INTRODUCTION
Information on life-history patterns of deep-sea fauna has been improved greatly in the past few years (Ramirez-Llodra, Reference Ramirez-Llodra2002; Young, Reference Young and Tyler2003). In the Atlantic Ocean, the echinoderms are an important component of the bathyal and abyssal fauna in terms of density, biomass and ecosystem function (Gage et al., Reference Gage, Pearson, Clark, Paterson and Tyler1983; Billett, Reference Billett1991; Ginger et al., Reference Ginger, Billett, Mackenzie, Neto, Boardman, Santos, Horsfall and Wolff2001; Wigham, et al., Reference Wigham, Hudson, Billett and Wolff2003), and their life history has been described for many species (Tyler et al., Reference Tyler, Grant, Pain and Gage1982a, Reference Tyler, Young, Billett and Giles1992, Tyler, Reference Tyler1983, Gage et al., Reference Gage, Tyler and Nichols1986, Galley et al., Reference Galley, Tyler, Smith and Clarke2008, Benítez-Villalobos & Díaz-Martínez, Reference Benútez-Villalobos and Dúaz-Martúnez2010, Ross et al., Reference Ross, Hamel and Mercier2013). A wide diversity of reproductive patterns has been reported for deep-sea seastars, from the most common quasi-continuous reproductive patterns to seasonal patterns (Tyler & Pain, Reference Tyler and Pain1982a, Reference Tyler and Painb). Also, there is a diversity in fecundity from high to low assets (Ramirez-Llodra et al., Reference Ramirez-Llodra, Tyler and Billett2002; Young, Reference Young and Tyler2003), and brooding or broadcasting strategies have been described (Mercier & Hamel, Reference Mercier and Hamel2008). In the deep Mediterranean Sea, the most abundant groups are fish and decapod crustaceans (Company et al., Reference Company, Maiorano, Tselepides, Plaity, Politou, Sardá and Rotllant2004; Danovaro et al., Reference Danovaro, Company, Corinaldesi, D'onghia, Galil, Gambi, Gooday, Lampadariou, Luna, Morigi, Olu, Polymenakou, Ramirez-Llodra, Sabbatini, Sardá, Sibuet and Tselepides2010; Tecchio et al., Reference Tecchio, Ramirez-Llodra, Sardá and Company2011; Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Ramirez-Llodra, Rotllant, Recasens, Murua, Quaggio-Grassiotto and Company2013a), and consequently deep-sea echinoderms have been less studied (Alvà, Reference Alvá1987; Mecho et al., Reference Mecho, Billett, Ramirez-Llodra, Aguzzi, Tyler and Company2014) and their life history remains mostly unknown.
Ceramaster grenadensis (Perrier, 1881) belongs to the Family Goniasteridae and presents a wide geographic and bathymetric distribution. It is present both in the Atlantic Ocean and the Mediterranean Sea between 200 and 2845 m depth (Clark & Downey, Reference Clark and Downey1992; Mecho et al., Reference Mecho, Billett, Ramirez-Llodra, Aguzzi, Tyler and Company2014). Its shallow bathymetric range is slightly deeper in the Mediterranean Sea, starting at 600 m, and its presence has been reported from the eastern, central and western Mediterranean basins (Koukouras et al., Reference Koukouras, Sinis, Bobori, Kazantzidis and Kitsos2007; Carlier et al., Reference Carlier, Le Guilloux, Olu, Sarrazin, Mastrototaro, Taviani and Clavier2009). This species is the most abundant seastar below 850 m in the north-western Mediterranean Sea (Mecho et al., Reference Mecho, Billett, Ramirez-Llodra, Aguzzi, Tyler and Company2014). Although C. grenadensis represents an important component of the deep benthic ecosystem, information about its general biology remains scarce. The trophic behaviour of some species within this genus has been studied, with some species (i.e. Ceramaster granularis (Retzius, 1783)) described as active sponge predators (Gale et al., Reference Gale, Hamel and Mercier2013) and others described as deposit feeders (i.e. Ceramaster patagonicus (Sladen, 1889)) (Anderson & Shimek, Reference Anderson and Shimek1993). Ceramaster grenadensis is a secondary consumer that may feed on decayed organic material (Carlier et al., Reference Carlier, Le Guilloux, Olu, Sarrazin, Mastrototaro, Taviani and Clavier2009). Life-history traits in relation to the reproductive biology are presently unknown. Thus, in this study we focused on the population distribution and the seasonal and bathymetric patterns of the reproductive biology of C. grenadensis along the middle and lower slope of the Catalan margin of the north-western Mediterranean Sea.
MATERIALS AND METHODS
Study area and sampling methods
From October 2008 to April 2013, ten oceanographic cruises were conducted in the north-western Mediterranean Sea. The sampling areas included the submarine canyons of Blanes, Palamós (also named La Fonera) and Cap de Creus, as well as the adjacent slope to the Blanes canyon (Figure 1). Because of the low number of individuals sampled inside canyon areas (N = 3), no comparative analyses between habitats were conducted. These cruises took place in the framework of three Spanish research projects (PROMETEO, DOSMARES and PROMARES), sampling at depths comprised between 900 and 2250 m every 150 m.
Fig. 1. Study area representing the three sampled canyons and adjacent open slopes in the north-western Mediterranean Sea. Dark lines represent trawl stations.
A total of 141 hauls were conducted with a single warp Otter-trawl Maireta System (OTMS, Sardà et al., Reference Sardá, Cartes, Company and Albiol1998) with a net length of 25 m fitted with a cod-end liner of 6 mm mesh size. A SCANMAR system was used to estimate the width of the net aperture. The average horizontal opening was of 12.7 ± 1.4 m. The height of the trawl mouth was estimated to be 1.4 m (Sardà et al., Reference Sardá, Cartes, Company and Albiol1998). As the SCANMAR system can only operate down to 1200 m depth, the same value for the mouth's width of the net was used also for hauls deeper than 1200 m. In addition, 53 hauls were conducted with an Agassiz trawl, made of a square steel frame with a mouth width and height of 2.5 and 1.2 m respectively, and a cod-end liner of 6 mm mesh size. The total swept area for both gears was of 10.3 km2.
Because of the low number of individuals available at each depth, and based on available literature on bathymetrics of structuring deep-sea communities (Quetglas et al., Reference Quetglas, Carbonell and Sanchez2000; Company et al., Reference Company, Maiorano, Tselepides, Plaity, Politou, Sardá and Rotllant2004), the individuals were grouped in four bathymetric strata comprising the whole depth range sampled as follow: 900–1050, 1200–1350, 1500–1750 and 2000–2250 m depth. Density and biomass were standardized to km2 estimated from vessel speed, distance from initial and final trawl positions and average of the mouth opening of the sampling gear.
Samples analysis
All specimens (N = 141, Table 1) were sorted on board. The major radius (R) from the anus to the tip of the D arm was measured for all individuals. The specimens were then fixed with 40% formalin diluted with seawater and neutralized with borax. To study the population size distribution, R sizes were grouped into 5 mm classes and percentage frequency calculated.
Table 1. Number of individuals sampled during the oceanographic cruises. See trawl information in Figure 1 (DM, DOSMARES trawls; PR, PROMETEO trawls).
After 30 days, the samples were transferred to 70% alcohol in the laboratory and weighed to the nearest ±0.01 g. The 141 specimens were dissected on the oral side, from which the five pairs of gonads were extracted and weighed to the nearest ±0.001 g. The same procedure was conducted for the pyloric caeca. The gonad index (GI) and pyloric caecum index (PCI) were calculated as follows:
$$\hbox{GI} = \displaystyle{\hbox{Gonad}\,\hbox{Mass}\,(\hbox{g}) \over \hbox{Female}\,\hbox{Mass}\,(\hbox{g})} \times 100$$
$$\hbox{PCI} = \displaystyle{\hbox{Pyloric}\,\hbox{caeca}\,\hbox{Mass}\,(\hbox{g}) \over \hbox{Female}\,\hbox{Mass}\,(\hbox{g})} \times 100$$
After being weighed, the gonads were preserved in 70% alcohol prior to histological preparation. In a subsample of 125 individuals, gonads were dehydrated in graded alcohols, cleared in Histoclear and embedded in paraffin wax. The processed gonads were sectioned at 7 μm and stained with haematoxylin and eosin. These 125 individuals were used for sex identification. Juvenile specimens were considered as those small individuals presenting undifferentiated gonads from histological sections. Of the 125 individuals, 71 were females, of which 42 were further analysed to describe oogenesis. The Feret diameter of 100 oocytes (whenever possible) sectioned through the nucleus was measured for each individual using the image analysis package SigmaScan Pro 5. The Feret diameter gives the diameter of a disc with the equivalent area of the measured object. Oocyte sizes were grouped in 50 μm classes and percentage of oocyte-size frequency was calculated for each individual. The mean and SD for each size class in the sample (depth and season) was calculated and plotted.
Statistical analysis
All the data were tested for normality using the Kolmogorov–Smirnov non-parametric tests. Two Analyses of Variance (ANOVA) were used to test population size class differences between depths and between seasons. Meanwhile, the difference in density, biomass, GI and PCI between depths and between seasons were tested by Mann–Whitney non-parametric tests. Chi-squared test was used to analyse sex ratios in relation to depth and season.
RESULTS
Density and biomass
The bathymetric distribution of Ceramaster grenadensis ranged from 900 to 2250 m depth, with mean density values not statistically different along depth strata (Figure 2A) (Kruskal–Wallis test, H 3 = 1.35, P > 0.05). Although not significant, the maximum mean density was higher at 1500–1750 m depth. By contrast, a decrease in population biomass with depth was observed (Figure 2B). The mean population biomass distribution showed significant differences between the two shallower strata (900–1050 and 1200–1350 m) (Mann–Whitney U test, U = 68, N 1 = 12, N 2 = 24, P < 0.01) and between the shallower and deepest strata (900–1050 and 2000–2250 m) (Mann-Whitney U test, U = 18, N 1 = 12, N 2 = 12, P < 0.01).
Fig. 2. Density (ind km−2) (A) and biomass (kg km−2) (B) of C. grenadensis by depth strata. The top and bottom of each box-plot represent 75% (upper quartile) and 25% (lower quartile) of all values, respectively. The horizontal line is the median. The ends of the whiskers represent the 10th and 90th percentiles. Asterisks represent outliers.
Although mean density and biomass were not significantly different between seasons (Kruskal–Wallis test, H 3 = 0.31, P> 0.5 for density; Kruskal–Wallis test, H 3 = 1.47, P > 0.5 for biomass), a general trend of lower values in summer was found (Figure 3A, B).
Fig. 3. Density (ind km−2) (A) and biomass (kg km−2) (B) of C. grenadensis by season. The top and bottom of each box-plot represent 75% (upper quartile) and 25% (lower quartile) of all values, respectively. The horizontal line is the median. The ends of the whiskers represent the 10th and 90th percentiles. Cross marks represent outliers.
Population size distribution
The population structure of C. grenadensis describes a normal distribution (Kolmogorov–Smirnov, D = 120, df = 137, P < 0.0001, Lilliefors significance correction), with most specimens (70%) presenting size classes between R = 20 and 35 mm (Major Radium size) (Figure 4A, B). The population size frequency distribution presented a depth-related pattern, with size decreasing significantly with depth (ANOVA, F (3,134) = 91.79, P < 0.0001), but with no significant differences in the mean size of the population between the two middle depth strata (1200–1350 and 1500–1750 m). A mean individual size of R = 38 mm was observed at 900–1050 m stratum decreasing to R = 15 mm mean individual size at 2000–2250 m stratum (Figure 4A). The adults were distributed along the whole bathymetric range, while juvenile specimens (considering juvenile specimens to be those presenting undifferentiated gonads) were limited to the deepest areas below 1750 m depth. At the deepest stratum (2000–2250 m), the ratio of adults/juveniles was nearly 1:1.
Fig. 4. Individual size frequency distribution by depth (A) and season (B). Dark grey bars, females; light grey bars, males; white bars, juveniles.
Seasonally, the mean individual size decreased from summer (R = 31.9 mm) to spring, when the smallest individuals were captured (R = 22.5 mm) (Figure 4B). Significant differences between summer and all the other seasons were found (ANOVA, F (3,123) = 6.637, P < 0.0001). The individual size-frequency distribution by season showed the presence of juveniles during winter and spring, while in summer and autumn only adult individuals were captured. The smallest individual was caught at 2250 m depth in spring, with a R = 6.34 mm while the largest specimen was collected at 1050 m depth in autumn with a R = 45.26 mm.
Sex ratio and size at first maturity
Of the 125 specimens sexed, 71 were female and 36 male, with 18 juveniles where sex could not be determined. Sex ratio was significantly biased towards females 2:1 (Chi-square test, χ2 = 0.5, P > 0.05). No significant differences were found in the sex ratio with depth or season (Chi-square test, χ2 = 0.5, P > 0.05). Minimum size at sexual maturity was R = 10.9 mm for females and R = 13.3 mm for males. All the 18 juvenile specimens were found deeper than 1750 m depth.
Gonad morphology
Ceramaster grenadensis presented macroscopically the typical gonad morphology of seastars, with five pairs of gonads per individual, one pair in each interradius. Each pair of gonads was suspended in the coelom and was attached to the body wall by a short gonoduct opening aborally at the gonopore. Macroscopically, mature ovaries and testes presented the same morphology of tufts of digitate tubules and could not be distinguished. Although differences in the colour of gonads were observed (from beige to light brown), there was no general colour pattern for males and females in this species, with both sexes presenting all ranges of colour (Figure 5A).
Fig. 5. (A) Macroscopic view of a C. grenadensis gonad; (B) Histological sections of gonads of mature testis; (C) Juvenile specimen with indeterminate gonad; (D) Immature female with oogonia (Oo); (E) Section showing previtellogenic oocytes (PV) and small vitellogenic oocyte (VS); (F) Section showing large vitellogenic oocytes (VL) and, previtellogenic oocytes (PV).
To be able to determine sex, the gonads had to be processed histologically and observed under a microscope. Males in maturity stages presented spermatozoa accumulated as dense masses of gametes in the lumen of the testes (Figure 5B). Juveniles were classified as indeterminate specimens and they were identified by the presence of immature follicles, with no distinguishable oocytes or spermatozoa (Figure 5C). In immature females, only oogonia were present, observed as small cells (<100 μm) with a large nucleus/cytoplasm ratio (Figure 5D). In females with developing ovaries, we observed all stages of oogenesis present at any single time, including oogonia, previtellogenic and vitellogenic oocytes (Figure 5E). The previtellogenic oocytes (23–209 μm), presented a central nucleus with an eccentric nucleolus. These cells could be identified because their cytoplasm stains in dark purple with haematoxylin due to their basophilic composition. The vitellogenic oocytes were larger cells (between 136 and 691 μm) with a smaller nucleus/cytoplasm ratio resulting from the accumulation of vitellum in the cytoplasm. The acidophilic cytoplasm of vitellogenic oocytes stained pale pink with Eosin. In the vitellogenesis stage, the oocytes were usually small or medium in size (from ±130 to ±450 μm) (Figure 5E). Meanwhile, in the maturity stages, females presented previtellogenic (<200 μm) and large vitellogenic oocytes (from 450 to 691 μm) (Figure 5F).
Oocyte-size frequency distribution
The females of C. grenadensis presented a broad range of oocyte stages, independently of season or depth (Figure 6A, B), suggesting that oogenesis is asynchronous in this species. There was a wide peak of previtellogenic oocytes (<250 μm) representing 60–72% of the total measured oocytes in all seasons and 53–81% of the measured oocytes in all depth strata. Small vitellogenic oocytes (from 250–350 μm) represented 10–18% of the measured oocytes in all seasons and 1–23% of the measured oocytes in all depth strata. Developing vitellogenic oocytes (350–500 μm) represented 14–22% in all seasons and 8–23% of the measured oocytes at all depths. The largest oocytes (500–700 μm) represented less than 2% of the measured oocytes for all seasons and between 0.5 and 8% for all depth strata.
Fig. 6. Bathymetric oocyte-size frequency distribution (A) and seasonal oocyte-size frequency distribution (B).
Gonad Index (GI) and Pyloric Caecum Index (PCI)
The bathymetric values for GI and PCI were not constant bathymetrically (Table 2). In females, GI presented higher values at 2000–2250 m depth (Figure 7A). Significantly higher mean GI and PCI values for females were reported at the lowest depth stratum (2000–2250 m depth) (Figure 7A) (Table 2). In males, the same pattern of higher mean values at lower depths was observed (Figure 7B). The mean GI was significantly higher at the deepest depth stratum (Table 3). Meanwhile, the mean PCI increased significantly at 1500–1750 m stratum (Table 3), with a maximum at 2000–2250 m (Figure 7B) (Table 3).
Fig. 7. Bathymetric changes in Gonad Index (GI) and Pyloric Caeca Index (PCI) (mean ± SE) in females (A) and males (B).
Table 2. Mann–Whitney U values for females’ GI and PCI values between depth strata. Significance of the U values (U) is indicated using: *P < 0.1; **P < 0.05; ***P < 0.001. Distance values with no asterisk indicate non-significant values.
Table 3. Mann–Whitney U values for males GI and PCI values between depth strata. Significance of the U values (U) is indicated using: *P < 0.1; **P < 0.05; ***P < 0.001. Distance values with no asterisk indicate non-significant values.
No significant differences were observed in the mean GI values of females throughout the year (Figure 8A, Table 4). A decrease in mean PCI was observed in summer (Figure 8A), but the differences were not statistically significant (Table 4). In males, we observed a seasonal pattern, presenting a significant increase of mean GI in summer and autumn (Figure 8B, Table 5), while there were no significant differences in mean PCI throughout the year (Figure 8B, Table 5).
Fig. 8. Seasonal changes in Gonad Index (GI) and Pyloric Caeca Index (PCI) (mean±SE) in females (A) and males (B).
Table 4. Mann–Whitney U values for females GI and PCI values between seasons. Significance of the U values (U) is indicated using: *P < 0.1; **P< 0.05; ***P < 0.001. Distance values with no asterisk indicate non-significant values.
Table 5. Mann–Whitney U values for males GI and PCI values between seasons. Significance of the U values (U) is indicated using: *P < 0.1; **P< 0.05; ***P < 0.001. Distance values with no asterisk indicate non-significant values.
DISCUSSION
The density of the seastar Ceramaster grenadensis was similar at the different depths sampled on the continental margin of the north-western Mediterranean Sea, with a slight, but not statistically significant, increase at 1500–1750 m. This pattern coincides with what has been reported for other echinoderms, such as the holothurian Mesothuria (Allantis) intestinalis (Ascanius, 1805) Östergren, 1896, found in large numbers at 1600 m depth in the Balearic basin (Cartes et al., Reference Cartes, Maynou, Fanelli, Papiol and Lloris2009), and other invertebrate species (i.e. Aristeus antennatus (Risso, 1816)) (Sardà et al., Reference Sardá, Company and Maynou2003). Sporadic food inputs at these depths provided by particulate matter transport down nearby submarine canyons (Sanchez-Vidal et al., Reference Sanchez-Vidal, Pasqual, Kerhervõ, Calafat, Heussner, Palanques, Durrieu de Madron, Canals and Puig2008) have been suggested as one of the possible factors for the higher presence of invertebrates at these intermediate depths (Sardà et al., Reference Sardá, Company and Maynou2003; Cartes et al., Reference Cartes, Maynou, Fanelli, Papiol and Lloris2009). The biomass pattern of C. grenadensis presented the inverse trend to density, with a decrease with depth and a significantly lower biomass found at 2000–2200 m compared with the other depth strata. This pattern coincides with the biomass patterns observed for the whole megafauna from the same area (Tecchio et al., Reference Tecchio, Ramirez-Llodra, Aguzzi, Sanchez-Vidal, Flexas, Sardá and Company2013). This decrease of megafaunal biomass with depth in the north-western Mediterranean has been related to the decrease of the megafauna individual size (Stefanescu et al., Reference Stefanescu, Lloris and Rucabado1993; Moranta et al., Reference Moranta, Stefanescu, Massutú, Morales and Lloris1998).
A decreasing trend of the individual size with depth was also observed for C. grenadensis. If we considered exclusively adult specimens, the observed bathymetric changes in size structure were in accordance with the deep-sea dwarfism theory (Harvey et al., Reference Harvey, Gage, Billett, Clark and Paterson1988), and found in the irregular echinoid Brissopsis lyrifera (Forbes, 1841) in the Atlantic Ocean. Mecho et al. (Reference Mecho, Billett, Ramirez-Llodra, Aguzzi, Tyler and Company2014) also found this bathymetric size pattern in individuals of B. lyrifera, with the smallest individuals found at the deepest distribution range (1750–2250 m depth) of this species in the Mediterranean Sea. This phenomenon has also been observed in Molpadia musculus Risso, 1826 from the same study area (Mecho, unpublished data). The smaller size observed in certain deep-sea species has been related to the low food availability in the deeper areas compared with more productive shelf and upper slope habitats (Rex et al., Reference Rex, Etter, Morris, Crouse, McClain, Johnson, Stuart, Deming, Thies and Avery2006). Food limitation could be particularly important in the deep Mediterranean seafloor, because of the oligotrophic nature of the Mediterranean waters (Company et al., Reference Company, Sardá, Puig, Cartes and Palanques2003; Zúñiga et al., Reference Zúñiga, Flexas, Sanchez-Vidal, Coenjaerts, Calafat, Jordá, García-Orellana, Puigdefábregas, Canals, Espino, Sardá and Company2009).
Juvenile specimens of C. grenadensis were only caught deeper than 1750 m depth. The lower bathymetric distribution of the juveniles compared with the adult population has been observed also for other seastars, such as Luidia sarsii sarsii Duben & Koren, 1845 and Pontaster tenuispinus (Düben & Koren, 1846), both in the Porcupine Seabight (Sumida et al., Reference Sumida, Tyler and Billett2001; Howell et al., Reference Howell, Billett and Tyler2002). Additionally, a previous study carried out in the western Mediterranean Sea (South Balearic Sea) reported juveniles of C. grenadensis below the adult lowest depth of distribution, thus extending the maximum depth range of distribution for this species down to 2845 m (Mecho et al., Reference Mecho, Billett, Ramirez-Llodra, Aguzzi, Tyler and Company2014). The factors driving this depth-related recruitment pattern in several species of deep-sea seastars is not fully understood and further information on larval distribution and dispersal in relation to environmental variables is necessary (Howell et al., Reference Howell, Billett and Tyler2002).
The gonads and pyloric caeca indices increased with depth in both sexes, with maximum values observed at the deepest stratum (2000–2200 m). A higher GI value at the deeper species distribution range has been found also in the Atlantic seastar Solaster endeca (Linnaeus, 1771) (Ross et al., Reference Ross, Hamel and Mercier2013). Ross et al. (Reference Ross, Hamel and Mercier2013) attribute these higher GI values to the larger oocyte size of the individuals distributed at their maximum depth of distribution. However, no bathymetric differences in oocyte size were found in the present study. We suggest that, in the Mediterranean Sea, the C. grenadensis individuals dwelling at greater depths probably increase their reproductive effort by reducing the somatic investment (small sizes) and devoting a higher amount of energy to oocyte production (i.e. higher GI values). The PCI values of the C. grenadensis females caught at 2000–2200 m depth were four times higher than values recorded at 900–1050 m. This high PCI recorded from individuals in the lower slope suggests that these females may be storing nutrients in the pyloric caeca in order to maintain a constant production of gametes in a food-limited environment (Benítez-Villalobos & Díaz-Martínez, Reference Benútez-Villalobos and Dúaz-Martúnez2010), allowing a constant transfer of energy from the pyloric caeca to the gonads (McClintock et al., Reference Mcclintock, Watts, Marion and Hopkins1995).
In males, the gonad index (GI) values were significantly higher in summer and autumn, while lower (but non-significant) pyloric caeca index values were reported in summer, coinciding with other studies (McClintock et al., Reference Mcclintock, Watts, Marion and Hopkins1995). In females, the GI values were similar at all seasons. Our results suggest a transfer of nutrients from the pyloric caeca to the gonads during periods of gonadal growth (summer–autumn) in males and a constant production of oocytes in females through the year. Higher reproductive activity in autumn has been described also for deep-sea fishes in the same study area (Fernandez-Arcaya et al., Reference Fernandez-Arcaya, Recasens, Murua, Ramirez-Llodra, Rotllant and Company2012, Reference Fernandez-Arcaya, Ramirez-Llodra, Rotllant, Recasens, Murua, Quaggio-Grassiotto and Company2013a, Reference Fernandez-Arcaya, Rotllant, Ramirez-Llodra, Recasens, Aguzzi, Flexas, Sanchez-Vidal, López-Fernández, García and Companyb). We observed that, in males, PCI values started to decrease in spring and the GI started to increase in the following season (i.e. summer). In females, PCI decreased in summer with a subsequent increase of GI in autumn. Thus, males seem to have a maximum gonad development before females. Unless the time between gametes release in male and females is usually shorter (i.e. minutes or hours) (Mercier & Hamel, Reference Mercier and Hamel2008) here this could be an adaptation to high fertilization success by ensuring that sperm is available when females start spawning.
In females, the distribution of oocytes remains constant throughout the seasons and depths, with the ovary mainly occupied by previtellogenic and medium vitellogenic oocytes and a low percentage of vitellogenic oocytes in the largest size range. This constant presence of all kinds of oocytes throughout the year was also reported in the Atlantic Ocean for Bathybiaster vexillifer (W. Thomson, 1873) (Tyler et al., Reference Tyler, Pain and Gage1982b) and Styracaster elongatus Koehler, 1907 (Benítez-Villalobos & Díaz-Martínez, Reference Benútez-Villalobos and Dúaz-Martúnez2010). The same pattern but bathymetrically has been reported for Henricia lisa A. H. Clark 1949 (Mercier & Hamel, Reference Mercier and Hamel2008), where the deepest specimens maintained a constant presence of all kinds of oocytes.
In deep-sea echinoderms, quasi-continuous reproductive patterns as well as seasonal patterns have been described (Ramirez-Llodra, Reference Ramirez-Llodra2002; Mercier & Hamel, Reference Mercier and Hamel2008; Baillon et al., Reference Baillon, Hamel and Mercier2011). For deep-sea Mediterranean echinoderms, the reproductive biology has only been described for one species, the echinoid Brissopsis lyrifera, in the Gulf of Lion. This species has a well-defined seasonal pattern of reproduction, with maximum gonad maturation and spawning in autumn (Ferrand et al., Reference Ferrand, Vadon, Doumenc and Guile1988), producing small eggs between 15 and 80 μm. The ovaries of C. grenadensis from the Mediterranean presented a wide range of oocyte sizes at all seasons and depths, with only a few mature gametes at any single time. This oogenesis pattern is characteristic of semi-continuous reproduction (Young, Reference Young and Tyler2003). Thus, C. grenadensis follows a similar reproductive pattern to that described for many deep-sea echinoderms (e.g. Mcclintock et al., Reference Mcclintock, Watts, Marion and Hopkins1995; Ramirez-Llodra et al., Reference Ramirez-Llodra, Tyler and Billett2002; Galley et al., Reference Galley, Tyler, Smith and Clarke2008), which has been related to the low food availability at great depths (Mcclintock et al., Reference Mcclintock, Watts, Marion and Hopkins1995). The deep Mediterranean seafloor is particularly food-limited because of the overlying relatively oligotrophic waters (Company et al., Reference Company, Sardá, Puig, Cartes and Palanques2003; Tyler, Reference Tyler and Tyler2003; Zúñiga et al., Reference Zúñiga, Flexas, Sanchez-Vidal, Coenjaerts, Calafat, Jordá, García-Orellana, Puigdefábregas, Canals, Espino, Sardá and Company2009). We suggest that this environmental food limitation together with the capacity for storing and redistributing nutrients from the pyloric caecum (high PCI values all year round) play key roles in shaping the semi-continuous reproductive pattern observed in C. grenadensis.
The diversity of eggs sizes in echinoderms, estimated from the largest oocyte size in the ovary, results in different nutrient input in the spawned egg and thus differences in larval development (Tyler et al., Reference Tyler, Campos-Creasy, Giles, Young and Eckelbarger1994; Eckelbarger & Watling, Reference Eckelbarger and Watling1995). In C. grenadensis, the mature oocyte sizes are relatively large (650–700 μm), similar to other studied species of deep sea seastars in Atlantic waters, such as Styracaster elongatus Koehler, 1907 and Hyphalaster inermis Sladen, 1883 (Ramirez-Llodra et al., Reference Ramirez-Llodra, Tyler and Billett2002; Young, Reference Young and Tyler2003; Benítez-Villalobos & Díaz-Martínez, Reference Benútez-Villalobos and Dúaz-Martúnez2010). This large egg size should provide enough energy to the embryo to develop as a lecithotrophic pelagic larva until settlement (Benítez-Villalobos & Díaz-Martínez, Reference Benútez-Villalobos and Dúaz-Martúnez2010). The lecithotrophic larvae do not need to feed in the water column (Shilling & Manahan, Reference Shilling and Manahan1994) and thus provide an advantage in food-limited environments such as the deep Mediterranean Sea (Eckelbarger & Watling, Reference Eckelbarger and Watling1995; Ramirez-Llodra, Reference Ramirez-Llodra2002).
In summary, the population structure and reproductive biology of C. grenadensis has been described for the first time. The results suggest that adult size and somatic growth in C. grenadensis is lower in the individuals distributed at the deeper range of the species, where food quantity and quality are lower. Nonetheless, the reproductive output was higher in the individuals distributed at deeper depths, suggesting a higher investment in reproduction in detriment of somatic growth. Seasonally, highest values of GI were observed in autumn, suggesting a higher spawning capacity in this season. The gametogenesis of C. grenadensis from the bathyal Mediterranean Sea is similar to that of other deep-sea echinoderms, and showed a quasi-continuous production of sperm and large oocytes, which has been reported as a common reproductive pattern for deep-sea echinoderms.
ACKNOWLEDGEMENTS
The authors would like to thank the Officers and Crews of the RV ‘Garcia del Cid’ and RV ‘Sarmiento de Gamboa’ and the scientific parties of the PROMETEO, DOSMARES and PROMARES cruises, for their contributions at sea.
FINANCIAL SUPPORT
This study was supported by three projects funded by the Spanish Government: PROMETEO (CTM2007-66316-C02/MAR), PROMARES – OASIS DEL MAR (Obra Social ‘la Caixa’) and DOSMARES (CTM2010-21810-C03-03). JA was a fellow of the Spanish Ramón y Cajal Programme. ERLL was partially funded by a Spanish Ramón y Cajal Programme.
