Annual reproductive cycle and fecundity of Scorpaena notata (Teleostei: Scorpaenidae)*

The family Scorpaenidae is of particular interest from the reproductive point of view, since it is made up of species with a wide variety of reproductive strategies, ranging from the most basic oviparity to matrotrophic viviparity. Wourms and Lombardi (1992) consider that, specifically within the genus Scorpaena, there is a shift from a primitive to a specialized mode of oviparity. Fertilization is still external and development is ovuliparous, but eggs are embedded in a gelatinous matrix. Scorpaena notata Rafinesque 1810 is a common species in rocky coastal habitats, and is found at depths of up to 700 meters. It is the object of semiindustrial and small-scale fishing (Whitehead et al., 1986; Fischer et al., 1987). It appears in the Mediterranean Sea and adjacent areas of the Atlantic, Madeira, the Azores and the Cabo Verde Islands. The southern limit of distribution seems to be Senegal (Eschmeyer, 1969). Until recently, most aspects related to its reproduction were unknown, but a recent study into its ovarian structure and process of oogenesis (Muñoz et al., 2002a) revealed that it is an ovuliparous species that shows several features that can be considered intermediate


INTRODUCTION
The family Scorpaenidae is of particular interest from the reproductive point of view, since it is made up of species with a wide variety of reproductive strategies, ranging from the most basic oviparity to matrotrophic viviparity. Wourms and Lombardi (1992) consider that, specifically within the genus Scorpaena, there is a shift from a primitive to a specialized mode of oviparity. Fertilization is still external and development is ovuliparous, but eggs are embedded in a gelatinous matrix.
Scorpaena notata Rafinesque 1810 is a common species in rocky coastal habitats, and is found at depths of up to 700 meters. It is the object of semiindustrial and small-scale fishing (Whitehead et al., 1986;Fischer et al., 1987). It appears in the Mediterranean Sea and adjacent areas of the Atlantic, Madeira, the Azores and the Cabo Verde Islands. The southern limit of distribution seems to be Senegal (Eschmeyer, 1969). Until recently, most aspects related to its reproduction were unknown, but a recent study into its ovarian structure and process of oogenesis (Muñoz et al., 2002a) revealed that it is an ovuliparous species that shows several features that can be considered intermediate between the most basic oviparity and the first transitional stages towards viviparity. Muñoz et al., (2002b) have already examined the histology of this species and the ultrastructure of the testes and the spermatogenic phases. These authors showed that the gonadal structure is intermediate between the restricted and unrestricted spermatogonial types of testes defined by Grier (1981Grier ( , 1993 and that the spermatogenesis is semicystic, which has been described in very few species of fish.
In this study, we aim to provide an in-depth analysis of the annual cycle of S. notata by studying the seasonal histological changes of the gonads and of various indices related to its reproduction. We also estimate and explain the fecundity of this species in relation to its specialized reproductive strategy.

MATERIALS AND METHODS
For the description of the different stages of maturity of the gonads, the ovaries and the testes were embedded in Histosec 56-58 pF (Merck) or in hydroxyethyl methacrylate, and were sectioned at between 4-10 µm, depending on the sex and the stage of maturity. Transverse and longitudinal sections were obtained for both sexes. The following stains were used for the samples kept in Histosec: haematoxylin-eosin for general histology; Mallory as a trichrome; PAS reaction (periodic acid-Schiff) for the demonstration of neutral mucopolysaccharides; and Alcian blue for acid mucopolysaccharides. The samples kept in methacrylate were stained with methylene blue-basic fuchsin, toluidine blue, and also PAS.
The stages of development in the oocytes were determined by following the criteria established by Wallace and Selman (1981) and West (1990). The ovaries were classified according to the more developed type of oocyte (West, 1990). The development stage of the testis was determined following the criteria laid down by Grier (1981).
In order to study the indices related to reproduction, we used 507 individuals of Scorpaena notata which were caught during a year-long period at various ports along the Costa Brava (northwest Mediterranean). The fish were fixed immediately after capture in 10% formaldehyde and preserved in 4% formaldehyde. The following parameters were analysed: -Sexual dimorphism. The total standard lengths and weights of the males and females were compared using an analysis of variance (ANOVA).
-Sex Ratio (SR = Number of males / Number of females) and the monthly variations of this index. We determined if the result was significantly different from 1 by means of the c 2 test.
All the indices were calculated as a function of the eviscerated weight, in order to avoid possible variations arising from differences in the digestive tract contents or energy reserves of the specimens. We calculated the indices, separating the fish according to sex and month of capture. Later, we observed any significant differences in the monthly variations by means of an analysis of variance (ANOVA). All the statistical analyses in this section were carried out in line with the criteria set out by Sokal and Rohlf (1995) with the programme SPSS 12.OS for Windows.
Fecundity was estimated in 45 females using the gravimetric method (Burd and Howlett, 1974;Hunter et al., 1985). The oocytes were separated by introducing samples of completely mature ovaries into Gilson's solution, as modified by Simpson (1951). The eggs were then filtered and once they had been sorted into different diameters they were counted. We repeated the process twice for each ovary. The individual or absolute fecundity refers to the number of eggs produced per female per year (Wootton, 1979), and can be defined as the number of mature oocytes present in the ovary immediately before spawning (Bagenal, 1973). In species that use multiple spawning, it is the number of oocytes destined for spawning, i.e. the ones that will mature during the current reproductive cycle, which are usually taken into account (Aboussouan and Lahaye, 1979). Therefore only oocytes with a diameter greater than the oocytes at the cortical alveoli stage were taken into account, since only these are considered to have been released in this reproductive cycle. This absolute fecundity tends to increase according to the size and age of the fish. Therefore, in order to facilitate the comparison we also calculated the relative fecundity, i.e. the number of eggs per unit eviscerated weight (Bagenal, 1978). In order to study the relationship between the fecundity and the size or the total weight of the individual, we used the linear regression analysis by means of the logarithm log Y = log a + b . log X. This is calculated using the least squares method and corre-sponds to an exponential function of the type: Y = a . X b . The significance levels are the same as we described above. Finally, we also determined the frequency distribution of the egg diameters.

Testes
The lobular structure of the testes can be seen clearly during November and December, since they are in the spermatogonial proliferation period. According to Grier (1981), they are characterized by the fact that the lobular lumens only contain a few spermatogonial cysts here and there on the periphery, which are always enclosed in the Sertoli cells which cover the inside of the seminiferous lobule. During the months between January and May, the testes are in the early recrudescence period: the lobular lumens are full of spermatocytes, and usually free of cysts (Fig. 1A). In May, the testes enter the mid-recrudescence period, and now contain germinal cells in all stages of development: spermatogonia, spermatocytes and spermatids. Already little groups of free spermatozoa can be seen in the lobular lumen (Fig. 1B). From June onwards, the lobules still show all the cited stages, but especially spermatids in various stages of development as well as spermatozoa: the testes are in the late recrudescence period (Fig. 1C). During the functional maturity period, which occurs from July to September, the lobules and all the ducts are full of sperm. A great quantity of PAS negative substance is detected within the lobular lumens (Fig. 1D). Finally, in September, there are no spermatozoa in any ducts or other regions of the testes because they are now in the post-spawning period. Spermatogonia become more and more abundant.

Ovaries
From November to March, oogonia and oocytes at various stages of development can be observed. This is the period of previtellogenis (Fig. 1E). The vitellogenic period begins approximately in June, when the largest oocytes exhibit yolk granules that grow progressively (Fig. 1F). During the period of maturation, between July and October, a lot of mature oocytes which are full of yolk granules as well as oocytes with migrated germinal vesicle and hydrated oocytes are detected (Fig. 1G). The ovary also contains postovulatory follicles, so it can therefore be assumed that spawning takes place within this period. During the periods of vitellogenesis and spawning, the internal epithelium of the ovarian wall has cytoplasmic projections and the lumen of the ovary contains PAS positive ovarian fluid, which is particularly abundant and viscous during spawning (Fig. 1H). Table 1 shows the averages obtained for the standard lengths and total weights of the males and females of Scorpaena notata. There are no significant differences between the values obtained for the two sexes.

Reproductive indices
The annual and monthly sex ratio values are shown in Table 2: 60.7% of the 507 individuals are male and the remaining 39.3% female, so the sex ratio is 1.5 which differs in a highly significant way from 1 (c 2 =24.434, g.d.l.=1, p=0.000). Figure 2 shows the annual development of the various indices we analysed in relation to the phase the gonad is in. The gonadosomatic index (GSI) shows highly significant differences for both sexes (ANOVA, p=0.000), and the maximum values appear between June and October in males, and between July and September in females. For the hepatosomatic index (HSI), which also shows high-REPRODUCTIVE BIOLOGY OF SCORPAENA NOTATA 557  ly significant monthly changes (ANOVA, p=0.000), the highest values appear from January to May in males, and from January to July in females. The condition of the individuals we studied (K) showed highly significant differences among the males but no significant differences among the females (ANOVA, p=0.000 and p=0.256, respectively). The profile of the condition factor is not shown in Figure    2 because in both sexes the annual development of the index is not very marked, with only slight decreases during late summer. Mesenteric fat was not detected in any of the specimens analysed.

Fecundity
The results obtained for absolute and relative fecundities are presented in Table 3. The distribution of the eggs in terms of frequencies of diameters is relatively open and often marked by two peaks. The relationship of the absolute and relative fecundity with the size and total weight of the specimens was significant in all cases.

DISCUSSION
During November, for Scorpaena notata both the ovaries and the testes are in the initial phases of development. In January, the hepatosomatic index (HSI) of the females begins to increase continuously and constantly, reaching much higher values than in the case of the males, but this is a common feature in many species of fish.
The males enter a phase of mid-recrudescence in May, which leads to a progressive increase in testicular activity as well as a significant decrease in their HSI. In the case of the females, the growth in GSI is more sudden and occurs later, beginning in July, which also brings with it a sharp fall in HSI in the following month. The transitory increase in the relative weight of the liver just before the increase in weight of the gonads suggests mobilization and processing of fatty acids and carbohydrates (Bruslé and González, 1996).
The spawning period of the scorpionfish is clearly delimited between July and October, a period which coincides, although it is slightly longer, with the period given for the Gulf of Leon (Duclerc and Aldebert, 1968) and the Algiers region (Siblot-Bouteflika, 1976). On the other hand, Fischer et al., (1987) stated that the reproduction of this species in the Mediterranean probably occurs in May, a hypothesis which we consider to be erroneous.
The condition of the individuals does not seem to be affected much by reproductive activity. The explanation probably lies in the major importance of the liver as an organ for storage, since no mesentery fat reserves were detected, nor were there seasonal changes in alimentary behaviour (Harmelin-Vivien et al., 1989) that could alleviate such a drain on resources. In the same vein, Shchepkin (1971) found that the liver of S. porcus had a high fat content compared to the muscles, and pointed out the major role of the liver as an organ for storing reserves to be used during reproduction. This information and the results about the evolution of the HSI obtained here suggest that the liver of the scorpionfish is the main provider of energy during the maturation and spawning periods.
The population we studied of Scorpaena notata is clearly dominated by males. The proportion of males is also higher in the Algiers region (Siblot-Bouteflika, 1976) andin Marseille (Kaim-Malka andJacob, 1985), although the difference is not as marked as it is in this study. In contrast, Bradai and Bouain (1991) studied the reproduction of two species of the same genus and found that, while S. scrofa had similar numbers of males and females, the population of S. porcus was clearly dominated by females, especially among larger-sized individuals. These authors felt that the results indicated a faster growth rate in the S. porcus females, a characteristic also attributed to S. guttata (Love et al., 1987). In the case of S. notata, the large inequality in the sex ratio cannot be connected with sexual differences in growth rates, since male dominance appears above all in medium-sized individuals (Muñoz et al., 1996). Furthermore, the morphometric analysis carried out showed no significant variations in size between males and females. However, the sex ratio obtained from a monthly population analysis also rules out the segregation of the sexes during spawning, a behaviour which in some species leads to different sex ratios of the unit (deMartini and Fountain, 1981;Alheit et al., 1984;Barbieri et al., 1992;among others). One possible explanation would be the existence of different distribution patterns between males and females, but there is insufficient data to confirm this idea.
When Scorpaena notata spawns, the number of eggs per female ranges from approximately 6000 to 33000. The analysis of the fecundity of another species of the same genus, S. porcus, carried out by Bradai and Bouain (1991), gave similar figures for total egg number and distribution of frequencies of diameters. The distribution of oocytes in various stages of development indicates that spawning is multiple, in such a way that the release of the more mature group is followed by the development and spawning of the following group. If we compare similar-sized individuals of S. notata captured at the beginning and end of the reproductive period (exam-ples with standard length = 140 mm from July and examples with standard length = 141 from September), we can see that the total egg number per female decreases as the spawning period goes on, a characteristic feature of a species with a determined fecundity (Greer Walker et al., 1994). At the same time, the decrease in relative fecundity would also make the worsening condition of the individuals clear, as a result of the drain on resources brought about by reproduction. These tendencies remain even if the individuals from September are larger (examples with standard length = 149 or 152), despite the fact that fecundity increases significantly as the standard length and the weight of the fish increases.
It should be noted that the fecundity of S. notata is relatively low compared with other species from the same Scorpaeniformes order, whether they are typically oviparous species, such as Trigla lyra (Muñoz et al., 2002c), with a maximum number of eggs which our own studies found to be around 108000 per female (Muñoz, 2001), or whether they are species of a much more viviparous nature, such as the zygoparous species Helicolenus dactylopterus, with a maximum egg count of 87000 per female (Muñoz and Casadevall, 2002). This low fecundity may be related to the specialized reproductive behaviour of the species studied in this paper. S. notata releases the spawn within a gelatinous mass segregated by an internal epithelium of the ovaric wall, which may have various functions, such as that it enables the spawn to float as well as provides mechanical protection and defence against predators (Muñoz et al., 2002a). However, if we combine the data we obtained with the data from our previous work on testicular structure and spermatogenesis of the same species (Muñoz et al., 2002b), another function of this gelatinous mass, which is perhaps the most important, becomes apparent: it keeps the spawn together. The abundant, viscous seminal fluid probably keeps the sperms together when they are released. If they are released onto the grouped mass of eggs within the gelatinous matrix, fertilization is assured, therefore reducing the need for the female to produce numerous eggs which would explain the low fecundity of the species. Observations on mating scorpionfish during certain times of the year seem to bear out this hypothesis.
It has been observed that the maximum egg diameter of Scorpaena notata is about 500 mm. This figure was obtained by both histological measurement and after being fixed in Gilson liquid.
However, it must be pointed out that the eggs of the same species that were found in the sea, floating within the gelatinous matrix, measured between 760 and 880 mm (Spartà, 1956;Kimura et al., 1989). This difference in size is probably partly due to the decrease in volume of the oocytes that occurs when they are fixed in formol (between 0 and 10% according to Fleming and Ng, 1987;Hislop and Bell, 1987; Lowerre- Barbieri and Barbieri, 1993), as well as the significant increase in volume in the eggs of some species when they come into contact with the marine environment.