INTRODUCTION
The coral Cladocora caespitosa (Linnaeus, 1767) belongs to the family Faviidae and is endemic to the Mediterranean Sea. It occurs from shallow waters to depths of about 40 m, where the amount of light still allows the process of photosynthesis for the symbiotic zooxanthellae. The hemispherical bush-like colonies may be found isolate, sparse or forming ‘beds’ (numerous colonies living more or less close to each other) or ‘banks’ (colonies connected together in a large formation more than 1 m high and covering several square metres) (Laborel, Reference Laborel1961; Zibrowius, Reference Zibrowius1980; Kühlmann et al., Reference Kühlmann, Chintiroglou, Koutsoubas and Koukouras1991; Schiller, Reference Schiller1993; Morri et al., Reference Morri, Peirano, Bianchi and Sassarini1994; Peirano et al., Reference Peirano, Morri, Mastronuzzi and Bianchi1998; Kružić & Benković, Reference Kružić and Benković2008). Cladocora caespitosa is physiologically and morphologically similar to typical tropical reef-building corals in being zooxanthellate, colonial and capable of forming extensive formations (Zibrowius, Reference Zibrowius1982; Schuhmacher & Zibrowius, Reference Schuhmacher and Zibrowius1985; Peirano et al., Reference Peirano, Morri, Mastronuzzi and Bianchi1998; Kružić & Požar-Domac, Reference Kružić and Požar-Domac2003; Kružić & Benković, Reference Kružić and Benković2008). Cladocora caespitosa is well adapted to the sea temperature seasonality of the Mediterranean Sea, experiencing large changes in temperature, from 8°C during winter and up to 28°C in summer (Schiller, Reference Schiller1993; Rodolfo-Metalpa et al., Reference Rodolfo-Metalpa, Richard, Allemand and Ferrier-Pagès2006; Kružić & Benković, Reference Kružić and Benković2008).
Banks of this coral formed by the fusion of numerous colonies are very rare at present, although true reefs of C. caespitosa were abundant during the Pleistocene epoch (Peirano et al., Reference Peirano, Kružić and Mastronuzzi2009). Large banks or ‘reefs’ of C. caespitosa, both living and fossil have been reported at several sites in the Mediterranean Sea: near the Tunisian coast (Zibrowius, Reference Zibrowius1980), in the Aegean Sea (Laborel, Reference Laborel1961; Laborel, Reference Laborel1987; Koukouras & Kühlmann, Reference Koukouras and Kühlmann1991; Kühlmann, Reference Kühlmann1996), as well as in Italy (Bernasconi et al., Reference Bernasconi, Corselli and Carobene1997). Morri et al. (Reference Morri, Peirano, Bianchi and Sassarini1994) found two living banks of C. caespitosa near Bonassola and Riomaggiore in the Ligurian Sea. In the Adriatic Sea, Abel (Reference Abel1959) described extensive beds of C. caespitosa in the Lim channel near Rovinj. Pax & Müller (Reference Pax and Müller1962) mentioned a bank of C. caespitosa in Veliko Jezero on the island of Mljet, but they never thoroughly explored it.
Although the National Park ‘Mljet’ protects the C. caespitosa bank from direct human impacts, this colonial coral could be endangered by global change-related disturbances such as mortalities linked to positive thermal anomalies (Kersting & Linares, Reference Kersting and Linares2012; Kružić et al., Reference Kružić, Sršen and Benković2012) or the presence of invasive algal species such as Caulerpa racemosa var. cylindracea (Sonder) Verlaque, Huisman and Boudouresque (Kružić & Benković, Reference Kružić and Benković2008). Mass-mortality events of the temperate coral Cladocora caespitosa were recorded in the north-western Mediterranean in the last two decades, mostly by polyp tissue necrosis (Rodolfo-Metalpa, Reference Rodolfo-Metalpa, Bianchi and Peirano2000, Reference Rodolfo-Metalpa, Bianchi, Peirano and Morri2005, Reference Rodolfo-Metalpa, Richard, Allemand, Bianchi, Morri and Ferrier-Pagès2006).
Important members of a coral reef community are coral-eating predators, as they can affect coral community structure and population dynamics. Coralliophilid snails are recognized as an important factor in structuring coral reef communities in tropical regions, although information on their population structure is still limited. In the Pacific region, the muricid gastropod from the genus Drupella and the crown-of-thorns sea star Acanthaster planci are examples of corallivores that undergo population outbreaks with significant impacts on coral populations (Glynn, Reference Glynn and Dubinsky1990; Bruckner et al., Reference Bruckner, Bruckner and Williams1997). Factors influencing the coral as a host may include nutritional quality, host frequency and distribution, predation pressure and competition (Cumming, Reference Cumming1999; Ritson-Williams et al., Reference Ritson-Williams, Shjegstad and Paul2003; Sotka, Reference Sotka2005; Johnston & Miller, Reference Johnston and Miller2007).
Gastropods from the genus Drupella feed exclusively on live coral tissue (Taylor, Reference Taylor1976, Reference Taylor and Morton1980), stripping it from the calcium carbonate skeleton and leaving distinctive, white feeding scars (Cumming & McCorry, Reference Cumming and McCorry1998). Large numbers of the corallivorous snail Drupella cornus (Röding, 1798) appeared on a 150 m artificial limestone quay in Eilat, Israel, following coral mortality (Shafir et al., Reference Shafir, Gur and Rinkevich2008). By July 2004, there were more than 200 Drupella on the colonies (30 cm in diameter) of several branching coral species, mainly on Acropora spp. The Caribbean corallivorous gastropod Coralliophila abbreviata (Lamarck, 1816) feeds on at least 14 species of scleractinian coral (Miller, Reference Miller1981). Several authors have reported small and localized tissue damage from the C. abbreviata predation (Ward, Reference Ward1965; Robertson, Reference Robertson1970; Knowlton et al., Reference Knowlton, Lang, Rooney and Clifford1981, Reference Knowlton, Lang and Keller1990).
Spada (Reference Spada1968) observed gastropod Coralliophila babelis Requien, 1848 (now accepted as Babelomurex cariniferus (Sowerby, 1834)) on Cladocora caespitosa colonies in the area of Lampedusa Island. The Indo-Pacific immigrant Quoyula madreporarum (Sowerby, 1832) (now accepted as Coralliophila monodonta (Blainville, 1832)) was found on Cladocora caespitosa colonies in the area of Sicily and the Adriatic Sea (Ghisotti, Reference Ghisotti1968; Di Natale, Reference Di Natale1978).
The corallivorous gastropod Coralliophila meyendorffii (Calcara, 1845) (family Coralliophilidae) that feeds upon a wide range of coral prey commonly occurs in the north-eastern Atlantic (including the Mediterranean Sea) (Figure 1A, B). Two main size-classes are encountered that show discrete differences in ecology. Large sized forms (adult females 8–40 mm in length) are found associated with sea anemones (Anemonia viridis Forsskål, 1775), while smaller forms (adult females 4–20 mm in length) are found on scleractinians (on colonial species such as Cladocora caespitosa, or solitary such as Balanophyllia europaea (Risso, 1826) or Leptopsammia pruvoti Lacaze-Duthiers, 1897) (personal observations). An extensive listing of Mediterranean prey species has not been published yet.
Fig. 1. (A, B) Coralliophila meyendorffi; (C) part of the Cladocora caespitosa bank at Veliko Jezero. scale bar: 5mm.
The results of an extensive research regarding a recent outbreak of gastropod Coralliophila meyendorffi (Calcara, 1845) in July, August and September 2010, that have caused considerable damage on the Cladocora caespitosa bank in Veliko jezero (the Mljet National Park, southern Adriatic Sea) are presented in this paper.
The objectives of the present study were to compare the feeding characteristics of Coralliophila meyendorffi on its coral host, the factors affecting host use, as well as the distribution of snails across a Cladocora caespitosa bank in the Mljet National Park. Since there were more damaged colonies in sheltered areas (with small influence of bottom sea currents), this study also tries to document the variations in gastropod size and abundance found on coral colonies in relation to the bottom sea currents measured on the coral bank.
MATERIALS AND METHODS
Study area
The special features of Mljet Island are its two marine lakes (Veliko and Malo Jezero). These lakes are in fact two depressions that were flooded by the sea during the postglacial period. In the largest part of the inlet Veliko Jezero (Great Lake) the present-day maximum depth is 47 m, while the smaller, inward part, Malo Jezero (Small Lake), is only 21 m deep. The lakes are connected to the open sea through a narrow channel constructed from stone blocks, about 60 m long, 12 m wide, and 3.5 m deep. Sea currents are quite strong in the channel during the tidal exchange between the marine lakes and the open sea. The Cladocora caespitosa bank in Veliko Jezero (42°46.05′N 17°22.30′E) occurs at depths from 6 to 18 m, covering an area of 650 m2, and thus forming one of the largest banks of C. caespitosa found in the Mediterranean Sea (Kružić & Požar-Domac, Reference Kružić and Požar-Domac2003; Kružić & Benković, Reference Kružić and Benković2008) (Figures 1C & 2A, B).
Fig. 2. (A) Maps of the Mediterranean Sea and Mljet National Park with location of the bank of the coral Cladocora caespitosa (arrow) studied in this investigation; (B) map of the Cladocora caespitosa bank in Veliko Jezero (Mljet National Park) with investigated sites. One square on the map is 25 m2.
Sampling methods
The coral bank of C. caespitosa was investigated by SCUBA diving during the summer season (July and August) from 2006 to 2010. Bottom sea currents were measured using a Flowatch flowmeter during high tide at fifteen stations where colonies were collected during 2010. Seawater temperatures at the investigated stations were recorded yearly at 1 h intervals using Onset Computers temperature data loggers (HOBO Pendant).
The first larger infestion of the C. caespitosa bank was observed during the night dive in 2010. Fifteen infested and damaged colonies, similar in size (colony diameter measuring around 3 cm), were collected and carefully placed in separate plastic bags and then brought to the laboratory in water-filled buckets. The mean number of corallites per colony was estimated, counting the number of corallites in ten replicated 5 × 5 cm squares (Peirano et al., Reference Peirano, Morri, Bianchi and Rodolfo-Metalpa2001; Kružić & Benković, Reference Kružić and Benković2008). Snails were isolated and numbered for each colony and their shell lengths were measured using calipers.
Statistical analysis
Mean value and standard deviation were calculated for morphological data. Prior to statistical analyses, the raw data were tested for violations of normality and homogeneity of variance assumptions. Normality was assessed using the normal probability plot procedure and the Shapiro–Wilk test, while homogeneity of variances was tested using Levene's test. Differences between sampling sites from the bank were investigated using one-way analysis of variance (ANOVA). The Pearson r correlation was used to test correlation between the investigated parameters. Statistical analysis and statistical graphics were performed using Statistica 9.0 for Windows.
RESULTS
During 2010, the mean monthly sea temperature at the bank in Veliko Jezero ranged from 12.93 to 27.35°C (mean value 19.14 ± 3.92°C). The maximum sea temperature values were measured in August and September (29.14°C and 28.35°C, respectively). The bottom sea currents at the Mljet bank, due to tidal exchange, ranged from 0.83 m/s during low tide to 1.03 m/s during high tide. During high tide, bottom sea currents ranged from 0.48 ± 0.18 m/s at Station M10 to 1.23 ± 0.29 m/s at Station M5 (Table 1).
Table 1. Stations and depth of investigated stations, bottom sea currents, affected corallites per colony of Cladocora caespitosa bank, snail density and snail shell length.
The highest mean number of corallites per colony in 25 cm2 was estimated at Station M12 (89.33 ± 7.65), while the lowest value was estimated at Station M1 (61.47 ± 3.09). Stations M1, M2 and M3 showed significant difference (ANOVA, P < 0.001) compared to other stations from the Cladocora caespitosa bank.
Coralliophila meyendorffi feeds on live coral tissue, stripping it from the calcium carbonate skeleton and leaving distinctive, white feeding scars. After predation, only the white coral skeleton remains (Figure 3A, B). The highest and lowest number of affected corallites was 98 (Station M9) and 29 (Station M5) corallites per colony, respectively (Table 1). The highest snail densities in a single colony recorded were 41 snails per colony at Station M15, and 39 snails per colony at Station M13 (Table 1). The lowest snail density sites were Stations M5 (nine snails per colony) and M7 (eight snails per colony). Statistical analyses showed a significant difference between the investigated stations (ANOVA, P < 0.001). The highest values of shell length were measured at Stations M1 (21.27 ± 6.28) and M4 (19.26 ± 3.39), while the lowest values were measured at Stations M12 (7.53 ± 2.03) and M15 (8.44 ± 2.48) (Table 1). Significant differences in shell length were found between these two groups of stations (ANOVA P < 0.001).
Fig. 3. (A, B) bleached polyps and tissue mortality (white skeleton) on Cladocora caespitosa caused by Coralliophila meyendorffi (arrow).
Snail density varied among the coral bank, depending on bottom sea currents. Individuals were highly aggregated in sheltered areas of the coral bank. A negative significant linear correlation (P < 0.05; r = −0.6642) existed between the number of corallites in 25 cm2 and bottom sea currents and also when comparing number of corallites per colony and bottom sea currents (P < 0.05; r = −0.6274) at the investigated bank (Figure 4A, B). Also, a negative significant linear correlation (P < 0.05; r = −0.8101) was revealed between snail density and bottom sea currents (Figure 4C), while a positive significant linear correlation (P < 0.05; r = 0.8933) existed between shell length and bottom sea currents (Figure 4D). A positive significant linear correlation (P < 0.05; r = 0.8154) existed between snail density and affected corallites at investigated stations (Figure 5).
Fig. 4. Relationships between corallite number (A), affected corallites (B), snail density (C), gastropod shell total length (D) and bottom sea currents at investigated stations.
Fig. 5. Relationship between snail density and affected corallites at investigated stations.
DISCUSSION
No ecological research on the family Coralliophilidae has focused on the Adriatic Sea on these current high-density populations. Large aggregations of corallivorous gastropod Coralliophila meyendorffii have been observed on colonies of Cladocora caespitosa and other hard and soft coral species in several locations in the Adriatic Sea (Kornati National Park, Telašćica Nature Park). The gastropod Coralliophila meyendorffi was previously only rarely found within the Mljet bank (only a few specimens were recorded during the cryptofaunal surveys on the Mljet coral bank in the summers of 2000 and 2006). We hypothesize that the gastropod outbreak could be related to higher sea temperatures measured in the past two decades (Cerrano et al., Reference Cerrano, Bavestrello, Bianchi, Cattaneo-Vietti, Bava, Morganti, Morri, Picco, Sara, Schiaparelli, Siccardi and Sponga2000; Garrabou et al., Reference Garrabou, Coma, Bensoussan, Bally, Chevaldonné, Cigliano, Diaz, Harmelin, Gambi, Kersting, Lejeusne, Linares, Marschal, Pérez, Ribes, Romano, Serrano, Teixidó, Torrents, Zabala, Zuberer and Cerrano2009). Thermal anomalies in the Mediterranean Sea could markedly affect invertebrates, such as corals. Sea temperature of 29.14°C measured in August of 2010 at 10 m depth (coral bank area) exceeds the normal maximum summer value for Veliko Jezero, which is 27°C. Thermal stress could be a considerable factor if the resulting tissue necrosis and tissue lysis of Cladocora caespitosa polyps were a trigger for the mass feeding of the gastropod Coralliophila meyendorffi. This prolonged period (in the Adriatic Sea during August and September) of elevated temperature could have directly triggered the development of opportunistic pathogenic organisms (viruses, bacteria, fungi), and also have favoured a corallivorous gastropod explosion. However, it is not clear whether gastropod density is stable over the year or whether there was an increase in the gastropod density during the warm summers. The reported coral mortality event in 2006 was presumably caused by the feeding activity of the gastropods, but also by the high sea temperatures in the area (one bleached colony as an isolated case). Positive correlation between number of snails and number of affected polyps proven in this study indicates strong relationship between feeding activity of gastropods and coral mortality.
The protection polyp nematocysts provide for this coral species is not impregnable, since Cladocora caespitosa polyps are also consumed by the fireworm Hermodice carunculata (Pallas, 1766), and occasionally by the parrotfish Sparisoma cretense (Linnaeus, 1758) (personal observations). The large surface-living polychaete H. carunculata, known as a voracious predator that feeds on soft and hard corals, has caused extensive damage on C. caespitosa colonies in the area of Mljet National Park. However, this damage was done on the colonies outside the Veliko Jezero. On coral bank no specimens of H. carunculata were observed.
In Veliko Jezero individuals of Coralliophila meyendorffi were observed mainly at night, because they usually hide when exposed to light and disappear before divers notice them. During feeding, C. meyendorffi completely removes the coral's tissue. Recently eaten or damaged areas usually become overgrown by red algae. It is important to note that feeding scars left by these coral predators could easily be mistaken for tissue necrosis (mostly caused by global sea warming (Rodolfo-Metalpa et al., Reference Rodolfo-Metalpa, Bianchi and Peirano2000, Reference Rodolfo-Metalpa, Bianchi, Peirano and Morri2005)) when observed during daylight when the predator is not visible. Feeding scars are characterized by areas of denuded coral skeleton. Single polyps and small areas of live tissue can often be found within the borders of the feeding area. The gastropods were seen on the Cladocora caespitosa colonies only in the feeding area. Only one colony that was affected by the Coralliophila meyendorffi feeding activity had bleached polyps (Figure 3). Polyp bleaching could be more related to high sea temperature, than directly caused by the gastropod.
Larger aggregations of gastropod C. meyendorffi have been observed in sheltered areas of the coral bank with weaker bottom sea currents (Figure 4; Table 1). Also, values of snail shell length were significantly larger in areas with stronger bottom sea currents. This proves that the larger snails better withstand the stronger bottom sea currents. At Mljet bank, the strong bottom sea currents have a direct impact on the morphology of coral colonies and calices diameter. In calm waters corallites are more or less elongated (usually with large interstitial space suitable for gastropod concealment during daylight), while in areas with strong sea currents, corallites are closely spaced and more irregularly shaped. The calices found in areas with strong sea currents had diameters of up to 9 mm (usually 4–5 mm). The gastropods were highly aggregated in sheltered areas of the coral bank (up to 41 specimens/dm2 at Station M15) with smaller clusters (up to 12 specimens/dm2) on coral colonies that were influenced by strong bottom currents (up to 1.23 ms−1 at Station M5). For lower values of snail density at the Stations M3, M5 and M7, the direct influence of strong bottom sea currents could be responsible.
Snails are often more abundant on massive corals (Hayes, Reference Hayes1990; Baums et al., Reference Baums, Miller and Szmant2003), proving that Cladocora caespitosa, as a bank, is a ‘superior’ host, especially in the Veliko Jezero. Cladocora caespitosa is a bushy coral (dendroid form), with parallel growing calices with interstitial spaces, ideal as a snail hideout during the day. Snails could also be found in algal mats along the coral colony. Smaller snails are highly cryptic in these environments, and the difficulty to detect them may also vary depending on the morphology of the coral colonies.
The density of corallivorous gastropods residing on different coral hosts may be self-regulating. Hayes (Reference Hayes1990) reported a field experiment in which snail aggregations on Acropora cervicornis (Lamarck, 1816) colonies were augmented with additional individuals, but aggregation size returned to its original size over time. The gastropod Coralliophila abbreviata normally feeds on the coral Montastrea annularis (Ellis & Solander, 1786). It can consume up to 9 cm2 of tissue in 24 h, and has a mean density of 13 ind./m2 of living coral in shallow water. Due to the low population density of C. abbreviata, and the fact that it feeds only rarely, no extensive damage to the reef corals was recorded (Hayes, Reference Hayes1990).
Over the past two decades, Cladocora caespitosa populations throughout the Mediterranean Sea have undergone widespread declines due to bleaching, pollution, algal blooms, and disease (Cerrano et al., Reference Cerrano, Bavestrello, Bianchi, Cattaneo-Vietti, Bava, Morganti, Morri, Picco, Sara, Schiaparelli, Siccardi and Sponga2000; Rodolfo-Metalpa et al., Reference Rodolfo-Metalpa, Bianchi and Peirano2000; Reference Rodolfo-Metalpa, Bianchi, Peirano and Morri2005; Kružić & Požar Domac, Reference Kružić and Požar-Domac2007; Kružić et al., Reference Kružić, Žuljević and Nikolić2008; Garrabou et al., Reference Garrabou, Coma, Bensoussan, Bally, Chevaldonné, Cigliano, Diaz, Harmelin, Gambi, Kersting, Lejeusne, Linares, Marschal, Pérez, Ribes, Romano, Serrano, Teixidó, Torrents, Zabala, Zuberer and Cerrano2009). Assessing and managing threats to their health will, therefore, be essential for C. caespitosa conservation in the region. This is the first observation that a gastropod caused mortality of a coral in the Mediterranean Sea, similar to those in the tropical coral reefs. The total ecological impact of Coralliophila meyendorffi feeding on corals is still unknown. Our observations highlight how the effects of small, cryptic coral predators may be mistaken for disease, and demonstrate the importance of observations at all times of day and night. A full understanding of predator and prey population dynamics, however, should be a prerequisite to any intervention in predator–prey interactions. Results from this study suggest that if sustained, such rates of predation could have an effect on the viability of these coral populations in Mljet National Park, but not as much as coral bleaching caused by high sea temperatures in the area. We hypothesize that the sea currents were the main factor influencing gastropod aggregation on some part of the coral bank. Elevated sea temperature possibly supported increase of C. meyendorffi population, but this phenomenon deserves more attention and research.
ACKNOWLEDGEMENTS
The authors wish to thank Dr Helmut Zibrowius from Centre d'Océanologie de Marseille (France) for his helpful suggestions and support. The authors also thank the staff of Mljet National Park and colleagues from the Marine Biology Laboratory, Faculty of Science, Zagreb and the Institute for Oceanography and Fisheries from Split for fieldwork and laboratory help.
