Age , growth , gonadosomatic index and diet composition of Crimean barbel , Barbus tauricus ( Actinopterygii : Cypriniformes : Cyprinidae ) , in a small Stream in NE Turkey

Background. Crimean barbel, Barbus tauricus Kessler, 1877, is a riverine cyprinid fish commonly found in welloxygenated streams with gravel bottom in the Black and Azov Sea basins. Its population has plummeted in the Salgir, Chornaya, and Alma rivers (Crimea) and hence this fish has been listed as Vulnerable on the IUCN Red List. The knowledge about its age, growth, length–weight relation, spawning period, and diet composition are either scarce or not available. This paper aimed to fill the existing gaps in the knowledge by describing selected biological characteristics of B. tauricus in the Çiftekavak Stream, in the outskirts of the city of Rize, NE Turkey. Materials and methods. Crimean barbel were collected by electrofishing (60 Hz pulsed DC) from April to November 2014. The total length (L, cm) and weight (W, g) of each specimen were recorded, and sagittal otoliths, gonads, and gut contents were then recovered. The length–weight relation (LWR) was calculated by a simple power function W=aLb. The age rings on sagittal otoliths were counted to determine fish age that was later used to analyse their growth by various growth models. The wet weight of gonads was used to calculate the gonadosomatic index (GSI). The gut contents were identified to the lowest possible taxonomic level and the contribution of a prey in the total diet composition was analysed by the occurrence frequency of prey groups (%O) and by numerical percentage frequency of prey groups (%N). Results. The age ranged from 0 to 4 years and more than 50% of the fish represented the 0-year group followed by 1-year group (21.9%) and 2-year group (13.5%). The von Bertalanffy growth model adequately described the correlation between the fish length and the age and indicated that females grew faster than males. The LWR identified negative allometric growth patterns in males and females. The higher values of GSI from males and females were recorded from April through July, while the lowest value of GSI was observed from September through November indicating the completion of the spawning season. A total of 14 prey items (including sand grains) were identified from the guts of Crimean barbel. The main prey items were Culex sp. (larva + pupa + adult), Chironomidae, followed by Ephemeridae and Zygoptera. They constituted up to >78%O (>95%N) of the diet. Conclusions. The results of this study will assess the conservative regulations and policies that will eventually provide a sustainable management of Crimean barbel stocks.


INTRODUCTION
Crimean barbel, Barbus tauricus Kessler, 1877 are widely distributed in the Black and Azov Sea basins (Kottelat andFreyhof 2007, Çiçek et al. 2015).Those fish are commonly found in well-oxygenated streams with a gravel bottom and high current velocity (Verep et al. 2006).Their occurrence within the known range of distribution has been well documented by various authors (Dobrovolov 1996, Kotlík and Berrebi 2001, Sarı et al. 2006).However, very little information exists on basic biological characteristics (e.g., spawning period) and feeding ecology of Crimean barbel (Kottelat and Freyhof 2007).Due to pollution, their population have plummeted in the Salgir, Chornaya, and Alma rivers (Crimea) and hence listed as a vulnerable species by the International Union for Conservation of Nature (Freyhof and Kottelat 2008).
The majority of the species of the subfamily Barbinae are bottom-feeders mainly consuming insects, including Chironomidae as the dominant prey (Collares-Pereira et al. 1996, Piria et al. 2005, Sapounidis et al. 2015).The spawning seasons of those fishes usually start in April and last until July (Herrera and Fernández-Delgado 1992, Kottelat and Freyhof 2007, Sapounidis et al. 2015).Females have a higher growth rate than males (Vitali and Braghieri 1984, De Silva et al. 1985, Herrera et al. 1988).
In this study, the basic biological characteristics of Barbus tauricus, including sex ratios, age, growth, spawning period, and length-weight relation (LWR) were examined for the first time in the Çiftekavak Stream, in the outskirts of the city of Rize, NE Turkey.This study intends to be a complement to the results obtained for LWR by the previous authors (Tarkan et al. 2006, Şahin et al. 2007, Gaygusuz et al. 2013a, 2013b).Also, the gut contents of B. tauricus were analysed to determine their diet composition during spring, summer, and autumn.The impact of fish sizes on their diet composition was also noted.

MATERIAL AND METHODS
Study area.The Çiftekavak Stream is a coastal stream located about 4 km west of the Rize city centre.It has a total length of 6 km with an approximate depth of 0.4 m, and a width of 4 m.It flows from Tuğlalı to Çiftekavak and empties into the Black Sea.The dominant bottom substrate of the stream is gravel.During the presently reported study the water flow velocity was 1 m • s -1 , the water temperature ranged from 15 to 22°C, pH was 7.5-8.0,and the dissolved oxygen amounted to 8.0-10.0mg•L -1 .No signs of industrial or anthropogenic pollutants were found.Also, all types of fishing are banned in the area.
In the laboratory, each fish was weighed (to the nearest 0.01 g) and its total length was measured (to the nearest 0.1 cm).Based on the total length, the fish were categorized into four length classes: 6.6-10.5 cm, 10.6-14.5 cm, 14.6-18.5 cm, and >18.6 cm.Finally, guts, gonads, and sagittal otoliths were recovered.The gut contents were promptly collected on Petri dishes and were identified to the lowest possible taxonomic level while the wet weight of the gonads was determined to the nearest 0.01 g.Furthermore, the sex of the collected specimens was determined by dissection of gonads.
Field samplings performed in this study were authorized by Ministry of Food, Agriculture and Livestock, General Directorate of Agricultural Research and Policies, Turkey.
Age determination.The age rings on whole sagittal otoliths were counted using a Nikon SMZ1000 stereomicroscope coupled to a Nikon DSFI1 digital camera at a magnification between ×0.8 and ×8.0.In order to make the growth rings clearly visible, each otolith was gently sanded using sandpaper.
Growth estimation.Several growth models were applied to the size-at-age data of Barbus tauricus to analysis their growth: (Schmalhausen 1926as cited in Ricker 1979) The respective formulas are given below: where L is the fish total length [cm] at the time t (age), L ∞ is the upper asymptotic total length [cm], K is the growth rate coefficient [year -1 ], t 0 and β is the hypothetical age.These parameters were estimated using PRIMER 6.0 (Plymouth Routines In Multivariate Ecological Research) software.
The best fit model to the size-at-age data was determined using Akaike Information Criterion (AIC) results (Akaike 1974) and the one provided the lowest AIC was chosen as an adequate growth model.Furthermore, the growth performance index (φ′) was calculated as indices of growth performance by following the formula using the values of K and L ∞ (Pauly and Munro 1984): ( ) ( ) Length-weight relation.The LWR of Barbus tauricus was determined by following formula: where W is the body weight [g], L is the total length [cm], a is the intercept, and b is the slope.The statistical deviation of estimated b from the isometric value (3.0) was tested by t-test (Pauly 1984).
Qualitative dietary analysis.The contribution of each prey item to the total gut contents was analysed by the occurrence frequency of prey groups (%O) and by numerical percentage frequency of prey groups (see reviews by Hyslop 1980, Cortés 1997).The abovementioned indices were calculated as: where n is the number of guts of a particular prey type, N s is the total number of guts containing prey, n i is the total number of prey in a food group, and N p is the total number of all prey groups.Dietary variation.The similarities in diet composition between different length classes during different seasons were estimated by dendrogram using the PRIMER 6.0 software package (Clarke and Warwick 2001).

RESULTS
Length frequency distribution and sex ratio.A total of 360 Crimean barbel were collected throughout the study whose sizes (total length) ranged between 5.0 and 24.7 cm.Females and males ranged in total length between 7.8 and 24.7 cm (n = 90) and 6.8-18.2cm (n = 149), respectively.Mean total lengths of females were significantly greater than males (t-test, P < 0.001).The total length of juvenile ranged between 5 and 10.9 cm (n = 121).The length frequency distributions of males and females also differed significantly (Kolmogorov-Smirnov two-sample test: d = 0.4635, P < 0.001).The dominant size classes in the length frequency distribution for males were 12 and 14 while the dominant size class for females was 13 cm.Among juveniles, the dominant size classes were 8, 9, and 10 cm (Fig. 1).Sex ratios for the population of Crimean barbel in the Çiftekavak Stream indicate a dominance of males with (female ÷ male) 0.60 ÷ 1 but it did not deviate significantly from 1 ÷ 1 (χ 2 = 0.263; P = 0.608).
Age and growth.Otoliths from 274 individuals were extracted and read successfully (Table 1).The majority of the individuals represented the 0-year age group (53.7%) and half of them were of juveniles.The estimation of growth parameters included all age groups.Based on AIC values, the correlation between the total length and age of B. tauricus was adequately described by the von Bertalanffy model.The faster growth rates were also indicated by the growth performance index that produced a higher value of φ′ for female (Table 2).
Length-weight relation.The LWR was determined for male and female separately and for all individuals that representing both sexes and the juveniles (Table 3).The estimated values of allometric coefficient indicated isometric allometry of growth of females and males.A  slightly positive allometry pattern, however, was obtained from combined data.Gonadosomatic index.The reproductive period was determined using the GSI values (68 females and 140 males).The highest values of GSI were found during April followed by the same pattern (in subsequent months).The GSI values of females were also observed to be higher from April through July with the highest value during June (Fig. 2).The lowest GSI values were found during September, October, and November.During these three months, the GSI values were found statistically similar for both males (ANOVA, F 2, 23 = 0.382; P = 0.687) and females (F 2, 5 = 0.189; P = 0.833).These findings indicated the completion of spawning season before September.
Overall diet composition.A total of 213 alimentary tracts from fish ranging from 6.8 to 24.7 cm of total length (12.46 ± 0.17 cm) were examined for diet composition.It turned out that 7% of guts were empty, 62.9% were moderately full, and >16.9% were half full, 9.95% were quite full, and 3.3% were very full (Fig. 3).
The diets of Crimean barbel included a total 14 prey types (including sand grains), of which seven belonged to order Diptera.The contribution of each prey type during different seasons as well as in overall diet compositions are summarized in Table 4. Diptera made up 68.31%O (80.31%N) of the overall diet.The four most predominant prey items were Chironomidae, Culex sp.(larva + pupa + adult), followed by Ephemeridae and Zygoptera.
Diet of Crimean barbel of different sizes in relation to season.The main prey group of different fish sizes remained Diptera constituting to >66%O (>65%N) of the total diet composition of all length classes (excluding 6.6-10.5 cm length class) during different seasons (Table 4).The diet composition of 6.6-10.5 cm length class contained 58%O (55%N) of Diptera.Similar to overall diet, the diets of all length classes (excluding 6.6-10.5 cm length class in spring) were largely made up by Chironomidae, Culex sp.(larva + pupa + adult), Ephemeridae, and Zygoptera (>83%O, >87%N).Moreover, the most dominant prey item was Culex sp.(larva + pupa + adult) in spring and summer seasons while during autumn it was replaced by Chironomidae.The higher amount of Nematoda gen.sp.(6%N, 17%O) and plant detritus (17%O) were recovered from the 6.6-10.5 cm length class in autumn.
Similarities in the diet composition.The dendrogram generated based on the results obtained showed diets of all different sizes of Barbus tauricus revealed >85% similarity in summer.Spring diet of 6.6-10.5, 10.6-14.5,     and >18.6 cm length classes showed >80% similarity.On the other hand, the diets of different sizes B. tauricus during autumn had relatively lower similarities and the diet of 6.6-10.5 and 14.5-18.5 cm length classes showed >67% similarity.The spring and summer groups separated from the autumn with 48.5% dissimilarity (Fig. 4).
Length-weight relation.The LWR of Barbus tauricus (all 360 specimens) evinced that they grow in slightly positive allometry pattern which was consistent with previous studies from Ömerli Dam Lake (Tarkan et al. 2006), Yeşildere Stream (Şahin et al. 2007), and Emet Stream, Porsuk Stream, and the Sakarya River at Sakaryabaşı (Gaygusuz et al. 2013a(Gaygusuz et al. , 2013b)).In this study, the LWR of males and females (calculated separately) showed that they grow following isometric allometry growth pattern (Table 3).In previous studies, the LWR of males and females were not reported separately, therefore estimating the length or weight of B. tauricus using the LWR parameters obtained from combined data (males, females, and juveniles) would result in erroneous estimation for male and female.Spawning period.The spawning season of Barbus tauricus in the Çiftekavak Stream started in April and lasted until July, which was in line with the results of Herrera and Fernández-Delgado (1992) on Luciobarbus sclateri, results of Vasiliou and Economidis (2005) on Barbus peloponnesius Valenciennes, 1842 and Barbus cyclolepis, and results of Sapounidis et al. (2015) on Barbus strumicae.Furthermore, according to Kottelat and Freyhof (2007) majority of Barbus spp.spawn in May-July.
Diet composition.Barbus tauricus exclusively feed on insects and the presence of plant detritus and sand grains (also Nematoda gen.sp.) in their gut contents evinced their feeding habits to be an omni-insectivorous bottom-feeder.The contribution of Diptera to the diet of B. tauricus as the first most frequent prey group along with the presence of plant detritus were in accordance with the results reported from previous studies on the feeding ecology of Luciobarbus bocagei (Steindachner, 1864) (see Magalhães 1993, Collares-Pereira et al. 1996), Barbus barbus (Linnaeus, 1758) (see Piria et al. 2005), Barbus cyclolepis (see Rozdina et al. 2008) and Barbus strumicae (see Sapounidis et al. 2015).Furthermore, Collares-Pereira et al. (1996) and Sapounidis et al. (2015) also reported the presence of sand grains in the gut contents of B. bocagei and B. strumicae.The results of the presently reported study stand in contrast to earlier work regarding the first most frequent prey item in the diets.The diets of aforementioned species were dominant by Chironomidae however the dominant prey item in the gut contents (overall diet composition) of B. tauricus was Culex sp.(larva + pupa + adult) or Chironomidae (Table 4).Based on seasonal data, during spring and summer, the diets were dominant by Culex sp. while in autumn the Chironomidae became the first most frequent prey items in the diet of B. tauricus.According to by Kottelat and Freyhof (2007) Crimean barbel move to deeper places with less current and stop feeding during winter.The presence of nematodes in the gut contents of Barbus sp. were also reported by Admassu and Dadebo (1997) from Lake Awassa, Ethiopia.

Fig
Fig. 1.Total length frequency distribution of females, males, and juvenile Crimean barbel, Barbus tauricus, sampled during spring, summer, and autumn from the Çiftekavak Stream, in the outskirts of the city of Rize, NE Turkey Table 1 Principal biometric parameters of Crimean barbel, Barbus tauricus, in the Çiftekavak Stream, in the outskirts of the city of Rize, NE Turkey

L
∞ = the upper asymptote, K = the growth rate, t 0 = the time when L = 0, I = the age at the inflection point, β = the size at time zero, φ′ = growth performance index, AIC = Akaike Information Criterion, VBGF = von Bertalanffy growth function.
as: the occurrence frequency of prey groups (%O) and by numerical percentage frequency of prey groups (%N).

Fig. 2 .Fig. 3 .Fig. 4 .
Fig. 2. Monthly gonadosomatic index values (Mean ± SD) of males and females of Crimean barbel, Barbus tauricus, caught from Çiftekavak Stream, in the outskirts of the city of Rize, NE Turkey Principal biometric parameters of Crimean barbel, Barbus tauricus, in the Çiftekavak Stream, in the outskirts of the city of Rize, NE Turkey Total length frequency distribution of females, males, and juvenile Crimean barbel, Barbus tauricus, sampled during spring, summer, and autumn from the Çiftekavak Stream, in the outskirts of the city of Rize, NE TurkeyTable1SE = standard error of the mean; ♂ = male, ♀ = female, J = juvenile.

Table 2
The growth parameters for Crimean barbel, Barbus tauricus, derived from different growth models are presented with growth performance index

Table 3
Total length and parameters of the length-weight relation (LWR) for Crimean barbel, Barbus tauricus, from in the Çiftekavak Stream and from other locations in Turkey TL = total length; a = constant (intercept), b = constant (slope of regression line), SE= standard error; ♂ = male, ♀ = female, B = both sexes including juveniles.

Table 4
Diet composition of four length classes of Crimean barbel, Barbus tauricus, in the Çiftekavak Stream, in the outskirts of the city of Rize, NE Turkey