Introduction

For several decades, occurrences of harmful benthic ‘microalgal’Footnote 1 blooms have strongly increased worldwide in frequency and intensity. This trend is likely to continue (Rhodes 2011). Among the Dinophyceae (kingdom Alveolata), these phenomena are observed in particular for the genus Ostreopsis. In the past, species of Ostreopsis were often observed in tropical and subtropical areas (Ballantine et al. 1988; Morton et al. 1992; Grzebyk et al. 1994). In 1995, the worldwide distribution of Ostreopsis spp. ranged between latitudes 28° N and 20° S (Faust et al. 1996). However, today, its distribution area has considerably expanded (Rhodes 2011) and extends between latitudes 45° N and 35° S. Over the last decade, massive proliferations of Ostreopsis spp. have become more common in temperate areas. Two species are mainly involved, Ostreopsis siamensis Schmidt and Ostreopsis ovata Fukuyo (Rhodes 2011). In the southern hemisphere, proliferation of Ostreopsis spp. has been reported in New Zealand (Chang et al. 2000; Rhodes et al. 2000; Shears and Ross 2009), Australia (Pearse et al. 2001; Rhodes 2011), Brazil (Granéli et al. 2002) and Hawaii (Parsons and Preskitt 2007). In the northern hemisphere, outbreaks of Ostreopsis spp. were reported in Japan (Taniyama et al. 2003; Rhodes 2011), Russia (Selina and Orlova 2010; Selina and Levchenko 2011) and throughout the Mediterranean Sea. Concerning the Mediterranean Sea, the authors referred to the observed material as Ostreopsis sp., Ostreopsis spp., O. ovata, O. cf. ovata, O. siamensis, and O. cf. siamensis. In the western Mediterranean, two distinct species have been reported, O. cf. siamensis from Catalonia, Andalusia and Sicily and O. ovata from Catalonia, the Balearic Islands, the Ligurian coast and the southern Tyrrhenian Sea (Penna et al. 2005, 2010; Battocchi et al. 2010; Bravo et al. 2010). The first observation of Ostreopsis sp. dates back to 1972 at Villefranche-sur-Mer (southern France) (Taylor 1979), but the first bloom event of O. cf. ovata occurred in 1998 in Tuscany (Italy) (Sansoni et al. 2003). Subsequently, Ostreopsis spp. outbreaks have been reported in the Balearic islands (Vila et al. 2001a, 2001b; Penna et al. 2005; Battocchi et al. 2010), the Ligurian Sea (Ciminiello et al. 2006; Mangialajo et al. 2008, 2011; Cohu et al. 2011), the Tyrrhenian Sea (Tognetto et al. 1995; Simoni et al. 2003; Penna et al. 2005; Zingone et al. 2006; Guerrini et al. 2010), the Adriatic Sea (Monti et al. 2007; Battocchi et al. 2010; Totti et al. 2010; Guerrini et al. 2010), the Aegean Sea (Aligizaki and Nikolaidis 2006; Aligizaki et al. 2008), Lebanon (Abboud-Abi Saab 1989; Abboud-Abi Saab and El-Bakht 1998) and Tunisia (Turki 2005). These massive proliferations of Ostreopsis spp. are usually associated with two other taxa of Dinophyceae, namely P. lima and Coolia spp. (Vila et al. 2001a; Aligizaki and Nikolaidis 2006) and occur in shallow waters close to the shoreline on macroalgal substrata. The environmental conditions required for blooming are unclear and seem to be very complex (Cohu et al. 2011). Species of the genus Ostreopsis produce palytoxin-like toxins and derivates (Ciminiello et al. 2008). This toxin is one of the most potent natural non-protein compounds known to date. Palytoxin-like compounds differ, depending on the Ostreopsis species incriminated: O. siamensis produces the ostreocin-d (Usami et al. 1995; Ukena et al. 2001) whereas Ostreopsis cf. ovata produces ovatoxin-a (Ciminiello et al. 2008). The impact of massive proliferations of Ostreopsis is harmful to both human health and ecosystems, such as causing negative effects on benthic ‘invertebrates’,Footnote 2 e.g. a loss of spines in the sea urchin Evechinus chloroticus (Valenciennes, 1846) (Simoni et al. 2003; Shears and Ross 2009; Shears and Ross 2010; Faimali et al. 2012; but see Blanfuné et al. 2012). Adverse effects on human health such as respiratory difficulties, conjunctivitis and skin irritation can occur either by direct contact with cells and by inhalation (Simoni et al. 2003; Brescianini et al. 2006; Ciminiello et al. 2006; Mangialajo et al. 2008; Tichadou et al. 2010) or by indirect contact through ingestion of seafood (Ito et al. 1996; Yasumoto 1998; Rhodes et al. 2002; Taniyama et al. 2003; Deeds and Schwartz 2010). To date, no intoxication by ingestion has been reported in the Mediterranean Sea (Aligizaki et al. 2008), and symptoms disappear at the same rate as the intensity of the blooms decreases (Durando et al. 2007). In urbanized areas, blooms of Ostreopsis spp. can induce beach closures or evacuation of the shoreline, resulting in economic loss for the tourism industry (Sansoni et al. 2003; Ciminiello et al. 2006; Mangialajo et al. 2008; Lemée et al. 2012). Furthermore, in the case of bloom events in mussel culture areas, the closure of fish farms can be imposed, also inducing economic loss.

Since 2008, further studies have been undertaken along the French Mediterranean coast, but only a few sites were studied; every year, blooms reappear during summer time. They start in June and never persist beyond August (Mangialajo et al. 2011; Cohu et al. 2011, 2013).

The goal of the present study was (i) to assess the distribution and abundance of Ostreopsis spp., together with associated Dinophyceae, along the French Mediterranean Coast from the Spanish to the Italian border and around Corsica Island (3 186 km, measured on a 1/2 500 map) and (ii) to investigate the possible relationship between the abundance of the studied dinophycean taxa and the macroalgal substrates.

Materials and methods

During summer 2011, three sampling campaigns were carried out: the first in June (weeks 23–24), the second in July (weeks 27–28) and the third in August (weeks 31–32). Seventy-nine sites were studied along the French Mediterranean coast (Fig. 1): 20 sites in French Catalonia and Languedoc (LC1 through LC20), 39 in Provence and the French Riviera (PR1 through PR39) and 20 in Corsica (CO1 through CO20). Only three campaigns were carried out, as the kinetics of the bloom was beyond the scope of the present study. Locality, acronym and coordinates of each study site are given in Table 1.

Fig. 1
figure 1

The 79 sites studied along the French Mediterranean coast during summer 2011

Full size image
Table 1 Acronym, locality and coordinates of sampling sites. See Fig. 1 for localisation
Full size table

Samples of ‘macroalgae’ were randomly collected on available hard substrates; the distance (as the crow flies) between samples was ~25–50 km, while closer samples (~5–10 km) were collected along the rocky shore of French Catalonia and in Provence and the French Riviera in order to assess the local variability. The depth was 0.3 to 0.5 m; depth was referred to the limit between the midlitoral and the infralitoral zones (according to Pérès and Picard 1964). This depth range corresponds to the maximum abundance of Ostreopsis spp. (Cohu et al. 2013). Macroalgae were gently placed in a plastic flask with the surrounding water in order to limit the loss of dinophycean cells. Samples were immediately fixed with 2 % acid Lugol’s solution and kept at 4 °C in the dark. The epibiontic cells were separated from the macroalgae by vigorous shaking. The surrounding water was first filtered on a 250-μm meshed grid, followed by filtration on a 100-μm meshed grid. This protocol was repeated twice by adding 300 mL of filtered sea water (10 μm) each time. The final volume of the surrounding water containing epibiontic cells was 1 L. Prior to being weighed (wet mass, ±0.01 g), the macroalgae were cleaned of detritus and roughly wiped with absorbent paper. Only three Dinophyceae taxa were taken into account in the assessment of cell abundance: Ostreopsis spp., P. lima (Ehrenberg) F. Steinand and Coolia spp. Cells were counted by means of an optical microscope using calibrated squared chambers (1 mL; Sedgwick Rafter®). Results were expressed as number of cells per gram of wet mass of macroalgae (cell g−1 WM).

Macroalgae that constitute the substrate of Dinophyceae belonged to Chlorophyta, Florideophyceae (kingdom Archaeplastida) and Phaeophyceae (kingdom Stramenopiles). Identification was done by one of the co-authors of this work (TT). In each sample, only taxa accounting for at least 20 % of the wet mass were taken into consideration here.

In order to compare macroalgal assemblage composition, as a function of the various factors of dinophycean abundance, regional effect on dinophycean abundance and potential competition/facilitation effect between the three studied taxa, we performed a permutational multivariate analysis of variance (PERMANOVA) (Anderson 2001) based on the S17 Bray Curtis similarity matrixes (Clarke and Warwick 1994; Clarke and Gorley 2006). P values were obtained by 999 permutations of residuals under a reduced model. The possible effects of the macroalgal substrate were analysed through a multivariate exploratory approach using non-metric multi-dimensional scaling (nMDS) (Clarke and Gorley 2006). Similarity measure matrixes were calculated from the initial data matrix containing, for each sample, the abundance of considered Dinophyceae after square root transformation. The chosen similarity measure was the S17 Bray Curtis similarity (Clarke and Warwick 1994; Clarke and Gorley 2006); nMDS represents samples as points in low-dimensional space such that the relative distances apart of all points are in the same rank order as the relative similarities of the samples. Finally, for each taxon, correlations of the main macroalgal substrate with the 2-D ordination plot of samples were plotted by displaying correlation vectors. Spearman correlation was used. Only samples with cell abundance higher than 1000 cell g−1 WM were considered for nMDS analyses.

Results

Apart from the three studied dinophycean taxa (see below), no other dinophycean taxon was ever found in significant amount at any site. It is of interest to note that before the beginning of the blooms of Ostreopsis spp., a short-lasting bloom of the diatom Coscinodiscus sp. (Coscinodiscophyceae) was very often observed.

Occurrence and abundance of Ostreopsis spp.

On the basis of visual examination, the collected material seems to correspond to O. ovata (small size) rather than to O. siamensis. However, in the absence of more accurate taxonomical characterization (electronic microscopy, genetic analysis), which would be beyond the scope of the present study, we hereafter refer the sampled material to Ostreopsis spp.

At 9 out of the 20 sites sampled in French Catalonia and Languedoc, Ostreopsis spp. were never observed during the 3 months of the study period (LC1, LC2, LC3, LC7, LC11, LC13, LC15, LC18 and LC20). Where present, it was very uncommon, with a maximum density of 56 cells g−1 WM (Table 2).

Table 2 Abundance of the studied dinoflagellates (cells g−1 WM macroalgae) along the Languedoc and French Catalonia coast and dominant macroalgae (at least 20 % of the wet mass). For macroalgal abbreviations, see text
Full size table

In Provence and the French Riviera, Ostreopsis spp. were was observed at all the study sites, with the exception of PR2. Its abundance generally conspicuously increased from June to August and was higher in the eastern part (PR31 through PR39), with density up to ~195,000 cells g−1 WM than in the western part, with density generally lower or very much lower than 10,000 cells g−1 WM (Table 3).

Table 3 Abundance of the studied dinoflagellates (cells g−1 WM macroalgae) along the Provence and French Riviera coast and dominant macroalgae (at least 20 % of the wet mass). For macroalgal abbreviations, see text
Full size table

In Corsica, Ostreopsis spp. were recorded at all the study sites. As in Provence and the French Riviera, a distinct seasonal pattern was observed, generally peaking in August. The density was higher along the west coast (CO7 through CO13: up to ~207,000 cells g−1 WM) than in other areas (always less than 5000 cells g−1 WM) (Table 4).

Table 4 Abundance of the studied dinoflagellates (cells g−1 WM macroalgae) along the Corsica coast and dominant macroalgae (at least 20 % of the wet mass). For macroalgal abbreviations, see text
Full size table

Occurrence and abundance of P. lima and Coolia spp.

P. lima and Coolia spp. were relatively abundant in the south-western part of French Catalonia and Languedoc (LC1 through LC6) and even conspicuously more abundant than Ostreopsis spp. In contrast, in the other study areas, they were generally less abundant than Ostreopsis spp. Massive proliferations were recorded at, e.g. PR29 (Coolia spp. ~25,000 in June) and PR14 (P. lima ~45,000 in July). No abundance higher than 100,000 cells g−1 WM was recorded (Tables 2, 3 and 4).

Regional effect on abundances of the three dinophycean taxa

Abundances of the three dinophycean taxa, Ostreopsis spp., P. lima and Coolia spp., were significantly dependent on the region (Table 5a). When focusing on Provence and the French Riviera, where abundances were high, the regional effect was no longer significant (Table 5b).

Table 5 Permutation multivariate analysis of variance (PERMANOVA) on square root transformation of the abundance of the three studied dinophycean taxa (Ostreopsis spp., Prorocentrum lima and Coolia spp.), as a function of the region (regional effect)
Full size table

Correlation between Ostreopsis spp., P. lima and Coolia spp. abundance

The abundance of Ostreopsis spp. was not correlated (positively or negatively) with the presence and abundance of the two other dinophycean taxa, P. lima and Coolia spp. (Table 6). That means there is neither competition between these taxa, for example access to a resource, nor inhibition through, e.g. toxic compounds produced by one of the dinophycean taxa, nor facilitation.

Table 6 Permutation multivariate analysis of variance (PERMANOVA) on square root transformation of abundance of Ostreopsis spp. in each sample, according to the abundance of the two other dinophycean taxa, Prorocentrum lima and Coolia spp. using S17 Bray Curtis similarities. Only Provence and French Riviera were considered, in order to avoid a possible bias due to the regional effect
Full size table

Effect of the macroalgal substrate

The macroalgal species composition of each sample was highly variable and the three dinophycean taxa were associated with 34 macroalgal taxa: 4 Chlorophyta (CR = Caulerpa cylindracea Sonder (=Caulerpa racemosa var. cylindracea); CL = Cladophora laetevirens (Dillwyn) Kützing, UI = Ulva intestinalis Linnaeus and UL = Ulva lactuca Linnaeus), 14 Florideophyceae (Rhodophyta) (AA = Asparagopsis armata Harvey, CE = Ceramium spp., C = Corallina elongata (J. Ellis & Solander) K.R. Hind & G.W. Saunders, GN = Gelidiella nigrescens (Feldmann) Feldmann & Hamel, GC = Gelidium crinale (Hare ex Turner) Gaillon, HV = Haliptylon virgatum (Zanardini) Garbary & H.W. Johansen, HI = Halopitys incurvus Hudson, JR = Jania rubens (Linnaeus) J.V. Lamouroux, L = Laurencia spp., LV = Liagora viscida (Forsskål) C. Agardh, PA = Parviphycus pannosus (Feldmann) G. Furnari, PC = Plocamium cartilagineum (Linnaeus) P.S. Dixon, P = Polysiphonia spp., WS = Womersleyella setacea (Hollenberg) R.E. Norris) and 16 Phaeophyceae (CV = Cladostephus verticillatus (Lightfoot) Lyngbye, CA = Cystoseira amentacea (C. Agardh) Bory de Saint-Vincent, CB = Cystoseira barbata (Stackhouse) C. Agardh, CBB = Cystoseira brachycarpa var. balearica (Sauvageau) Giaccone, CCO = Cystoseira compressa (Esper) Gerloff & Nizamuddin, CC = Cystoseira crinita Duby, DP = Dictyopteris polypodioides (A.P. De Candolle) J.V. Lamouroux, DDD = Dictyota dichotoma (Hudson) J.V. Lamouroux var. dichotoma, DDI = Dictyota dichotoma var. intricata (C. Agardh) Greville, DF = Dictyota fasciola (Roth) J.V. Lamouroux, DI = Dictyota implexa (Desfontaines) J.V. Lamouroux, EC = unidentified Ectocarpales, HF = Halopteris filicina (Grateloup) Kützing, HS = Halopteris scoparia (Linnaeus) Sauvageau, PP = Padina pavonica (Linnaeus) Thivy, TA = Taonia atomaria (Woodward) J. Agardh).

Table 7 shows that the abundance of the three taxa of Dinophyceae significantly differed according to the species composition of the macroalgal substrate. Samples composed of J. rubens and H. scoparia exhibited the highest abundance of Ostreopsis spp. (>100,000 cell g−1 WM) in contrast to samples composed of U. lactuca, C. elongata, Laurencia spp., D. fasciola and T. atomaria which presented lower abundances (1000 to 10,000 cell g−1 WM). Coolia spp. followed the same trend, unlike P. lima that showed the opposite trend (Fig. 2).

Table 7 Permutation multivariate analysis of variance (PERMANOVA) on square root transformation of the abundance of the dinophycean taxa (Ostreopsis spp., Prorocentrum lima and Coolia spp.), as a function of the macroalgal composition of each sample (all regions and dates), using S17 Bray Curtis similarities
Full size table
Fig. 2
figure 2

Two-dimensional nMDS. a Correlation vectors of taxa of macroalgae vs. Dinophyceae abundance; correlation >0.3 (Spearman). All samples considered. b Two-dimensional nMDS ordination plot on abundance of Ostreopsis spp. in each sample. c Two-dimensional nMDS ordination plot on abundance of Prorocentrum lima in each sample. d Two-dimensional nMDS ordination plot on abundance of Coolia spp. in each sample. Each circle represents a sample. Abundance of Dinophyceae: light grey circles = 1000 to 10,000 cell g−1 WM; dark grey circles = 10,000 to 100,000 cell g−1 WM; black circles = >100,000 cell g−1 WM. 2D stress: 0.01

Full size image

When the regional effect was removed, i.e. when the nMDS ordination was focused upon Provence and the French Riviera (excluding therefore the other regions), a larger number of macroalgae were associated with dinophycean abundance. In addition to the former species, U. intestinalis, C. amentacea, D. dichotoma var. intricata and D. dichotoma var. dichotoma emerged (Fig. 3). The two varieties of Dictyota and C. amentacea were associated with high densities of Ostreopsis spp. and Coolia spp. and low densities of P. lima, while the opposite was the case with U. lactuca, C. elongata, D. fasciola and T. atomaria. In addition, the status of three macroalgae was slightly changed (J. rubens, Laurencia spp. and H. scoparia).

Fig. 3
figure 3

Two-dimensional nMDS. Correlation vectors of taxa of macroalgae vs. Dinophyceae abundance (see Fig. 2b–d); correlation >0.3 (Spearman). Only samples from Provence and the French Riviera. 2D stress: 0.01

Full size image

Discussion and conclusion

Our study clearly showed the extensive distribution of Ostreopsis spp. along most of the French Mediterranean coast. In French Catalonia and Languedoc, abundance was conspicuously low, more than two orders of magnitude less than in Provence, French Riviera and Corsica; this cannot be due to the extension of beaches in Languedoc, as density was also low along the neighbouring rocky shores of French Catalonia. In addition, density was relatively high on the eastern shore of Corsica, with a similarly long extension of sandy beaches (Tables 2, 3 and 4). The reason is therefore a geographical parameter, such as temperature, salinity or the invasional process. The temperature, which is considered to enhance Ostreopsis blooms (Vila et al. 2001a, 2001b; Granéli et al. 2011; Mangialajo et al. 2011; David et al. 2012; Cohu et al. 2013), is lower in the Gulf of Lions than in Provence, French Riviera and Corsica (Nykjaer 2009; Schaeffer et al. 2011; Shaltout and Omstedt 2014). Low salinity possibly hinders Ostreopsis spp. (Pistocchi et al. 2011; Pezzolesi et al. 2014; Grossel et al. 2013); it is slightly lower in the Gulf of Lions than in other study regions due to the Rhone River plume (Diaz et al. 2008; Brasseur et al. 1996). Finally, the progression of Ostreopsis spp. could be a time-dependent process (the Ostreopsis invasion has not yet reached this area), rather than a habitat-linked one. Such contrasting abundances, at small scale (i.e. regional scale), were not highlighted by previous authors, who did not study such an extensive area at regional scale.

In addition to regional variability, there is also local spatial variability. Abundance was highly contrasted over short distances (a few kilometres), as already noted by previous authors (Aligizaki and Nikolaidis 2006; Cohu et al. 2011, 2013; Mangialajo et al. 2011).

The ecology of Ostreopsis spp. blooms, in relation with environmental parameters, has been investigated by several authors (Vila et al. 2001a, 2001b; Penna et al. 2005; Mangialajo et al. 2008; Totti et al. 2010; Accoroni et al. 2011, 2012; Cohu et al. 2011). However, the bloom determinism seems to be complex and is still far from being fully understood. In Provence, the French Riviera and Corsica, the abundance of Ostreopsis spp. increased from June to August. Previous authors have already emphasized this trend (Mangialajo et al. 2008, 2011; Cohu et al. 2011, 2013). These blooms are of the same magnitude as those previously recorded in summer in Provence, French Riviera and the Ligurian Sea (Mangialajo et al. 2008, 2011; Cohu et al. 2011, 2013). Our results confirm that high densities and blooms are linked with high sea surface temperature (SST); their frequency in the French Riviera may be related to the low frequency of wind events, which induce rapid cooling episodes of the sea surface water (Schaeffer et al. 2011).

In the Mediterranean Sea, Ostreopsis spp. have been recorded in the water column (Taylor 1979) and on a variety of substrates, such as seagrasses, macroalgae, dead biotic material, reefs and sand (Faust et al. 1996; Turki 2005; Aligizaki and Nikolaidis 2006; Mangialajo et al. 2008; Battocchi et al. 2010; Totti et al. 2010; Cohu et al. 2013; Mabrouk et al. 2014). As far as macroalgae are concerned, previous authors did not distinguish the species of macroalgae (Mangialajo et al. 2008, 2011; Cohu et al. 2011) or only considered a very few species (Cohu et al. 2013), with the exception of Battocchi et al. 2010. Cohu et al. (2013) noted the ability of Dictyota spp. to support very high abundance of O. cf. ovata during blooms, but were unsuccessful in establishing significant preferences. Here, 34 taxa of dominant macroalgae hosting the species were identified. They belong to three higher taxa (Chlorophyta, Florideophyceae and Phaeophyceae) and to a variety of morphological types (articulate, fleshy, bushy, flat and erect). We have recorded blooms mainly on two varieties of D. dichotoma, C. amentacea and H. scoparia bearing, as epibionts, a large amount of the articulated coralline J. rubens. In contrast, abundance was significantly lower on U. lactuca, D. fasciola, T. atomaria and C. elongata. It is worth noting that our results explain the failure of previous authors (Cohu et al. 2013) to establish a correlation between dinophycean abundance and the presence of Dictyota spp.: In fact, this taxon encompasses species and varieties that present opposite trends (Fig. 3).

What could be the reasons for the correlations (positive or negative) between the abundance of dinophycean cells and the dominant macroalgae within the sample? Several features may account for these differences. (i) The morphology of macroalgae can provide a wide span of available surface area for epibiontic cells, depending upon, e.g. the high density of branches (as in H. scoparia and C. amentacea) vs. flat and sparsely branched blades (as in Ulva spp., Dictyota spp. and T. atomaria). The fact that D. dichotoma var. dichotoma and D. dichotoma var. intricata, on one hand, and D. fasciola and T. atomaria, on the other hand, exhibited opposite trends (Fig. 3) suggests that morphology is not the sole parameter involved in the observed correlations. (ii) Many macroalgae produce a variety of biologically active compounds (e.g. Fenical 1983; Baker 1984; Paul and Fenical 1986; Hay 2009; Michalak and Chojnacka in press). Because of their properties (antimicrobial, antiviral, antifungal and anti-herbivore toxins), these compounds can hinder the settlement of epibiontic organisms, both multicellular and unicellular. T. atomaria, which was associated with low densities of Ostreopsis spp. and Coolia spp., exhibits high antimicrobial and antiviral activities (Caccamese et al. 1980; but see Salvador et al. 2007). D. fasciola contains several sesquiterpenes (Amico et al. 1979), in addition to antiviral activities (Caccamese et al. 1980); sesquiterpenes are known to exert antiherbivory properties (Boudouresque et al. 1996; Estell et al. 2008). Solid extracts of C. elongata, when lyophilized, exert in summer antimicrobial and antifungal activities (Salvador et al. 2007). As far as macroalgae associated with high abundances of Dinophyceae are concerned, D. dichotoma var. intricata and H. scoparia lack in summer antimicrobial and antifungal activities (Salvador et al. 2007). This is consistent with the behaviour of the herbivorous sea urchin Paracentrotus lividus (Lamarck, 1816), which strongly to moderately prefers these species when browsing, behaviour usually indicative of the scarcity of defence compounds (Boudouresque and Verlaque 2013). However, the production of active compounds cannot account for all of our results (Fig. 3). C. amentacea and D. dichotoma var. dichotoma exhibit antimicrobial activities (Salvador et al. 2007; Alghazeer et al. 2013), while being highly associated with Dinophyceae; in addition, the former species is strongly preferred and the latter moderately to strongly preferred by sea urchins (Boudouresque and Verlaque 2013). (iii) Ostreopsis spp. and associated dinophyceans can be co-consumed by herbivores, together with the macroalgal substrate (Bire et al. 2013). If small multicellular metazoans (invertebrates) and/or unicellular eukaryotic grazers actively feed on these dinophyceans, they would control their abundance in the regions or on the macroalgae where they dwell. Unfortunately, nothing is known about the direct consumption of the studied dinophyceans.

Finally, Ostreopsis spp. are frequently associated with two other Dinophyceae, P. lima and Coolia spp., as already highlighted by the authors (Penna et al. 2005; Turki 2005; Zenetos et al. 2005; Aligizaki and Nikolaidis 2006; Streftaris and Zenetos 2006; Zingone et al. 2006; Aligizaki et al. 2009; Ingarao et al. 2009; Ignatiades and Gotsis-Skretas 2010). Our results confirmed this frequent association of the three dinophycean taxa; they did not show any obvious competition or facilitation between them.

Our study clearly questions the dispersal of Ostreopsis spp. Obviously, the dispersal seems to be effective but its mechanisms remains unclear; further studies are need to disentangle the role of currents, drifting debris (Masó et al. 2003) or aerosols (Ciminiello et al. 2014) in the dispersal of the species and to determine at which stage of development (cell, dormant cysts or non-dormant cysts) it occurs. The understanding of these mechanisms would contribute to explaining the scarcity of the species in French Catalonia and Languedoc.