Introduction
The golden grey mullet, Chelon auratus (Risso, 1810) is a major commercial resource of Turkish waters distributed from the Eastern Atlantic from Morocco to Norway, and the Mediterranean Sea and Black Sea (Nelson, Reference Nelson2016). This is a euryhaline and eurythermal species tolerant to large salinity and temperature ranges and inhabiting lagoon entrances and lower estuaries (Ben–Tuvia, Reference Ben-Tuvia1986; McDowall, Reference McDowall1988; Blaber, Reference Blaber1997). It attains 59 cm, and plays an important role in the food web, with a trophic level of 2.8 (Froese & Pauly, Reference Froese and Pauly2019).
Studies of this species and other mugilids in the Mediterranean Sea have focused mainly on their age and growth (Hotos, Reference Hotos2003; Hotos & Katselis, Reference Hotos and Katselıs2011; Hotos, Reference Hotos2019) and reproduction (Slastenenko, Reference Slastenenko1956; Brusle, Reference Brusle and Oren1981; Gözükara, Reference Gözükara2000; Hotos et al., Reference Hotos, Avramidou and Ondrias2000; Kraljevic et al., Reference Kraljević, Dulčić, Pallaoro and Matić-Skoko2011). Despite these numerous studies including the Aegean Sea and Black Sea off Turkish coasts (Bilgin et al., Reference Bilgin, Bircan, Sümer, Özdemir, Çelik, Ak and Bayraklı2006; Ilkyaz et al., Reference Ilkyaz, Firat, Saka and Kinacigil2006), nothing is known about their growth and reproduction in the Sea of Marmara.
There are several species of grey mullets (Mugilidae) living in the Sea of Marmara, C. auratus, Liza ramada, Liza saliens and Mugil cephalus (Deveciyan, Reference Deveciyan1926; Aksıray, Reference Aksiray1954; Slastenenko, Reference Slastenenko1956; Karakulak & Yıldız, Reference Karakulak and Yıldız2016), and C. auratus and M. cephalus are the most valuable with the highest catches. Their abundance is higher in spring and autumn when they enter into estuaries of rivers which discharge into the Sea of Marmara (Zengin et al., Reference Zengin, Güngör, Güngör, İnceoğlu, Düz, Benli, Kocabaş, Ceylan, Dağtekin, Demirkol and Çolakoğlu2017). These species are caught during the open commercial fishing season (from September–March) by seiners and trammel nets (Yildiz & Karakulak, Reference Yıldız and Karakulak2016). Recreational fishermen catch C. auratus in the Golden Horn using hand-lines all year round (Iwano & Öztürk, Reference Iwano and Öztürk2012).
Mullet catches gradually declined between 2000 and 2018 in the Sea of Marmara (TUIK, 2019). Unfortunately, mullet landings are generally not recorded to the species due to difficulties of identification, and do not incorporate recreational catches. This study aimed to shed some light on the reproductive biology, age and growth of C. auratus from the Golden Horn Estuary.
Materials and methods
The study was carried out in the Golden Horn Estuary, Istanbul Strait, from January to December 2017. The estuary is 8 km long and ~0.9 km wide at the southern portion of the Bosphorus Strait. Two streams, Alibeyköy and Kağıthane, provide the main freshwater inflow (Figure 1). However, the inflow from these tributaries has been severely reduced due to the construction of several dams in the watershed (Kınacı et al., Reference Kınacı, Inan, Aydın, Yüksel, Sevim, Arıkan and Topacık2004; Albayrak et al., Reference Albayrak, Balkıs, Balkıs, Zenetos, Kurun and Karhan2010).
Fig. 1. Study area (Golden Horn Estuary).
Fish samples were collected randomly each month from fishers around the estuary using set nets, cast nets and hand-lines, and were immediately covered in ice and transported to the laboratory for further investigation. A total of 431 mullets were collected, among which 197 (46%) were females, 213 (49%) were males, and in 21 (5%) the sex was not assigned with certainty. Total length ranged from 13.4–46.8 cm. Age estimation was performed by reading annual growth rings on 390 otoliths. The remaining otoliths were not suitable for reading due to the presence of vateritic structures.
Total length (TL), total weight (W), and gonad weight (GW) were recorded in each fish to the nearest 0.1 mm and 0.1 g, respectively. Sex of individuals was determined by macroscopic examination of the gonads using the maturity scales defined by Holden & Raitt (Reference Holden and Raitt1974).
The difference in mean length and weight between sexes was estimated using the Kruskal–Wallis-test.
The length–weight relationship was calculated by the equation
$$W = {\rm a}{\rm L}^{\rm b}\comma \;$$where W is the total body weight (g), L is the total length (cm), and a and b are coefficients (Le Cren, Reference Le Cren1951). Analysis of covariance (ANCOVA) was performed to test the differences for length–weight relationship between sexes (Zar, Reference Zar1999).
The gonadosomatic index (GSI) and condition factor (CF) were calculated as GSI = (GW⁄W) × 100 (Barber & Blake, Reference Barber, Blake, Shumway and Parsons2006) and CF = (W/L 3) × 100 (Ricker, Reference Ricker1975), where GW is the gonad weight, W is the total weight, and L is the total length.
Sagittal otoliths were removed, cleaned and stored dry in polyethylene bags for subsequent steps to determine age. Age was estimated by counting annual growth rings on otoliths. Before the age determination step, the otoliths were immersed first in ethanol (70%) and then in glycerine and xylol solutions for about 5 min. A Leica DC 500 camera system connected to Leica S8 APO stereo microscope and image analysis program (Leica Application Suite Version 4.3.0) was used for otolith imaging with reflected light over a black background (Uysal, Reference Uysal1992). Ages were confirmed by two separate experienced age readers.
Length at age was described by the von Bertalanffy growth model
$$L_{\rm t} = { L_{\infty} }\left( {1\!-\!e^{-{\rm k}(t-t_0)}} \right) $$where L t is the length at age t, L ∞ is the asymptotic length, k is the growth coefficient, and t 0 is the age at ‘zero’ length (von Bertalanffy, Reference von Bertalanffy1938). The growth performance index (φ′, phi-prime) was used for comparing the results with those of other studies (Pauly & Munro, Reference Pauly and Munro1984),
$${ \varphi{\prime}} = {\rm lo}{\rm g}_{ 10}k + {\rm 2lo}{\rm g}_{ 10}{ L_{ \infty }}$$The length at maturity (L 50), was estimated for both sexes using a logistic function that was fitted to the proportion of sexually mature individuals collected during the peak of spawning period by each size class using a non-linear regression following King's (Reference King1995) formula
$$P = 1/\lcub 1 + {\rm exp}\,\lsqb {-}r \lpar {L-{\rm Lm}} \rpar \rsqb \rcub \comma \;$$where P is the proportion mature in each size class, r (−b slope) is a parameter controlling the slope of the curve and Lm is the size at which 50% of fish are mature (Saila et al., Reference Saila, Recksiek and Prager1988).
The χ2 test was used to compare sex ratio (F:M) from the expected 1:1 ratio (Sümbüloğlu & Sümbüloğlu, Reference Sümbüloğlu and Sümbüloğlu2005).
All statistical analyses and plots were generated using R (R Development Core Team, 2018) with FSA (Ogle et al., Reference Ogle, Wheeler and Dinno2019), ggplot2 (Wickham, Reference Wickham2016) and sizeMat (Torrejon-Magallanes, Reference Torrejon-Magallanes2019) packages.
Results
Sex ratio and length distribution
The observed sex ratio (F:M) was 1:1.08. It was not significantly different from the expected ratio of 1:1 (P > 0.05). There were no statistical differences between male and female length (Figure 2).
Fig. 2. Total length (TL) distribution by sexes from Golden Horn Estuary.
Size, weight, age and growth
The LWR was calculated as W = 0.0127L2.89 for all individuals, R 2 = 0.9204 (W = 0.0099 × L2.97 for females, R2 = 0.9391; W = 0.0156 × L2.82 for males, R 2 = 0.9149). The ANCOVA test indicated that there were no significant differences between the slopes (b) of equations estimated for females and males (P < 0.05).
Captured fish age ranged from 1–10 years, with 2-year-old fish predominating in the total sample (43%) (Figure 3). The number of fish decreased with increasing age. The estimated von Bertalanffy growth parameters were: L ∞ = 57.52 cm, K = 0.10 year−1, t 0 = −2.24 year for pooled data (Figure 4).
Fig. 3. Age frequencies of C. auratus from otolith readings.
Fig. 4. The von Bertalanffy growth curve for all samples (N = 431).
Reproduction
The gonadosomatic index (GSI) values calculated monthly for both sexes (Figure 3) exhibited two annual peaks: in spring (March–April) and in autumn (September–November), suggesting the existence of two spawning periods per annum with autumn reproduction being relatively more important. Decline in the condition factor (CF) was observed after each spawning peak, but was more strongly expressed after the spring spawning (Figure 5).
Fig. 5. Box-plot of monthly changes in the Gonadosomatic Index (GSI) and Condition Factor (CF) calculated for both sexes.
The length at maturity (L 50) was estimated as 26.2 cm for males and 24.1 cm for females (Figure 6) that correspond to 3-year-old fish in females and 4-year-old fish in males.
Fig. 6. The length at first sexual maturity of female and male individuals (A: Male, B: Female).
Discussion
The age, growth and reproductive biology of C. auratus were studied for the first time in the Sea of Marmara. While the sex ratio (M:F) tended to be biased to females in some studies elsewhere, e.g. 1:1.87 in Aegean Sea, 1:1.22–1.42 in Caspian Sea (Ilkyaz et al., Reference Ilkyaz, Firat, Saka and Kinacigil2006; Fazli et al., Reference Fazli, Daryanabard, Abdolmaleki and Bandani2008; Ghaninejead et al., Reference Ghaninejad, Abdolmalaki and Kuliyev2010), it was closer to theoretical (1:1) in the Adriatic Sea (Bartulovic et al., Reference Bartulović, Dulčić, Matić-Skoko and Glamuzina2011) and in this study. These variations in the sex ratio are probably related to differences in the age composition of the stock linked to higher mortalities in older males (Cooper et al., Reference Cooper, Zapata, Barrutia and Ramirez1983), and also food availability and environmental conditions.
The LWR in fishes is likely influenced by several factors such as: season, habitat, gonad maturity, sex, diet, stomach fullness, health, preservation techniques and length variability of specimens (Reis & Ates, Reference Reis and Ateş2019). LWR in fish in the Sea of Marmara was similar to that estimated for other areas of the Mediterranean Sea (Table 1) with ‘b’ parameter being somewhere in the middle of the observed range of between 2.49 (from the Gulf of Lion) and 3.26 (from Korinthiakos Gulf).
Table 1. The length–weight relationship parameters of C. auratus obtained from other studies (F, Female; M, Male; U, Unsexed; J, Juvenile)
The growth parameters for C. auratus have been reported from the Mediterranean Sea by several authors. The asymptotic length (L ∞), calculated from this study was similar to that in the relatively brackish Black Sea (Alexandrova, Reference Alexandrova1964) and Caspian Sea (Fazli et al., Reference Fazli, Daryanabard, Abdolmaleki and Bandani2008) rather than to the rest of the Mediterranean and the Atlantic Ocean with relatively high salinity. The growth coefficient (K = 0.1) in our study was within the known range (0.079–0.311 – Table 2), indicating a slow-growing and long-lived species, with relatively low yield-per-unit stock, because of a lower production/biomass ratio. The oldest C. auratus aged in this study of 10 years was just under the maximum reported age of 11 years in the Caspian Sea. The growth performance index (phi-prime Φ′) of the mullet from the Sea of Marmara was one of the lowest among studied populations (Table 2).
Table 2. The von Bertalanffy growth parameters (L ∞, asymptotic mean length; k, growth rate; t 0, hypothetical age at zero length) and growth performance index values (Φ′) obtained from different areas for C. auratus.
The CF values depend on age, sex, season and habitat (Hotos et al., Reference Hotos, Avramidou and Ondrias2000) so are difficult to compare. However, the CF of mullet from adjacent estuaries of the Black Sea (Bilgin et al., Reference Bilgin, Bircan, Sümer, Özdemir, Çelik, Ak and Bayraklı2006) mostly ranged between 0.7 and 1.1 (lowest ~0.3), while the majority of CF values ranged between 1.0 and 1.5 (lowest 0.68) in our study. The higher CF found in the Sea of Marmara may be related to the high nutrient capacity of the Golden Horn estuary (Albayrak et al., Reference Albayrak, Balkıs, Balkıs, Zenetos, Kurun and Karhan2010).
This study demonstrated that the main spawning peak in the Golden Horn Estuary occurs during the same time period reported from other studies conducted in the Black and Aegean Seas and off Western Greece (Hotos et al., Reference Hotos, Avramidou and Ondrias2000; Bilgin et al., Reference Bilgin, Bircan, Sümer, Özdemir, Çelik, Ak and Bayraklı2006; Ilkyaz et al., Reference Ilkyaz, Firat, Saka and Kinacigil2006); from August–September to November–December. It was supposed that the reproductive period of this species is highly temperature dependent and occurs between 7–10°C in July and August (Çelikkale, Reference Çelikkale1991). However, this study demonstrated that a much higher temperature of ~19.8°C was associated with the main peak spawning period in September and relatively high temperatures of 11.7–13.7°C with the minor spawning peak (March and April) (oceanographic data from Dorak, Reference Dorak2010), implying that C. auratus spawns in warmer temperatures in the Golden Horn Estuary. Existence of the second spawning peak probably is caused by high water productivity combined with high temperatures accelerating metabolism.
Previous studies indicate that water temperature affects mullet length-at-maturity (Brusle, Reference Brusle and Oren1981; Cambrony, Reference Cambrony1983; Katselis et al., Reference Katselis, Minos, Marmagas, Hotos and Ondrias1994; Bilgin et al., Reference Bilgin, Bircan, Sümer, Özdemir, Çelik, Ak and Bayraklı2006; Ilkyaz et al., Reference Ilkyaz, Firat, Saka and Kinacigil2006). The length-at-maturity of C. auratus estimated in other areas varied between 24 and 34 cm (Campillo, Reference Campillo1992; Fazli, Reference Fazli1998; Hotos et al., Reference Hotos, Avramidou and Ondrias2000; Bilgin et al., Reference Bilgin, Bircan, Sümer, Özdemir, Çelik, Ak and Bayraklı2006; Ghaninejad et al., Reference Ghaninejad, Abdolmalaki and Kuliyev2010) with our data lying closer to the lower boundary. It is consistent with the relatively high temperature of the Sea of Marmara. In practice it means that the current minimum legal landing size of C. auratus in Turkey of 30 cm imposed by fisheries regulations (BSGM, 2016) should be reduced to 26 cm, at least for the Sea of Marmara.
In the Golden Horn Estuary, C. auratus is an economically important resource for small-scale and recreational fishers. This study filled in some missing gaps on its basic population parameters and spawning aspects of this stock necessary for successful management of its fisheries. However, additional research on its biology, ecology and fisheries are advised to better understand the sustainability of its exploitation.
Acknowledgements
This work was supported by Istanbul University BAP, Project number: FYL-2016-21263. We thank Aylin Ulman from Mersea Marine Conservation Consulting for helping us with English editing.
