 1996;Lehtonen et al., 1996;Yılmaz and Ablak, 2003;Specziár, 2005;Balık et al., 2006;Kangur et al., 2007) Pikeperch often become piscivorous in their first summer, although fish may constitute a consid-erable proportion of the diet already when the pikeperch is 20-30 mm in length (Lehtonen et al., 1996).Sander lucioperca is non-native in Lake Eğirdir.Ten thousand pikeperch fingerlings were introduced by the then Hydrobiological Institute of the University of Istanbul to the lake in 1955 (Campbell, 1992).When the pikeperch were introduced, there were 10 species in Lake Eğirdir.It reproduced rapidly and became the dominant fish species in the lake.At the beginning of the 1990s, the silver crussian carp, Carassius gibelio were introduced into the lake, followed later by tench Tinca tinca in 1996 and sand smelt Atherina boyeri in 2003 (Balık et al., 2004;Balık et al., 2006;Çubuk et al., 2006;Küçük et al., 2009).These introduced fish species naturally affected the populations.Sander lucioperca directly affected the fish community structure though the effects of predation.The food and feeding habits of pikeperch in Lake Eğirdir were thoroughly investigated by Campbell in 1992; later data are scarce (Becer and İkiz, 1997;Ekmekçi and Erk'akan, 1997;Balık et al., 2006), as is information on the feeding ecology of Sander lucioperca populations in Turkey.For this reason, this research investigated the diet and prey selectivity of pikeperch in Lake Eğirdir, compared the IRI values obtained from pikeperch of different size-classes, and compared these results with other data.

MATERIALS AND METHODS
Investigations were conducted at Lake Eğirdir, which is situated in Southern Anatolia.Its maximum depth is 13 m.With its 47 250 ha surface area, lake Eğirdir is an oligotrophic lake, situated 918 m above sea level (Yarar and Magnin 1997).Pikeperch were collected from January 2010 to December 2010.A total of 241 individuals were caught monthly at four selected sampling sites (Fig. 1).Fishing nets with mesh sizes of 10,16,20,25,30,35,40,45,50,55,60,65,70,80 and 100 mm were used.All samples were carried out in the morning.The fork length (FL) of pikeperch was measured to the nearest millimeter and weighed (W) to the nearest gram.The stomachs of the specimens were immediately preserved in a plastic barrel containing 4% buffered formalin (Buijse and Houthuijzen, 1992;Balık et al., 2006), and their contents were analyzed in the laboratory.Prey items were counted and identified to the lowest taxonomic level using the identification keys of references (Geldiay and Balık 2002;Demirsoy 1997Demirsoy , 1998)).The contributions, by weight and number of a given prey category to each stomach content, and finally to the food of all pikeperch in the sample, were calculated (Hyslop, 1980).The fullness index (FI) was determined to investigate the variations in feeding intensity, using the equation: where, W SC is the weight of the stomach contents and W F is the total weight of the fish.Fish specimens were divided into five length groups (10-20 cm, 20-30 cm, 30-40 cm, 40-50 cm, >50 cm) to analyze their size-dependent dietary variations.The percentage of the IRI (Pinkas et al., 1971), and three dimensional graphical representation (Cortes, 1997) were used to express prey importance: where W i and N i are the total net weight and number of prey, and O i is the number of stomachs containing prey i.To estimate prey preference of pikeperch, the prey selection index V proposed by Pearre (Pearre, 1982) was calculated.The index was determined as follows:

− =
where Va is Pearre's index for pikeperch selection of species a, a d is relative abundance of species a in the diet, b e is the relative abundance of all other species in the environment, a e is the relative abundance of species a in the environment, and b d is the relative abundance of all other species in the diet: The selection index (Va) is statistically tested using the chi-squared test: (X 2 =n*V 2 ), where n=a d +a e +b d +b e .Diet similarity among length classes, months and seasons were estimated using the Schoener Overlap Index (C) (Schoener, 1970): where pxi and pyi are the proportions by number of prey type i in the diets of groups (length or season) x and y, respectively.If the C value is bigger than 0.80, the diet of the 2 groups is similar.

The diet composition of pikeperch
During the study, pikeperch ranged from 21.6 cm to 77.0 cm in fork length (FL) with a mean value of 28.90 cm, and their total weight ranged from 105 to  .The diet of pikeperch in Lake Eğirdir was comprised of food that included prey fish species, insects and other organisms (Table 1).Prey fish were found in the stomachs of 97 pikeperch (O = 59.88%,N = 97.29%),represented by A. anatoliae, K. caucasica, Pseudophoxinus egridiri, Pseudorasbora parva, S. ispartensis, A. boyeri and C. gibelio.

Seasonal variation of diet composition
A seasonal variation was shown in the diet of pikeperch (Fig 5).Prey fish species were predominant during all seasons.However, A. boyeri made up an important numerical value during winter (N= 100%), while its proportion was small in terms of weight.It was also present during all seasons.A. boyeri was an important prey only in winter in terms of relative importance (IRI = 33.43%).The importance of prey fish species changed seasonally, but A. anatoliae, K. caucasica, P. egridiri and P. parva were also absent in the fall and winter diets.C. gibelio was numerically the most important prey during the fall (N = 46.67%,W = 29.29%,O = 13.04%);however it was absent in spring and winter.According to the seasonal index of relative importance (IRI), A.

Diet variation by fi sh size
As shown in Fig. 3, the most important preys of individuals in the 20-30 cm and >50 cm classes con-sisted of some fi sh species, fi sh remains, remains of organisms and Nematoda.In addition to prey fi sh, pikeperch in the 20-30 cm length class also fed on A. boyeri, A. anatoliae, K. caucasica, C. gibelio, fi sh remains and the remains of organisms.In the stomachs of pikeperch of in the >50 cm length class, only fi sh remains of organisms and Nematoda were found In fi sh >50 cm sized pikeperch not consumed prey fi sh species.Also, cannibalism was not seen in the diets of all individuals (Fig 6).

Prey selection and feeding strategies
A. boyeri, K. caucasica, A. anatoliae and S. isparten-sis were the most abundant prey fi sh in the lake, accounting for 88.26% of fi sh; they comprise 93.98% of all prey fi sh consumed by pikeperch.P. egridiri, P. parva and C. gibelio comprise 3.31 % of all prey fi sh ingested by pikeperch although they accounted for 9.77% of the total fi sh in the lake.According to the prey selectivity index, P. parva, S. ispartensis, A. boyeri and C. splendens were positive but their selection indices were not statistically signifi cant (p>0.05). A. anatoliae and Chironomus sp. were a common prey in the lake but negatively selected by the pikeperch.P. egridiri and C. gibelio were negatively selected; their selection indices were not statistically significant (p>0.05). A. anatoliae accounted for 19.39% of the prey fish in the lake but comprises 4.82% of the diet.It was thus negatively selected and its selection index was significant (V=-0.223,X 2 = 9.997 and p<0.01).Similary, Chironomus sp., was negatively selected; its selection index was statistically significant (V=-0.297,X 2 = 17.665, p<0.01) (Fig 7).

Similarity index
Pikeperch had a similar feeding strategy in March, May, June and November because of a high Schoener overlap index (C>0.8)(Table 2).Diet composition in the group over 30 cm in length was different from that in the other length groups as the C value was smaller 0.8.C; the pikeperch in all stations had different diet compositions because of the low values of C (C<0.8).The feeding strategy in spring and summer was similar (Table 3).
The mean cannibalism proportion pikeperch populations were reported as 0.14% in Lake Ijssel (10), 96% in Lake Eğirdir (13), 20.9% in Lake Beyşehir (Balık, 1999), 40.5% in Hirfanlı Dam Lake (2), 0.1% in Lake Beyşehir (4), 0.6% in Lake Eğirdir (Balık et al., 2006).The cannibalism proportion in this research was 0% for the pikeperch population.It seems that in recent years the population of prey fishes increased along with the increasing density of sand smelt (Atherina boyeri) in Lake Eğirdir.In conclusion, this research shows that pikeperch, S. lucioperca, feed on very dense fish species, and the diet and feeding feature changes by month, season, fish length and habitat.Also, in past decade, cannibalism among pikeperch has not occurred, because the populations of the sand smelt (Atherina boyeri) are abundant in the lake.

Fig. 1 .
Fig. 1.Map of study area, with the location of sampling stations a= a d +a e , b=b d +b e , d=a d +b d , e=a e +b e .

Table 1 .
Diet of pikeperch (Sander lucioperca) in Lake Eğirdir, (N): prey number (W): prey weight, (O): frequency of occurrence, and (IRI): relative importance index 61±1.05 in winter.It was determined that 33.2 % of them were empty.The percentage of pikeperch containing prey was lowest in summer (52.94%) and above 70.0% in other seasons.The proportion of pikeperch containing prey was 76.76% in smallsized pikeperch (20-30 cm) and 13.28% in mediumsized pikeperch (30-40 cm).The maximum fullness index was observed in January and July, while the minimum fullness index appeared in August and November(Fig 2).The pikeperch captured in the 4 th station had the highest fullness index(Fig 3)

Table 2 .
Schoener overlap index of pikeperch sampled in different months

Table 3 .
Schoener overlap index of pikeperch according to size classes, stations, and seasons

Table 4 .
Comparison of food in different pikeperch (Sander lucioperca) populations