Main Growth Parameters of a Threatened Species Chondrostoma holmwoodii ( Boulenger , 1896 ) from Tahtalı Reservoir , İzmir , Turkey

A diverse genus of the Cyprinidae family, genus Chondrostoma Agassiz, 1832 has a wide distribution. More than half of the species distributes in Turkey, however there is little biological information about them. The aim of this study is to investigate the population parameters of Eastern Aegean Nase Chondrostoma holmwoodii and to evaluate the risks for the species in Tahtalı Reservoir. Fish sampling was carried out at 8 different sampling points at Tahtalı Reservoir in 2014 with multimesh gillnets. Population parameters such as age and sex composition, length frequence analysis and von Bertalanffy growth function were investigated. A total of 215 specimens of C. holmwoodii was sampled. Total length of the specimens varies between 4.3 28.2 cm and total weight 1.05 271 g. Age composition of the sampled specimens varies between 0 to V and most of the specimens were age-III. The von Bertalanffy growth parameters for C. holmwoodii was estimated as; L∞= 395.30 mm (SD=63.80), K=0.17 (SD=0.05) and L0=46.45 mm (SD=9.41). In conclusion, main problem for C. holmwoodii population in Tahtalı Reservoir is predation risk by Perca fluviatilis Linnaeus, 1758. In order to offer conservation strategies for C. holmwoodii Tahtalı Reservoir population, diet, population structure, and life history traits of P. fluviatilis need to be studied in detail.

Evaluation of life history traits of a threatened species can be an important tool to identify population trends, leading to take successive management actions for conservation (Park et al. 2003;Riberio et al. 2008).They are also valuable tool for fisheries biologists to determine the effects of exploitation on economic fish species (Jennings et al. 1999) or by policy makers to determine the time/limits of fishing activities (Redpeth et al. 2009).In addition to these, life history traits are being used in order to predict the probability of success of invasive/introduced species (Rosecchi et al. 2001;Copp et al. 2016).
Growth is a critical trait causing determination of population demography, evolution and community interactions (Urban 2007).However, some serious critics about the model has been reviewed (Roff 1980), the von Bertalanffy growth function is the most commonly used growth function in fisheries analyses (Haddon 2011).
Tahtalı Reservoir is an important drinking water supply for İzmir province and in order to keep water quality of the lake in the desired limits, alternative land use politics have been investigated (Ay 2001).In addition to these, lake is also highly protected for fisheries activities.A total of 8 species distributes in the lake (OSİB SYGM 2015).One of the residents of the lake, C. holmwoodii is listed under Vulnerable (VU) status and it has been reported that population size and also mature individuals are decreasing (Freyhof 2014) Aim of this study is to investigate the population parameters of the threatened species C. holmwoodii in Tahtalı Reservoir, located in Küçük Menderes River Basin, İzmir, Turkey, report growth parameters for the species and compare results with other studies.

Materials and Methods
Fish sampling was carried out at eight sampling stations at Tahtalı Reservoir.Sampling was conducted at three different seasons, spring, summer and fall in 2014.Multimesh gillnets with 12 different mesh size ranging between 5 mm to 55 mm were used according to modified TS EN 14757.Dead specimens following sampling were fixed in a 4 % formalin solution and transferred to alcohol after fixation.Insize digital caliper was used to measure the total length (TL; cm) and digital scale to measure weight (W; g).Sex determinations were done with visual examination of the gonads of the sampled specimens.

Age-Length
Scales between lateral line and dorsal fin were used for determination of age.More than 25 % of the sampled specimens (n:54), covering total length range of the sample were aged by three independent readers.The multinomial logistic regression model (Gerritsen et al. 2006) from 'nnet' package (Venables and Ripley 2002) developed for R 3.4.0(R Core Team 2014) was used to predict ages of the unaged specimens.

Growth Model
Due to low sample size of small specimens, von Bertalanffy growth function by von Bertalanffy (1938) was used in where  0 is replaced by  0 ; In the equation, {|} is the mean length at age t and  ∞ , K,  0 are the coefficients to be estimated from the function, they represent asymptotic length ( ∞ ), relative growth coefficient (K) and theoretical length when fish age is zero ( 0 ).
Results from the model were then bootstrapped and given in histograms with confidence intervals.All analyses were conducted with FSA package (Ogle 2014) developed for R 3.4.0(R Core Team 2014).

Results
A total of 215 specimens of C. holmwoodii was sampled.Total length of the specimens varies between 4.3 cm and 28.2 with a mean value of 19.17 cm (SD=0.42)and total weight (W) varies between 1.05 and 271 g.According to a subset of 92 specimens, female/male ratio is 1.00/0.94and 16 of them were identified as immature.Deviation from 1:1 ratio is not significant at p<0.05.

Age-Length
Age of all specimens were predicted from the aged specimens (n:54).Age composition of the sampled specimens varies between 0 to V. Total length variability of the age groups is illustrated at Figure 1 and given in detail in Table 1.Most of the specimens (33%) were age-III with a mean total length of 189 mm (SD=24.16).Mean length of second lowest age, which was found out to be age-0, was 63.5 mm (SD=19.8).Only one specimen was found to be age-I.Total length of the 29 (6 %) age-II specimens were 148 mm with a standard deviation of 6.33 mm.Mean total length of age IV and V specimens were found out to be 219 (SD=16.11)and 249.5 (SD=13.83)mm respectively.
Bootstrapped von Bertalanffy growth parameter results were given as histograms in Figure 2 with confidence intervals.Residuals of the model fit with frequencies shown in Figure 3.

Discussion
Reviewing the extent literature is beyond the scope of this study however, using different parameterizations, controlling correlations between model parameters is gaining a momentum.Fitting fish growth data with a von Bertalanffy growth function is really hard, model fit can be inaccurate because of missing data (e.g.young classes) and characteristics of the model (Ogle 2016).In order to deal with lack of young fish classes (Age I), original version of the model where t0 is replaced with L0, is used in this study rather than Beverton (1954).According to Cailliet et al. (2006) L0 has serious advantages over t0, especially for evaluation of fit.According to an experimental study with C. nasus, length after 63 days of hatch is averagely 3.47 cm (Schludermann et al. 2009).Our estimate of L0=4.65±0.94cm seems to represent not the best but an acceptable fit.
Result of this study significantly differs from the results presented by İlhan et al. (2010).Authors reported  ∞ = 29.19cm, t0 = -3.258,k = 0.177 from a population whose fork length ranged between 13 -24 cm.When we calculate L0 from these values, it equals to 12.79 cm, which is larger than prediction of 4.65 cm.Also  ∞ results from this study is lower than prediction of 39.53 cm.These differences are probably due to sampling equipment selectivity.Specimens sampled in this study ranged from 4.3 cm to 28.2 cm and this range is probably giving a better estimation.
According to a study conducted with the same specimens allometric coefficient b of the C. holmwoodii was 3.04, the b value showed a significant deviation from value 3 (p < 0.05) indicating a positive allometry (Yerli et al. 2016)  According to our observations, main problem for C. holmwoodii population in well protected Tahtalı Reservoir is predation risk by Perca fluviatilis Linnaeus, 1758.Multimesh gillnets sampled specimens from 4.3 cm to 28.2 cm; however, there is only one specimen between 10 -13 cm range which falls to age I of the species.Similar observation was reported by İlhan et al. (2010).Lack of this age group might be due to evasion of young members of the population to deeper parts of the lake to avoid predation risk.Another possible explanation for this situation might be that this is a population response to decrease interspecific competition.However detailed predator-prey relationship studies need to be conducted in order to answer these questions and offer a conservation strategy.
In order to offer conservation strategies for C. holmwoodii Tahtalı Reservoir population, diet, population structure, and life history traits of P. fluviatilis need to be studied in detail.By this way, in addition to getting this important data about P. fluviatilis, its population size will be controlled too.

Figure 1 .
Figure 1.Age-Length of the C. holmwoodii in Tahtalı Reservoir (Dot size indicate percentages)

Figure 2 .
Figure 2. Frequency for bootstrapped von Bertalanffy growth parameters with confidence limits

Table 1 .
C. holmwoodii mean total lengths at age

Table 2 .
von Bertalanffy growth paramater estimates for C. holmwoodii .İlhan et al. (2010)reported that the b value for C. holmwoodii in Tahtalı Reservoir similar to this result (3.13).On the other hand, the b value of different species belonging to the genus