INTRODUCTION
Albacore, Thunnus alalunga (Bonnaterre, 1788), is a highly migratory species widely distributed in the major oceans of the world, including the Mediterranean Sea, between latitudes 50°N and 40°S (Collette & Nauen, Reference Collette and Nauen1983). For stock assessment purposes, six independent albacore stocks are assumed worldwide, i.e. North and South Pacific, Indian, North and South Atlantic, and Mediterranean stocks (Arrizabalaga et al., Reference Arrizabalaga, Costas, Juste, González-Garcés, Nieto and López-Rodas2004). The Mediterranean stock is considered to be independent from the North Atlantic stock, as based on independent spawning areas, tagging experiments, morphological differences, different growth rates and age of first maturity (ICCAT, 2010). Moreover, studies on feeding habits (Bello, Reference Bello1999; Consoli et al., Reference Consoli, Romeo, Battaglia, Castriota, Espósito and Andaloro2008; Salman & Karakulak, Reference Salman and Karakulak2009; Goñi et al., Reference Goñi, Logan, Arrizabalaga, Jarry and Lutcavage2011; Cardona et al., Reference Cardona, Álvarez de Quevedo, Borrell and Aguilar2012) and recent genetic studies (Davies et al., Reference Davies, Gosling, Was, Brophy and Tysklind2011; Montes et al., Reference Montes, Iriondo, Manzano, Arrizabalaga, Jiménez, Pardo, Goñi, Davies and Estonba2012; Albaina et al., Reference Albaina, Iriondo, Velado, Laconcha, Zarraonaindia, Arrizabalaga, Pardo, Lutcavage, Grant and Estonba2013) support that the Mediterranean T. alalunga stock is an independent population that completes its entire life cycle in the Mediterranean Sea. The International Commission for the Conservation of Atlantic Tunas (ICCAT) stated that the Mediterranean T. alalunga is a data-poor stock, and recommended conducting research on its basic life history parameters (Anon., 2012).
Although the life cycle and migration routes of T. alalunga in the Mediterranean Sea have been little studied, three spawning grounds have been identified based on larval studies: (1) eastern Mediterranean (Aegean Sea) (Piccinetti et al., Reference Piccinetti, Piccinetti-Manfrin and Soro1996); (2) central Mediterranean (Ionian, Tyrrhenian and Adriatic Seas) (Potoschi et al., Reference Potoschi, Cavallaro, Sturiale and Lo Duca1994; Piccinetti et al., Reference Piccinetti, Piccinetti-Manfrin and Soro1996; De Ruggieri et al., Reference De Ruggieri, Vlora, Pastorelli and Marano1997); and (3) western Mediterranean (Balearic Sea) (Dicenta et al., Reference Dicenta, Piccinetti and Piccinetti-Manfrin1975; Dicenta, Reference Dicenta1977; García et al., Reference García, Alemany, De la Serna, Oray, Karakulak, Rollandi, Arigò and Mazzola2005; Alemany et al., Reference Alemany, Quintanilla, Velez-Belchí, García, Cortés, Rodríguez, Fernández de Puelles, González-Pola and López-Jurado2010). Thunnus alalunga is a multiple spawner with asynchronous oocyte development and indeterminate spawning strategy (Otsu & Uchida, Reference Otsu and Uchida1959). Like all other tuna species, spawning occurs when the sea surface temperature exceeds 24°C (Schaefer, Reference Schaefer, Block and Stevens2001). Current available information on the reproductive biology of Mediterranean T. alalunga is scarce (Arena et al., Reference Arena, Potoschi and Cefali1980; Megalofonou, Reference Megalofonou1990; Akayli et al., Reference Akayli, Karakulak, Oray and Yardimci2013), whereas several studies have been published on the stock-related reproductive characteristics of North and South Pacific T. alalunga and on the histological analyses of their gonads (Partlo, Reference Partlo1955; Ueyanagi, Reference Ueyanagi1957; Otsu & Uchida, Reference Otsu and Uchida1959; Otsu & Hansen, Reference Otsu and Hansen1962; Ratty et al., Reference Ratty, Laurs and Kelly1990; Ramon & Bailey, Reference Ramon and Bailey1996; Chen et al., Reference Chen, Crone and Hsu2010; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013, Reference Farley, Hoyle, Eveson, Williams, Davies and Nicol2014).
An accurate evaluation of the reproductive phases of the gonads should be based on histological examination (Hunter & Macewicz, Reference Hunter, Macewicz and Lasker1985), which is especially important in relation to females sampled after the spawning season when the resting ovaries may appear to be immature if gonad indices or macroscopic techniques are used (Schaefer, Reference Schaefer, Block and Stevens2001). Thus, this study used both histological and biometric data to examine various aspects of the reproductive biology of T. alalunga in the western Mediterranean, which is one of the most important spawning grounds of tuna species in the Mediterranean Sea. The study aims were to examine the size frequency, sex ratio, reproductive phase, spawning season and minimum length at maturity of T. alalunga caught in the western Mediterranean Sea.
MATERIALS AND METHODS
The data used in this study were collected from fishing tournaments of the recreational fishery (RF) and from the albacore longline fishery (LLALB) in the western Mediterranean Sea. Recreational and commercial fisheries both target T. alalunga in this area (Figure 1).
Fig. 1. Study area. The western Mediterranean Sea showing the fishing grounds of albacore longline fishery (dashed line) and of recreational fishery (circles around the base ports) during the fishing tournaments. Locations (according to the month) are shown where Thunnus alalunga ovaries were collected on albacore longline vessels. Base ports: A, S'Estanyol; B, Sóller; C, Cala D'Or; D, Port Balís; E, Torredembarra; F, Dénia; G, Jávea; H, Torrevieja.
Fishing description and sampling collection
Trolling is the fishing method used by recreational fishery and consists of using fishing lines (30 or 50 lb) with hooks and artificial lures. Fishing tournaments take place during the daytime (approximate effective fishing time: between 08:00 and 17:00 hours) and the fishing grounds extend 60 nautical miles around the base ports (Figure 1). Thunnus alalunga were caught during 21 fishing tournaments (38 fishing days) between 2005 and 2012 (Table 1). In this period, the tournaments were mainly held between mid-June and mid-July, but were only held in September in 2005 and 2006. A total of 287 positive fishing operations (sets with albacore catch) were recorded between 2006 and 2012 by scientific observers on fishing boats. The main characteristics of the gear are as follows: shallow mainline, i.e. hook depth ranges from 20 to 50 m; between 2000 and 6000 hooks are usually set in each fishing operation; and the most commonly used bait is frozen sardine Sardina pilchardus and round sardinella Sardinella aurita. The fishing operation lasts around 24 h, and thus the effective fishing time includes the day and night hours. Further details on the LLALB can be found in García-Barcelona et al. (Reference García-Barcelona, Ortiz de Urbina, De la Serna, Alot and Macías2010). In 2008, Spanish legislation on the LLALB (Order APA/254/2008, 7 February of Ministry of Agriculture, Fisheries and Food) prohibited the use of albacore longline gear between 15 October and 31 May. However, in 2010, a new order (Order ARM/2023/2010, 26 July of Ministry of Environment, and Rural and Marine) allowed its use from 1 May to 30 September. Fortunately, historical data are available for the period March–December, before the current legislation was introduced.
Table 1. Number of individuals caught (N), individuals measured (number and range of fork length), males and females, ovaries sampled and ovaries histologically examined of Thunnus alalunga caught at fishing tournaments in the western Mediterranean Sea by month (pooled across years).
Sampling from fishing tournaments was conducted by qualified staff from the Spanish Institute of Oceanography (IEO), and by scientific observers on longline fishing vessels. According to the ICCAT recommendations, onboard observers were assigned to vessels based on a standard ICCAT time-area stratum (5° × 5° × month). Positive fishing operations alone were used in the current study to identify the fishing areas where the commercial albacore longline fishery usually operates (Figure 1).
In total, 3952 T. alalunga were measured of the 4016 fish caught during the recreational fishery (RF) tournaments and 12,152 were measured from the LLALB catch (Tables 1 and 2). Fork length (L F) was measured either to the nearest 0.1 cm (fish caught by the RF) or to the nearest cm (fish measured on longline vessels). Total weight (W T) was measured to the nearest 0.1 kg. During the fishing tournaments in 2005 and 2006, observers on sports vessels monitored the catch and recorded fish that were not retained. The observers reported that no T. alalunga were discarded during any of the tournaments. In the subsequent years up to 2012, surveys of the tournament participants were conducted at landing sites and also showed that no small T. alalunga were caught and discarded.
Table 2. Number of individuals measured (N) and ovaries histologically examined of Thunnus alalunga caught by LLALB in the western Mediterranean Sea by month (pooled across years).
All the fish were sampled to identify their sex when the total catches at fishing tournaments were small, whereas a subset was randomly selected when a large number of fish were caught. Macroscopic examination of the gonads was performed to determine the sex of 1486 fish caught by the RF (Table 1) and the sex ratio was calculated as the ratio of females to males. A total of 755 ovaries were removed and the weights of the gonads (N = 753) (W G) were recorded to the nearest g. A 2−3 cm cross-section from the central part of the right or left lobe was fixed in Bouin's fluid for 4 h, and preserved in 70% ethanol. A total of 55 females, 54 males and five undetermined fish were dissected on longline vessels and their gonads were fixed and preserved as previously described. Histological examination showed that the five individuals of undetermined sex were four immature females and one male. Given that an examination of testes did not form part of the study aim, the results concerning the immature male are not discussed further. Figure 1 shows the localities by month where females were caught by the LLALB.
Gonadosomatic index (GSI) was calculated for males and females sampled in both fisheries to determine the spawning period. In order to compare our results with those reported in other albacore studies, two alternative formulations were used:
$$\hbox{GS}{\hbox{I}^{\rm a}}={10^4} \times {W_G} / {L_F}^3\comma \;$$where GSIa is the GSI calculated from the ratio of gonad weight to fork length times 104; W G in grams and L F in cm.
$$\hbox{GS}{\hbox{I}^{\rm b}}=100 \times {W_G} / {W_T}\comma \;$$where GSIb is the GSI calculated from the ratio of gonad weight to total weight times 102; W G and W T both in grams.
Laboratory processing and histological classification
A total of 587 ovaries were histologically examined (Tables 1 and 2). A representative portion of the preserved ovarian tissue (from the tunica albuginea to the ovarian lumen) was dehydrated in ascending concentrations of ethanol, cleared with n-butanol, and embedded in paraffin. Sections were cut at 10 µm and stained with Mallory's trichrome stain. The most advanced group of oocytes (MAGO) was determined for each ovary; primary growth, lipid-stage, early vitellogenic, medium vitellogenic, advanced vitellogenic, migratory nucleus and hydrated oocyte stages (Figure 2). The presence of postovulatory follicles (POFs) and atretic follicles was also recorded to determine the reproductive phases of the ovaries (Schaefer, Reference Schaefer1998). Subsequently, histological criteria based on the work of Schaefer (Reference Schaefer1998) and Farley et al. (Reference Farley, Williams, Hoyle, Davies and Nicol2013) were applied to determine the maturity status of each female and its ovarian phase. Females were classified as immature if their ovaries only contained primary growth oocytes or early vitellogenic oocytes, but no signs of atresia (Figure 2A); whereas they were classified as mature if the ovaries contained vitellogenic, migratory nucleus or hydrated oocyte stages, POFs, atresia of vitellogenic oocytes (alpha or beta atresia) and/or late stages of atresia. Ovaries of mature females were classified into the following five ovarian phases: Developing phase, when they contained early vitellogenic oocytes as MAGO, no POFs and no atresia; spawning capable phase if ovaries showed medium or advanced vitellogenic oocytes as MAGO, some atresia of vitellogenic oocytes could be present, but no POFs (Figure 2B); spawning phase, when they showed evidence of past spawning (POFs present) or imminent spawning (migratory nucleus or hydrated oocytes present), and atresia, when present at all, only in limited amounts (Schaefer, Reference Schaefer1998; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013) (Figure 2C–F); regressing phase, when they had lipid-stage or early vitellogenic oocytes, abundant alpha and/or beta atresia, and no POFs (Figure 2G); and regenerating phase, if ovaries contained primary growth or lipid-stage oocytes, late stages of atresia, and no POFs (Figure 2H). As not only oocytes in the latter two advanced developmental stages but also the presence of POFs reflects spawning tunas (Schaefer, Reference Schaefer1998; Chen et al., Reference Chen, Crone and Hsu2010; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013), spawning ovaries were further classified into four subphases according to the MAGO present in the ovary: medium vitellogenic subphase, advanced vitellogenic subphase, migratory nucleus subphase and hydrated subphase.
Fig. 2. Reproductive phases of the ovaries of Thunnus alalunga caught in the western Mediterranean Sea. Immature (A), spawning capable (B), spawning (C–F) [medium vitellogenic subphase (C), advanced vitellogenic subphase (D), migratory nucleus subphase (E), hydrated subphase (F)], regressing (G), and regenerating (H) showing the different developmental stages of oocytes and follicles: PG, primary growth oocyte; LS, lipid-stage oocyte; E-Vit, early vitellogenic oocyte; M-Vit, medium vitellogenic oocyte; A-Vit, advance vitellogenic oocyte; MG, migratory nucleus oocyte; HY, hydrated oocyte; POFs, postovulatory follicles; At, atretic follicles. Histological sections were stained with Mallory's trichrome. Scale bar = 500 µm.
Spawning frequency
The estimation of the spawning frequency of female T. alalunga from the western Mediterranean Sea was calculated following the postovulatory follicle method of Hunter & Macewicz (Reference Hunter, Macewicz and Lasker1985), which is commonly used in tuna (Hunter et al., Reference Hunter, Macewicz and Sibert1986; Farley & Davis, Reference Farley and Davis1998; Medina et al., Reference Medina, Abascal, Megina and García2002; Chen et al., Reference Chen, Crone and Hsu2006, Reference Chen, Crone and Hsu2010). This method is based on the assumption that POFs are resorbed within 24 h after spawning, which has been demonstrated in South Pacific T. alalunga (Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013). Spawning frequency was estimated as the inverse of the spawning fraction, which is defined as the number of females whose ovaries contained POFs divided by the total number of mature females. The spawning fraction was also estimated using mature active females (females whose ovaries are either in the spawning-capable or spawning phase), i.e. mature specimens whose ovaries were reproductively inactive (either in the regressing or regenerating phase) were not included.
Statistical analyses
All statistical analyses were conducted using R statistical software (R Core Team, 2013).
A two-sample Kolmogorov–Smirnov test was performed to test for differences in length distributions from LLALB data between months (pooled across years) with sufficient sample sizes (May–September). The same approach was applied to test for differences in length distributions (pooled across months and years) between the RF and the LLALB. A binomial generalized linear model (GLM) was used to examine the effect of L F on the proportion of females. A significance level of α = 0.01 was assumed in all statistical tests.
RESULTS
Size composition
During the study period, observers on the LLALB boats measured a total of 12,152 T. alalunga. Their sizes ranged between 33 and 109 cm L F (mean ± SD = 72.4 ± 8.3 cm) (Figure 3A). No statistically significant differences in size distribution between months were found (P > 0.01). The sizes of the 3952 specimens from the RF ranged between 57 and 103 cm L F (mean ± SD = 70.7 ± 7.0 cm) (Figure 3B). Statistically significant differences in size distributions between fisheries were found (P = 0.01). Only 5.4% of the 16,104 T. alalunga measured fell outside the length range of 60–90 cm L F.
Fig. 3. Fork length frequency distributions of Thunnus alalunga caught by albacore longline fishery (N = 12,152 fish) (A) and by recreational fishery (N = 3952) (B) in the western Mediterranean Sea.
Sex ratio
The overall sex ratio was 1.1♀:1♂ (n = 1486 fish). In total, 99.5% were measured (n = 1478 fish; 779 females and 699 males). The size of males ranged between 57.5 and 98.0 cm L F (mean = 74.7 ± 8.3 cm) and females ranged between 57.2 and 90.5 cm L F (mean = 69.4 ± 5.7 cm). Significant differences in the proportion of females were found between length classes (P < 0.01). The predictions of the GLM analysis showed that females predominated in the length classes less than 70 cm L F, whereas males were more abundant in the length classes greater than 75 cm L F; that is, the sex ratio was 1:1 for a 72.6 cm L F fish and for a length class of between approximately 71 and 74 cm L F (95% confidence interval) (Figure 4).
Fig. 4. Predicted trend in sex ratio (proportion female) with fork length. Dashed lines indicate confidence interval (±95% CI).
Histological classification
With the exception of the developing phase, all ovarian phases were represented in the ovarian samples examined (Figure 2). Histological analysis classified 583 females as mature and only four as immature. Immature females were caught in July; the smallest was 52 cm L F and the largest was 60 cm L F, whereas mature females ranged between 56 and 90 cm L F. Figure 5 shows the gonad phases of mature females by month. The absence of ovaries in the developing phase (the phase before the spawning capable phase) may have been due to the number of females sampled (four) between April and May. The only three ovaries collected in May were in the spawning capable phase. The majority of females were caught in June, July and August and their ovaries presented POFs (N = 548). Then, spawning ovaries were classified into the four subphases: Medium vitellogenic subphase (24.6%); advanced vitellogenic subphase (43.6%); migratory nucleus subphase (28.6%); and hydrated subphase (only 3.1%). Ovaries in the hydrated subphase represented 1.2% (n = 6) of the spawning ovaries analysed from the RF and 33.3% (n = 11) from the LLALB. A total of 32 ovaries (5.5% of mature females) were classed as either regressing or regenerating; only one of these ovaries was collected in August (Table 3 and Figure 5).
Fig. 5. Monthly relative frequency of gonad phases (by microscopic examination) for mature female Thunnus alalunga in the western Mediterranean Sea (immature fish are not included). Numbers on bars represent the number of individuals.
Table 3. Gonadosomatic index values for mature female according to their gonad phase. GSIa = ratio of gonad weight to fork length times 104; GSIb = ratio of gonad weight to total weight times 102.
Spawning season and spawning frequency
The GSI was calculated for 803 females (753 from the RF and 50 from the LLALB) and 736 males (682 from the RF and 54 from the LLALB) (Table 4). The mean monthly GSI values calculated by using the two formulas increased gradually from May and reached a peak during June and July (Figure 6). Although few ovaries (N = 39) and testes (N = 41) were sampled between August and November, a clear downward trend in GSI values was observed (Table 4 and Figure 6).
Fig. 6. Monthly variations of the gonadosomatic index (GSI) for mature female Thunnus alalunga (A) and for male Thunnus alalunga (B). GSIa = ratio of gonad weight to fork length times 104; GSIb = ratio of gonad weight to total weight times 102.
Table 4. Gonadosomatic index ranges for female and male Thunnus alalunga by month (pooled across years). GSIa = ratio of gonad weight to fork length times 104; GSIb = ratio of gonad weight to total weight times 102.
The maximum values of the GSI calculated in relation to the length (17.4) and weight (8.7) were observed in ovaries in the hydrated subphase (Table 3). Mean maximum GSI values were also observed in ovaries in the hydrated subphase (mean ± SD = 11.6 ± 3.7), whereas the minimum values were observed in ovaries in the regenerating phase (mean ± SD = 1.5 ± 0.6) (Table 3). The combined monthly GSI values for females and males and histological examination of the ovaries suggest that the spawning season for albacore in the western Mediterranean Sea is between June and August.
Histological analysis showed that all ovaries classified as spawning presented POFs. The POFs found in ovaries in the hydrated subphase were small and scarce (Figure 2) with the exception of one ovary in which the POFs were fresh and few hydrated oocytes were observed. The spawning fraction calculated considering all the mature females was 0.94 and the mean spawning interval was 1.06 days (i.e. 1/0.94). The proportion of active females whose ovaries had POFs (548) to all active females (551 ovaries, in phase III and IV) was 0.99, and the spawning frequency 1.01 days (i.e. 1/0.99). If only mature females caught during the peak spawning months (June and July) (N = 537) are considered, then the spawning fraction is 1; therefore, in these months, T. alalunga from the western Mediterranean Sea are capable of spawning every day.
DISCUSSION
Although T. alalunga is a resource of high economic value for both large-scale industrial and small-scale artisanal fisheries, the reproductive biology of the Mediterranean population of T. alalunga remains poorly understood (Juan-Jordá et al., Reference Juan-Jordá, Mosqueira, Freire and Dulvy2013). In the present study a total of 16,104 specimens were measured and 587 females were examined to determine their maturity status and to gauge their spawning season in the western Mediterranean Sea. Thus, this study presents new information on aspects of the reproductive biology of this species in the western Mediterranean that could be of value to future fishery management.
Only 5.4% of the large number of T. alalunga measured was less than 60 cm L F and greater than 90 cm L F. Specimens over 90 cm L F have been frequently observed in other oceans, e.g. up to 110 cm L F in the South Pacific Ocean (Otsu & Hansen, Reference Otsu and Hansen1962; Ramon & Bailey, Reference Ramon and Bailey1996; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013), 120 cm L F in the Indian and North Pacific Oceans (Wu & Kuo, Reference Wu and Kuo1993; Chen et al., Reference Chen, Crone and Hsu2010), and even 130 cm L F in the South Atlantic Ocean (Anon., 2012). Although statistically significant differences were found in size distributions between the RF and the LLALB, the length distribution pattern in the present study is consistent with that observed in other fisheries, including artisanal ones, throughout the Mediterranean Sea [e.g. Aegean Sea and Antalya Bay (eastern Mediterranean): 54−101 cm L F (Megalofonou, Reference Megalofonou1990, Reference Megalofonou2000; Karakulak et al., Reference Karakulak, Özgür, Gökoğlu and Emecan2011); Ionian, Adriatic and Tyrrhenian Seas (central-southern Mediterranean): 32−111 cm L F (Arena et al., Reference Arena, Potoschi and Cefali1980; Megalofonou, Reference Megalofonou2000; Marano et al., Reference Marano, De Zio, Pastorelli, Rositani and Ungaro2005; Consoli et al., Reference Consoli, Romeo, Battaglia, Castriota, Espósito and Andaloro2008; Di Natale et al., Reference Di Natale, Mangano, Potoschi and Valastro2011); and the present study (western Mediterranean): 33–109 cm L F].
The overall estimated sex ratio for T. alalunga in the western Mediterranean was 1.1♀:1♂. Similar results were reported for T. alalunga in the south-western Atlantic, South Pacific (Ratty et al., Reference Ratty, Laurs and Kelly1990; Anon., 2012) and eastern Mediterranean Sea (calculated from Megalofonou, Reference Megalofonou2000; Table I). However, T. alalunga males were more abundant than females in autumn in the Aegean Sea, with an estimated sex ratio of 0.48♀:1♂ (Megalofonou, Reference Megalofonou1990). The analysis of sex ratio by length showed that the sex ratio was close to 1:1 for length classes between 71 and 74 cm L F, with females being predominant in small length classes. A predominance of T. alalunga females in length classes less than 75 cm L F was found in the eastern Mediterranean, suggesting that females cease growing at smaller sizes than males (Karakulak et al., Reference Karakulak, Özgür, Gökoğlu and Emecan2011). A greater abundance of males in larger length classes (above 95 cm L F to 100 cm L F), has also been reported in T. alalunga from the Pacific Ocean (Chen et al., Reference Chen, Crone and Hsu2010; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013) and in other tuna species (Schaefer, Reference Schaefer1998; Aranda et al., Reference Aranda, Medina, Santos, Abascal and Galaz2013). According to Schaefer (Reference Schaefer, Block and Stevens2001) and Schaefer et al. (Reference Schaefer, Fuller and Miyabe2005), the almost complete absence of females in tuna species within larger size classes seems to be related to differences in natural mortality and vulnerability to capture rather than to differential growth. However, recent studies have reported that male T. alalunga from the eastern Mediterranean Sea (Megalofonou, Reference Megalofonou2000; Karakulak et al., Reference Karakulak, Özgür, Gökoğlu and Emecan2011) and the North and South Pacific (Chen et al., Reference Chen, Shimose, Tanabe, Chen and Hsu2012; Williams et al., Reference Williams, Farley, Hoyle, Davies and Nicol2012) grow faster and reach greater sizes at age than females, suggesting that once the size at maturity is reached, reproductive investment rather than somatic growth is higher in females than in males (Chen et al., Reference Chen, Shimose, Tanabe, Chen and Hsu2012; Williams et al., Reference Williams, Farley, Hoyle, Davies and Nicol2012; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013).
Histological analysis of the ovaries and GSI values (for males and females) suggest that the spawning season in the western Mediterranean Sea is from June to August. These results are in line with a larval study conducted in the Balearic Sea by Alemany et al. (Reference Alemany, Quintanilla, Velez-Belchí, García, Cortés, Rodríguez, Fernández de Puelles, González-Pola and López-Jurado2010), who suggested that the peak T. alalunga spawning season occurs in mid-July. The estimated spawning season duration of around 3 months was similar to that observed for male T. alalunga in the eastern Mediterranean Sea (Akayli et al., Reference Akayli, Karakulak, Oray and Yardimci2013), but much shorter than the 6 and 7 months reported for T. alalunga in the tropical waters of the South Atlantic (Anon., 2012) and the North and South Pacific (Chen et al., Reference Chen, Crone and Hsu2010; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013). It is known that T. alalunga larvae in the western Mediterranean Sea avoid temperatures lower than 23°C (Alemany et al., Reference Alemany, Quintanilla, Velez-Belchí, García, Cortés, Rodríguez, Fernández de Puelles, González-Pola and López-Jurado2010), which are only exceeded in summer in this area (Vargas-Yáñez et al., Reference Vargas-Yáñez, García Martínez, Moya, Tel, Parrilla, Plaza, Lavín, García, Salat, López-Jurado, Pascual, García Lafuente, Gomis, Álvarez, García Sotillo, González Pola, Polvorinos, Fraile Nuez, Fernández de Puelles and Zunino2010). Thus, the spawning period of T. alalunga in the Mediterranean Sea will rarely exceed 4 months, due to the inherent climatology of the Mediterranean Sea (the environmental conditions suitable for spawning are a surface water temperature of over 24°C; Schaefer, Reference Schaefer, Block and Stevens2001). These differences between the Mediterranean and oceanic populations in the duration of the spawning season are also found in other species. For example, it has been reported that the spawning period of little tunny Euthynnus alletteratus (Rafinesque, 1810) in the central Mediterranean Sea is between June and September (Hajjej et al., Reference Hajjej, Hattour, Allaya, Jarboui and Bouain2010), whereas in the Atlantic Ocean the spawning season is between about April and November (Collette & Nauen, Reference Collette and Nauen1983). Likewise, the spawning period of the swordfish Xiphias gladius is shorter, between June and August, in the western Mediterranean Sea (De la Serna et al., Reference De la Serna, Ortiz de Urbina and Macías1996) and between December and June in the north-west Atlantic Ocean (Arocha & Lee, Reference Arocha and Lee1996). Our findings are in line with those of Arocha (Reference Arocha2007), who reported that these differences are due to the fact that the 24°C isotherm only appears in the Mediterranean between late spring and late summer.
Given that the majority of the ovaries examined were collected from the RF, which operates during the day, the scarcity of females with hydrated oocytes would suggest that spawning takes place at night as reported for other tuna species (Schaefer, Reference Schaefer, Block and Stevens2001) including South Pacific T. alalunga (Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013). These results are in line with those of Hunter et al. (Reference Hunter, Macewicz and Sibert1986), who suggested that sampling should be conducted after 21:00 hours in order to capture a significant number of skipjack tuna Katsuwonus pelamis (L. 1758) with hydrated oocytes. Furthermore, the higher proportion of ovaries in the hydrated subphase in fish caught by the LLALB (N = 11) compared with those caught by the RF (N = 6) could be due to the fact that tournament fishing takes place during the day, whereas sets from the LLALB can also be placed during the night. Another explanation for these results may be due to the differences in the potential time-lapse between the moment the fish bite the hook and their death, i.e. the fish die almost immediately in a sports vessel, whereas tuna species such as albacore could be caught hours before eventually dying in a LLALB, and thus the stress caused by the capture could induce the hydration process (Hunter et al., Reference Hunter, Macewicz and Sibert1986). With the exception of one ovary in which the POFs were fresh and only few hydrated oocytes were observed, the small size and the scarcity of POFs in ovaries in the hydrated subphase together with the high surface water temperature of 25.2°C ± 1.3 (mean ± SD) from June to August in the western Mediterranean (www.ma.ieo.es/gcc) suggest that POFs are reabsorbed before the following batch.
All the ovaries in the spawning phase presented POFs. Thus, assuming that POFs do not persist for more than 24 h after ovulation, which has been confirmed in K. pelamis (Hunter et al., Reference Hunter, Macewicz and Sibert1986), T. albacares (Schaefer, Reference Schaefer1996) and South Pacific T. alalunga (Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013), the spawning frequency of T. alalunga in the western Mediterranean Sea would be almost daily. Similar spawning rates have been reported for other tuna species in ‘prime reproductive condition’: 1.11 days in K. pelamis (Hunter et al., Reference Hunter, Macewicz and Sibert1986) and 1.09 days in Thunnus obesus (Lowe 1839) (Nikaido et al., Reference Nikaido, Miyabe and Ueyanagi1991) in the western Pacific; 1.14 days in T. albacares in the eastern Pacific (Schaefer, Reference Schaefer1996); 1.1 days in Southern bluefin tuna Thunnus maccoyii (Castelnau 1872) in the south-east Indian Ocean (Farley & Davis, Reference Farley and Davis1998); and ~1.15 days in Atlantic bluefin tuna Thunnus thynnus (L. 1758) in the Balearic Sea (Medina et al., Reference Medina, Abascal, Megina and García2002; Aranda et al., Reference Aranda, Medina, Santos, Abascal and Galaz2013). However, spawning frequency during the peak spawning activity of North and South Pacific T. alalunga has been estimated at 1.7 days and 1.3 days, respectively (Chen et al., Reference Chen, Crone and Hsu2010; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013). Furthermore, differences have been found in estimated spawning frequencies according to the type of fishing gear used and the vertical distribution of spawning tuna (Itano, Reference Itano2000; Medina et al., Reference Medina, Abascal, Aragón, Mourente, Aranda, Galaz, Belmonte, De la Serna and García2007). Thus, to address the issue of possible sampling bias, it would be very useful to conduct a comparative study on the estimated spawning frequency of Mediterranean T. alalunga based on catches using non-shallow gear.
The estimated GSI values of T. alalunga from the western Mediterranean (current study) are strikingly different from those estimated in other oceans. The mean GSIa value for the hydrated subphase was 11.6 in T. alalunga from the western Mediterranean, whereas it was 5.3 in North Pacific T. alalunga (Chen et al., Reference Chen, Crone and Hsu2010). Similarly, the highest GSIa value of 17.4 in T. alalunga from the western Mediterranean is higher than the value of 5.2 (Ramon & Bailey, Reference Ramon and Bailey1996) or the value of nearly 7 (Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013) in South Pacific T. alalunga. Likewise, the largest GSIb values found in male and female T. alalunga in the western Mediterranean were 8.6 and 9.2, respectively, which are much higher than the values of 0.97 and 2.23 in male and female T. alalunga in the Indian Ocean, respectively (Wu & Kuo, Reference Wu and Kuo1993). However, our results are in line with those of Akayli et al. (Reference Akayli, Karakulak, Oray and Yardimci2013). These authors studied male T. alalunga in the eastern Mediterranean Sea and found similar mean GSIb values in June and July (greater than 3.0).
It has been suggested that tuna species would reach the spawning grounds following a migration pattern structured by age and size: larger or older fish would arrive in advance and would therefore have a longer spawning season than smaller or younger fish (Schaefer, Reference Schaefer1998; Zudaire et al., Reference Zudaire, Murua, Grande and Bodin2013). Our results show that there were no significant differences in size distributions among months, therefore indicating that the T. alalunga spawning in the western Mediterranean does not follow the spawning migration pattern suggested above and that estimated size at maturity would not differ based on a time-limited sampling scheme. Given that only five immature individuals were found, it was not possible to estimate the length at 50% maturity (L50) using the most common method, i.e. fitting a logistic model to the proportion of mature fish by age-length class. Histological analysis showed that the minimum size at maturity for females was 56 cm L F. Similar minimum size at maturity were reported by Akayli et al. (Reference Akayli, Karakulak, Oray and Yardimci2013) for males (63 cm L F) in the eastern Mediterranean and by Arena et al. (Reference Arena, Potoschi and Cefali1980), who estimated the age at first maturity of 2 years (mean size 66.3 cm L F) and the minimum size of 62 cm L F in the Tyrrhenian Sea. In contrast, a minimum size at maturity of between 71 and 90 cm L F has been reported in North and South Pacific T. alalunga (Otsu & Uchida, Reference Otsu and Uchida1959; Otsu & Hansen, Reference Otsu and Hansen1962; Ratty et al., Reference Ratty, Laurs and Kelly1990; Ramon & Bailey, Reference Ramon and Bailey1996; Chen et al., Reference Chen, Crone and Hsu2010; Farley et al., Reference Farley, Williams, Hoyle, Davies and Nicol2013) and the L50 of 87 cm L F has recently been estimated for South Pacific T. alalunga by Farley et al. (Reference Farley, Hoyle, Eveson, Williams, Davies and Nicol2014). These estimates are in agreement with those found in T. alalunga in the Indian Ocean by Kikawa & Ferraro (Reference Kikawa and Ferraro1967) and by Wu & Kuo (Reference Wu and Kuo1993), who estimated that T. alalunga reached maturity at a minimum length of between 90 and 96 cm L F, and between 83 and 90 cm L F, respectively. Based on studies by Bard, the L50 and age at 50% maturity (A50) of North Atlantic T. alalunga are assumed to be 90 cm L F and 5 year (Bard, 1981 in ICCAT, 2010). The L50 of South Atlantic T. alalunga is estimated to be 92 cm L F for females and 100 cm L F for males (Anon., 2012). Therefore, T. alalunga in the Mediterranean Sea reaches sexual maturity at a smaller size than all of their counterparts in all other regions worldwide; suggesting different reproductive tactics between the Mediterranean T. alalunga population and the oceanic ones. Other species also follow this pattern of earlier maturation. The smallest L50 of X. gladius, 142 cm LJFL (lower jaw – fork length), has been found for the Mediterranean population (De la Serna et al., Reference De la Serna, Ortiz de Urbina and Macías1996). In contrast, in the Indian Ocean the L50 for female X. gladius was estimated at 170 cm LJFL by Poisson & Fauvel (Reference Poisson and Fauvel2009) and in the North Atlantic Ocean it was estimated at 179 cm LJFL by Arocha & Lee (Reference Arocha and Lee1996). However, in the South Atlantic the L50 has been estimated at 146.5 cm LJFL, which is slightly larger than that estimated for the Mediterranean population (Mejuto, Reference Mejuto2007).
In summary, the present study provides new information on some key aspects of the reproductive biology of T. alalunga in western Mediterranean Sea. The results suggest that the reproductive traits (lower maximal length, higher GSI values during the spawning season, lower minimum length at maturity and shorter spawning season) of T. alalunga in this area differ from those determined for the oceanic populations. These differences are consistent with those found between Mediterranean and North Atlantic T. alalunga in previous studies, such as different growth rates (Megalofonou, Reference Megalofonou2000), different gill parasites (Mele et al., Reference Mele, Merellab, Macias, Gómez, Garippab and Alemany2010) and genetic differences (Arrizabalaga et al., Reference Arrizabalaga, Costas, Juste, González-Garcés, Nieto and López-Rodas2004; Nakadate et al., Reference Nakadate, Viñas, Corriero, Clarke, Suzuki and Chow2005; Davies et al., Reference Davies, Gosling, Was, Brophy and Tysklind2011; Montes et al., Reference Montes, Iriondo, Manzano, Arrizabalaga, Jiménez, Pardo, Goñi, Davies and Estonba2012; Albaina et al., Reference Albaina, Iriondo, Velado, Laconcha, Zarraonaindia, Arrizabalaga, Pardo, Lutcavage, Grant and Estonba2013), suggesting a reproductive isolation of the Mediterranean T. alalunga (Davies et al., Reference Davies, Gosling, Was, Brophy and Tysklind2011). It may be the case that the reproductive strategy of Mediterranean T. alalunga would represent an adaptation to the environmental conditions of the Mediterranean Sea. Additional studies incorporating other spawning sampling localities within the Mediterranean Sea are needed for further and better understanding of the differences in reproductive strategies between the T. alalunga Mediterranean population and the oceanic populations.
ACKNOWLEDGEMENTS
We would like to acknowledge the skippers and crews of the fishing vessels who worked voluntarily with the IEO-OP and the scientific observers for sample collections on the longline vessels. We also thank Federació Balear de Pesca i Casting, the Federación Española de Pesca y Casting, all the sports judges, the recreational fishing clubs (S'Estanyol, Dénia, Torrevieja, Jávea, Sóller, Cala D'Or, Port Balís and Torredembarra) and the fishing participants for their collaboration during the samplings. We are very grateful to the IEO located in Palma de Mallorca for providing facilities during the samplings in Mallorca, our colleagues of the Large Pelagic Fisheries department of Málaga, S. Mele, T. Dobinson and P. Bárcenas for their assistance. We would also like to thank the anonymous referees whose suggestions improved the submitted manuscript.
FINANCIAL SUPPORT
This study was financially supported by the Instituto Español de Oceanografía projects GPM-3, GPM-4 (included an IEO-PhD grant to S. Saber) and GPM-1213, and by the DCF (Data Collection Framework) (EC) No. 199/2008.
