Introduction

Pinna nobilis L., 1758 is the largest Mediterranean bivalve mollusc and is usually found anchored among Posidonia oceanica shoots, hidden by the leaves of the seagrass (Vicente 1990) and with the shell strongly overgrown by epibionts (Zavodnik 1967). In general, typical densities of individuals are relatively low. Zavodnik et al. (1991) estimated a mean of nine individuals/100 m2 in the Adriatic Sea and Vicente et al. (1980) and Moreteau and Vicente (1982) one individual/100 m2 in Port Cros National Park (France). In south Spanish coasts, densities of ten individuals/100 m2 are common (Templado 2001). Only in particular habitats, such as some littoral lagoons or sheltered coasts, the density reaches higher values of up to six individuals/m2 (De Gaulejac and Vicente 1990; Catsiki and Catsilieri 1992), although these are seldom observed exceptional populations. Therefore, great sampling efforts are necessary to obtain an inventory of individuals, because large underwater areas must be explored and the utilization of indirect remote-video techniques are not always suitable, since the majority of individuals are hidden by the leaves of P. oceanica. Consequently, they must be located directly by SCUBA diving, whose technical limitations (mainly limited working time, water transparency and weather conditions), reduce the performance of the surveys. As indicated by Richardson et al. (2004), it is difficult to perform the necessary repeated surveys to study the species ecology. Therefore, only a few surveys on this subject have been recently developed (cf. Porcheddu et al. 1998; Siletic and Peharda 2003; Katsanevakis 2005) and many aspects of the species ecology remain still unknown.

Vicente (1990), Ramos (1998) and Templado (2001) have also indicated the fragmentary knowledge of P. nobilis biology, even in fundamental aspects. Previous studies have evidenced the existence of peculiar features of the population structure in this species, some of which have not been deeply studied up to date. Some of them are concerned with its spatial distribution: depth-related size segregation, i.e. the occurrence of smaller individuals, and absence of larger ones, in shallower waters (Zavodnik 1967; Vicente et al. 1980; Moreteau and Vicente 1982; Vicente 1990; Vicente and Moreteau 1991; Barberá et al. 1996, unpublished data; Templado 2001), patchiness distribution (Combelles et al. 1986; Butler et al. 1993), and common orientation of the valves in some localities (Combelles et al. 1986; De Gaulejac and Vicente 1990; Butler et al. 1993).

The first hypothesis postulated to explain the size segregation of P. nobilis refers to the displacement capacity of the individuals. According to this idea, young individuals would implant at shallower sites and would move towards deeper ones during ontogeny (Zavodnik 1967; Vicente et al. 1980; Moreteau and Vicente 1982; Vicente 1990; Vicente and Moreteau 1991; Barberá et al. 1996, unpublished data; Templado 2001). However, the only prove of the change of position is the localization of some old individuals lying on the sand and later re-implanted some metres away from their original location (Moreteau and Vicente 1982; Butler et al. 1993; Vicente, personal communication) and the experimentally proved capacity of re-implantation of small individuals (up to 20 cm in total length) (De Gaulejac and Vicente 1990). The supposed mechanism for this movement would be crawling through the substratum, in a similar way to many other Pteriomorphia (Smeathers and Vincent 1979; Price 1983; Reimer and Tedengren 1997; Taylor et al. 1997) either attaching new byssus filaments in the direction of the movement and liberating the old rear ones, or shedding the entire byssus complex and producing a new one for re-implantation as indicated by Price (1983) and Taylor et al. (1997). The displacement using the foot as is done by many Veneroida is less probable since in the Pinnidae the foot is reduced and too small in relation to shell size (Yonge 1953). Other ideas suggested to explain this peculiarity of P. nobilis population structure are related to higher mortality of large individuals in shallower waters, either by amateur collection (Vicente 1990) or by the increased hydrodynamic stress respect to deeper sites (García-March 2003).

Combelles et al. (1986) suggested that the patchiness distribution is mainly related to habitat availability (e.g. the presence of Posidonia meadows), while the common orientation of the valves is the result of an increased mortality rate of inadequately oriented individuals in relation to some factor (i.e. deep currents or the slope of the bottom).

Anyway, the agreement of the displacement hypothesis makes it necessary to discriminate between those ecological features actually produced by differences in mortality and those due to an active displacement of individuals according to some kind of gradient. However, if the displacement hypothesis is not accepted, one question remains: what is causing the generally observed size segregation with depth?

In order to characterize the main factors influencing P. nobilis population structure from an holistic point of view and improve the understanding of P. nobilis autoecology, different aspects of the species demography were studied in two groups on individuals inhabiting at different depths (at −6 and −13 m) in the same P. oceanica meadow. On these sites, the individuals were inventoried, tagged and their positions accurately recorded inside known areas. Displacement evidences, common valve orientation, density of individuals, spatial distribution, size distribution, growth and mortality were studied periodically during a 5 years period (from July 1997 to July 2002).

Furthermore, the effects of collection of large individuals by amateur divers and the influence of hydrodynamic stress on the population structure were also considered. The results of the experiments carried out to investigate the shell drag coefficients and drag force supported by the individuals of both areas, as well as the estimation of P. nobilis dislodgement force, are described in García-March et al. in press.

Material and methods

The research was developed on a dense population of P. nobilis located in Moraira bay (Alicante, E of Spain), at the south of La Nao Cape, at 38°41′N and 0°8′W, from July 1997 to July 2002 (Fig. 1). Moraira shoreline is dominated by medium and low cliffs formed by ancient dunes of Würmiens eolianites, with small sandy beaches between them (Sanjaume 1985).

Fig. 1
figure 1

Location of Moraira bay in Western Mediterranean with sampling stations A1 (−13 m) and A2 (−6 m), where all Pinna nobilis individuals were inventoried and catalogued for their subsequent study

Full size image

Posidonia oceanica meadows are frequent in shallow waters in the coast of Alicante where they reach depths of c. 20 m and are characterized by high foliar density (Torres et al. 1990a, b). Preliminary exploration of Moraira Posidonia meadows reported a high foliar index up to 500 shots/m2, according to Giraud’s (1977) method, and a bathymetric range between 3 and 22 m. Observations on meadow topography indicate that it is irregular at shallow levels, with many channels, pools and small front steps, and more homogeneous at deeper levels, with some high front steps (between 1 and 2 m high) and large channels among large P. oceanica patches at some places.

Pinna nobilis was occasionally found in hollows among large rocks or in sandy sheltered areas at shallower levels. However, most of the individuals were observed in the meadow itself. In a previous study, García-March (2005) put in evidence that the density of individuals increased along a depth interval between 3 and 15 m, from one individual/100 m2 at 3 m, 6.5 ± 3.1 at the range 5–10 m and to 10.7 ± 4.3 individuals/100 m2 at the range 10–15 m. Likewise, mean size also augmented with depth increment, 25.8 ± 10.3 cm at 3 m, 35 ± 11.9 cm at the depth range 5–10 m and 49.1 ± 9.8 cm in the depth interval of 10–15 m. Therefore, the zone was selected as an optimal locality to study the evolution of a P. nobilis population and to test the different hypothesis proposed to explain the size segregation, spatial distribution model and common shell orientation.

Two areas of study, A1, located at 13 m depth and with a surface of 1,000 m2, and A2, located at 6 m depth and with a surface of 314 m2, were selected for the research (Fig. 1, 2). A smaller surface in A2 was studied, since individuals displacement was expected to be higher in shallower levels and, thus, surveys were performed more often on this site (each 3 months instead of each year for A1).

Fig. 2
figure 2

Design of the two areas where all Pinna nobilis individuals were inventoried in Moraira bay: A1, surface = 1,000 m2; A2, surface = 314 m2

Full size image

All individuals located in the studied areas were tagged and their positions recorded respect to the vertex of triangles formed by stainless-steel rods driven in the bottom. Three triangles were deployed in A1 and two in A2. The smallest triangle in A2 was placed near a group of five individuals to increase precision in the case of occurring displacement of only a few centimetres (Fig. 2). The individuals were located searching inside concentric circles of different radius (5, 10 and 15 m). Divers placed marking buoys near each individual found during the localization surveys. Each specimen was registered, labelled with slackly tight bridles and positioned measuring its distance to the triangle sticks.

The antero-posterior length outside the substratum (unburied length, UL), the maximum and minimum dorso-ventral lengths, usually called width (W and w, respectively), and the orientation of the valves respect to the hinge line of each specimen were measured (Fig. 3). In situ measurements were performed with a tree calliper for the period 1997–1998 and with the multi-calliper proposed by García-March et al. (2002) from 1999 to 2002. Total sizes were estimated with the equation by García-March and Ferrer (1995).

Fig. 3
figure 3

Biometry of Pinna nobilis: W maximum width; w minimum width; UL unburied length; h buried length; Or orientation; b bend

Full size image

The study of spatial distribution was carried out using the position of the individuals in both areas. The analysis of the data was developed using the Poisson’s series as mathematical model and the χ2-test (goodness-of-fit) for agreement with a Poisson series (Elliott 1983). The surface of each area was subdivided on the paper in 9 m2 grids because of the adjustment of the units to the scale of the distribution study, since the occurrence of individuals inside the grids becomes a rare event, what agrees with the premises of the statistical method used for the study.

To evaluate the differences of mortality between depths, the percentage of dead individuals and the total mortality coefficient (z) (Gulland 1988) were calculated in each area after the first 3 years of survey. To discriminate if mortality is size-dependent, all individuals (alive and dead) of each area were grouped in 15 cm size-sets, and comparisons between expected and observed mortalities were done by means of a χ2-test.

Shell-length data of the different individuals during the period 1997–2002 were adjusted to the Von Bertalanffy growth equation, according to the method proposed by Wang (2004) for multiple recaptures, when initial ages for all individuals are unknown. To test if the growth patterns are the same in each group of individuals, the generalised T2-Statistic (Anderson 1971) was used.

Finally, orientation of the shell was measured taking as reference point the bend (b) formed by the hinge-line of the animals (Fig. 3). The measurements were fitted to the following four 90° intervals of the cardinal points: N(315°–45°), S(135°–225°), E(45°–135°) and W(225°–315°). Two tests were carried out for each group to discriminate if the orientation of the hinge-line is significant or not. In the first case, the four orientations are compared, N, S, E and W and, in the second one, the intervals are compared in groups of 2: N–S against E–W. To check if some relationship exists between the orientation of the individuals and the wave exposure of the shore, the wave exposure index proposed by Thomas (1982), which reflects the shoreline exposure to hydrodynamism according to the slope of the beach, the fetch and the intensity and direction of preferential winds in the area, was calculated. Wind data were recorded from the nearest meteorological station in San Antonio Cape Lighthouse of the Spanish Institute of Meteorology (SIM). To test the significance of the results, orientation distributions were compared to a uniform distribution using a χ2, as indicated by Fisher (2000) for grouped data.

The individuals were monitored annually at A1 and every three months at A2, resulting in a total of 6 and 21 samplings for each group, respectively. All the individuals of both areas were measured in each survey during the period 1997–2000. Afterwards, only controls of subgroups of animals were performed. For that reason, the study of size segregation, spatial distribution, mortality and orientation was developed using the data of July 2000, whereas for the studies of growth and movement the 1997–2002 period data.

Results

All the inventoried individuals of both areas could be easily relocated at each new survey from their measured distances to the stainless-steel sticks. Soon, it was evident that the repetition of observations after short time periods to control displacement of individuals was not necessary, because individuals were always in the same previously observed location. After the 5 years of study, all live specimens remained in the same place of the first observation.

As it can be deduced from the displacement data, neither individual changed its position during the monitored period. Consequently, to perform the orientation study, only the orientation data of individuals alive in the last complete survey (year 2000) were necessary for the analysis. The results show that it is not possible to discard random shell orientation in A1 (−13 m) (P = 0.663 and 0.375 for the orientations N, S, E and W and N–S versus E–W, respectively). By contrast, the differences are highly significant in A2 (−6 m) (P = 0.007) when N–S against E–W orientations are considered, and non significant when the four 90° intervals N, S, E and W are considered (P = 0.055), although in this last case χ2-test is very close to significant results. Indeed, in A2, 16 of the 20 individuals have their valves oriented towards N–S direction (Fig. 4). In this context, Thomas’s (1982) exposure index shows that the maximum wave exposure lays between WSW and ESE in both areas, being more intense at 6 m depth, where the A2 is located (Fig. 5) (Table 1).

Fig. 4
figure 4

Shell orientation of alive Pinna nobilis in both areas

Full size image
Fig. 5
figure 5

Comparison between Thomas’s (1982) exposure index and Pinna nobilis shell orientation. Numbers inside figures representing the total of individuals oriented on each 90° orientation interval (N, S, E and W)

Full size image
Table 1 Observed (O) and expected (E) number of individuals for each area, oriented in the pre-established range of orientations respect to the cardinal points in the year 2000
Full size table

As expected from previous observations performed in Moraira bay, density of individuals is higher in A1 at 13 m depth (103 specimens alive, 10.3 individuals/100 m2) than in A2 at 6 m depth (20 specimens alive, six individuals/100 m2), closely matching with the previous density estimations in the same depths cited above (García-March 2005). The current result, together with the previous data, confirms that individuals are progressively more aggregated as depth increases.

Nevertheless, the study of the individual’s spatial distribution in both inventoried areas shows not enough evidence to affirm that the specimens are not randomly distributed (A1, χ2=0.41, df = 3, P > 0.5, A2, χ2=0.98, df = 2 and 0.3 < P < 0.5). Therefore, although the individuals tend to be more aggregated with the depth increment, at the smaller scale of the studied areas they tend to be neither aggregated nor evenly distributed.

In the same way, the size distribution also shows significant statistical differences between both sites, as expected from previous author observations. Mean size is higher at 13 m depth (A1) than at 6 m depth (A2) (A1, x = 47.2 ± 9.86 cm; A2, x = 29.8 ± 7.39 cm, P < 0.001). Not only the group of individuals with smaller specimens is located in the shallowest zone, but also the proportion of small individuals is greater and the minimum sizes are also smaller in this site (Fig. 6). Furthermore, in situ observations have clearly shown that at −6 m individuals are smaller than at −13 m, both in the studied areas and in adjacent places, and direct observations on line transects perpendicular to the shore show that larger individuals are found deeper.

Fig. 6
figure 6

Size distribution of live Pinna nobilis in both areas in year 2000: a area 1; b area 2

Full size image

With regard to the study of in vivo growth, repeated measurements of the inventoried individuals soon indicated different growth rhythms in both areas, as it was subsequently demonstrated by the fitting of the Von Bertalanffy growth equations. Differences were observed both at the rate at which maximum asymptotic size is reached [K (y−1)] and in the maximum size reached [Linf (cm)]. However, striking inverse differences are observed. In A2, at 6 m depth, the growth rate at which the asymptotic size is reached is distinctly superior to that of A1, whereas the asymptotic size is clearly inferior than that of A1 (Fig. 7) [A1, L = 56.03(1 − e−0.17t); A2, L = 37.59(1 − e−0.40t)], indicating that individuals grow to similar sizes during the first 5–6 years but subsequently there is a sharp growth brake in A2. The differences between growth curves are significant, with a P < 0.001. Therefore, there are important differences in growth rates between the two groups of individuals, with smaller maximum size at the lower depths, and consequently, “dwarf” individuals occur in A2, compared to those of A1. The different parameters of the Von Bertalanffy equations, as well as their confidence limits can be observed in Table 2.

Fig. 7
figure 7

Von Bertalanffy growth equations of Pinna nobilis located in the two surveyed areas: A1, area 1 at 13 m depth, L = 56.03(1 − e−0.17t); A2, area 2 at 6 m depth, L = 37.59(1 − e−0.40t)

Full size image
Table 2 Estimated parameters of the Von Bertalanffy growth curve for both groups of individuals
Full size table

Annual mortality was homogeneous during the first 2 years and increased drastically in 2000 in both sites (Mean annual mortality A1 = 10.7 ± 6.4, A2 = 5.3 ± 4 individuals) (Table 3). Total mortality has however, shown important differences between both sites, since much more individuals died in A2 than in A1, in relation to total individuals density (A2: 16 dead individuals out of 36, 44.4%; A1: 32 dead individuals out of 135, 23.7%); total mortality coefficient being zA2=0.20 and zA1=0.09, respectively. Regarding to size-group mortalities on each area, significant differences occur in A1 (0.05 < P< 0.01, df = 2), where the individuals up to 30 cm died much more often than expected from a random distributed mortality on each size group, and the individuals of the 31–45 cm size-group died considerably less often than expected. The same tendency does not occur in A2 and the mortality is concentrated on the larger individuals (31–45 cm), although the differences among the observed and expected deaths are not statistically significant (Table 4). Total mortality coefficients calculated for each size group show important differences between sites for the range 31–45 cm (zA1(31–45)= 0.05, zA2(31–45)= 0.25), being more similar for the range 0–31 cm (zA1(0–31)= 0.25, zA2(0–31)= 0.15) (Table 3).

Table 3 Mean (Y) and total number of dead individuals, total mortality coefficient (z) and percentage of dead individuals (%) in each size group in the period 1997–2000 in both area 1 (A1) and area 2 (A2) groups of individuals
Full size table
Table 4 Observed and expected death of P. nobilis in each area in the year 2000, assuming mortality as a random event. A1, area 1, 0.01 < P < 0.05, df = 2. A2, area 2, 0.3 < P < 0.5, df = 1
Full size table

Discussion

The population structure of P. nobilis in Moraira bay follows similar trends to those observed before in other P. nobilis populations, with a depth-related size segregation, contagious distribution (as an increment in number of individuals with depth) and valve orientation in some sites (Zavodnik 1967; Vicente et al. 1980; Moreteau and Vicente 1982; Vicente 1990; Vicente and Moreteau 1991; Butler et al. 1993; Barberá et al. 1996, unpublished data; Templado 2001). However, the additional data on mortality, small-scale spatial distribution, individuals displacement and growth, allow completing the general view of the structure of P. nobilis population in Moraira bay. The result is a global picture where causes and effects can be related, except for some particular ecological features, beyond the scope of the present study, that should be investigated in future surveys.

The observation of young individuals in both shallow (6 m depth) and deep levels (13 m depth) proves that larvae settle along all the depth range, as already was supposed by Combelles et al. (1986), after the observation, in the P. nobilis population of Port Cros, of young individuals also implanted in deep zones. Larval behaviour could also have some influence on population structure of P. nobilis, increasing the prevalence of small individuals in the shallowest sites, similarly to Tellina spp. and Mulinina lateralis, where the active maintenance of a preferred depth keeps the spat within 5 km of shore at Duck (NC, USA) (Shanks and Brink 2005). However, in the present study, the proximity of both areas to the shore, < 1 km, and one from the other, circa 500 m, with no unevenness in the meadow between them, makes it possible to assume an homogeneous recruitment in both areas. Also, if enhanced recruitment at the shallow area had been of relevance, a higher number of individuals, increased by the presence of a large amount of new recruits, should have been observed in A2 respect to A1. Leaving aside the short period after settlement (whose characteristics are presently unknown), in which some degree of movement or reorientation is perfectly possible, the displacement survey has shown that once a young individual is large enough to be detected, it remains in the same position until its death. Mussels which have the ability of crawling on the substrate and release the byssus, are able to produce a new byssus complex in periods ranging from hours to few days (Glaus 1968; Price 1983; Côté 1995; Uryu et al. 1996; Reimer and Tedengren 1997). However, in Croatian waters, P. nobilis needed 4–5 months to regenerate the dissected byssus (Mihailinovic 1955). It is difficult to assume that P. nobilis could shed the byssus voluntarily, since the individuals would be unattached for a long period, during which predators such as Octopus vulgaris would have more facilities to handle detached specimens (Fiorito and Gherardi 1999). Also, the slow production of byssus threads by P. nobilis indicates that displacements by means of the byssus filaments would be extremely slow, apart from the fact that displacement should be done through the substratum, not crawling on it. Moreover, in the present survey, inventoried individuals neither moved nor changed their orientation. Consequently, displacement can be definitely discarded as the cause of observed population structure. On the other hand, the process of site selection during larva settlement seems to be a random event, not influenced by the proximity or distance to conspecifics, as it is indicated by the results of the spatial distribution study.

After implantation, Von Bertalanffy growth equations indicate that fast growth in the first 4–5 years of life takes young individuals to a total antero-posterior length of 20–30 cm. Individuals inhabiting the shallowest sites reach this size faster, but later suffer a sharp growth decrease limiting their maximum sizes to around 37 cm of total length. The exact causes of the growth stop are not known, although excessive resuspended matter and higher hydrodynamic stress in shallower areas could be the responsible factors (García-March, et al. in press). In concordance with the present results, Richardson et al. (1999) observed considerable differences in growth rates and maximum sizes in four populations of P. nobilis from Almería (South–East of Spain, Western Mediterranean). Although they did not consider this aspect deeply, their data show that the individuals from Rodalquilar, whose population inhabited the shallowest site (< 5 m depth), also showed the smallest maximum size (< 35 cm) and the smallest estimated maximum size (45.7 cm), together with the highest K value (0.28), in the Von Bertalanffy growth equation. Thus, although new surveys in other populations are necessary to valuate the frequency of such growth differences, the agreement between the present results and those obtained by Richardson et al. (1999) indicates that the differences in growth, related with depth, could be usual in many other populations.

Under a practical perspective, the observed differences in growth parameters mean that the size-age distributions of populations located at different depths in the same meadow cannot be compared, because the greatest the depth difference, the biggest the difference between size–age relationships among populations and the possibility to confuse adult individuals of shallower populations with young specimens of deeper ones. What is more, the most exposed the shore, the most important these differences would be. Moreover, the growing tendency of individuals total size in relation to depth, together with the differences in growth equations in both groups of individuals, show a gradual evolution between growth parameters of populations located in shallow waters and those located at deeper ones, introducing the necessity of a new form of calculating P. nobilis growth taxes, as an integrated continuous evolution of growth parameters with the increase of depth. For that reason, subsequent studies about growth and age of P. nobilis populations should be stratified in narrow depth intervals as the best approximation.

On the other hand, mortality data shows that the group of small individuals is under a higher selective pressure throughout the Moraira bay and many of them died before reaching larger sizes, as demonstrated by the high values of total mortality coefficients (z) of this size-group in both areas. However, the destiny of the survivors above 30 cm length is drastically different on each site. Whereas in A1 the individuals between 31 and 45 cm have the minimum z of all size-groups in both areas, hence being in an optimal location for their survival, in A2 these groups of individuals, which represent the largest specimens at 6 m depth, have a relatively huge z (0.05 vs. 0.25, respectively) (Table 3), being strongly decimated. These observations are in consistence with the idea suggested by Combelles et al. (1986) that some factor, with anthropic or natural origin, could be conditioning the survival of the largest individuals in the shallowest areas. It is improbable that collection of large individuals in A2 is the cause of the adult mortality in this shallow area of Moraira bay, since the zone is far from the shore, individuals are completely hidden by the leaves of P. oceanica and amateur divers seldom approximate to the location of the experimental area.

However, the shell drag force experiments simultaneously developed, have shown that, given a wave height, individuals of shallower levels are subjected to higher drag forces than those located at deeper ones in Moraira bay. Also, lateral valve orientation suppose a drag force increment of more than fourfold respect to the dorso-ventral position of the valves (García-March et al. in press). Thus, above a minimum threshold of water speed, more frequently reached in shallower levels (i.e. A2), both large individuals and those badly oriented with regard to the main water direction, are in disadvantage for survival. Therefore, according to Thomas’s (1982) shore exposure index, those specimens oriented E–W in Moraira bay are suffering much more hydrodynamic stress than those oriented N–S. Indeed, A2 preferential valve orientation matches with the optimum dorso-ventral orientation N–S in this bay. These results are consistent with those observed by Combelles et al. (1986), since the preferential orientation N–S coincides with the slope of the bottom. However, the fact that valve orientation matches with the slope of the bottom does not mean that mortality of individuals with improper orientation respect to near bottom currents bringing food, is the causing factor in Moraira bay, chiefly because it is only observed in the shallowest site. Furthermore, the combination of a N–S (dorso-ventral) orientation together with the smallest sizes in A2, strongly reduce mean drag force, being always lower than that observed in A1, whatever the wave height and wave period are (García-March et al. in press). Consequently, all indicates that, in the shallowest sites, maximum total sizes and shell orientation are controlled by hydrodynamic stress, causing the death of larger and improperly oriented individuals, and by some kind of endogenous biological limitation of the maximum size, as suggested above.

In this context, it has been demonstrated that extreme wave heights, produced during huge storms, have an essential regulation capacity over the population structure of P. nobilis, inasmuch as they kill the largest individuals inhabiting deeper waters, seldom subjected to such high water speeds (García-March, et al. in press). These events, the strongest probably occurring each 22 years, according to the climate cycles observed in the western Mediterranean (Duarte et al. 1999), produce a cycled decline in population density by the punctual mortality increase of large individuals and, what is more important, a drastic reduction of the mean age of the population, probably killing many adult specimens before reaching their potential maximum age.

The irregular topography of the P. oceanica meadow, with many channels, pools and front steps, can alter water direction and speed in some sites (Granata et al. 2001), so that some individuals can take advantage of those niches and grow larger or with theoretical improper orientations than expected from hydrodynamic forces calculations. In fact, in Moraira bay some large individuals are observed in the corner of a front step in an apparently sheltered zone. Therefore, in the “picture” of Moraira P. nobilis population structure, sometimes large individuals are observed among many smaller ones.

Summarising, in the general view of Moraira P. nobilis population, mean sizes (not strictly age-related), and number of individuals, gradually increase with depth throughout the bay, due to the reduction of hydrodynamic stress and the consequent decrease of its selective pressure. The general reduction of stress does not influence small-scale spatial distribution, since it occurs at a large scale, affecting similarly all the individuals from the same depth range, except those inhabiting special niches, as indicated above. In a global view, it seems that juvenile individuals preferentially settle at lower depths and afterwards move toward deeper areas as they grow; the occurrence of dwarf individuals in the shallowest zones strongly contributes to this impression.

Likewise, the observation of size segregation with the increment of depth in many other Mediterranean coasts could be also consequence of the effects of hydrodynamic stress on individual’s survival and growth. In fact, they could be the causes of the population structure observed by Combelles et al. (1986) in Port Cros.