Hydrozoan species richness in the Mediterranean Sea: past and present

Introduction

Species lists and distribution records are fundamental to biodiversity research (Costello et al. 2001; Mora et al. 2011). Biodiversity research has a long history in the Mediterranean Sea, which is one of the best known seas globally. Yet, reliable fauna inventories and easily accessible regional species lists are only available for a few taxonomic groups. Lists of nominal species, such as the ERMS (European Register of Marine Species; Costello et al. 2001), do not provide resolution of records at the regional scale (Boero 2002; Occhipinti-Ambrogi et al. 2011). Both historical and ecological factors contribute to high biodiversity in the Mediterranean (Bianchi & Morri 2000; Bianchi 2007). In particular, pronounced seasonality is thought to facilitate the spatial coexistence of species by creating niche differentiation in time, and by allowing the occurrence of species with temperate and tropical affinities in the same basin (Coma et al. 2000).

The biodiversity of the Mediterranean Sea is changing dramatically, partly linked to global warming, which is thought to contribute to the establishment of tropical nonindigenous species in the basin (Coll et al. 2010; Lejeusne et al. 2010). The arrival of new species of tropical provenance increases the regional species pool, being possibly counterbalanced by regressions of resident species that are adapted to colder water (Boero et al. 2008). The diversity of the Hydrozoa in the Mediterranean Sea is well known (Bouillon et al. 2004, 2006; Schuchert 2006, 2007, 2008a,b, 2009, 2010), and Hydrozoa can be considered a good proxy for marine biodiversity, being widely represented both in the plankton and in the benthos. Here we limit our treatment to non-siphonophoran Hydrozoa (NSH); they are not a formally recognized taxon, but the Siphonophora have, traditionally, been studied separately. NSH, both as polyps and/or medusa stages, are common in all oceans and seas, and syntheses of their global distribution are available for the medusa stage (Kramp 1959, 1961, 1968).

The aim of this paper is to review the knowledge about the diversity of Mediterranean NSH. We assess current estimates of the size of the species pool, and examine whether species might have become locally or regionally extinct.

Material and Methods

Our list of NSH species of the Mediterranean Sea is based on the monograph by Bouillon et al. (2004), on other taxonomic revisions (e.g. Schuchert 2001, 2006, 2007, 2008a,b, 2009, 2010), and on a revison of the Mediterranean NSH that is currently in preparation (Gravili C. & Boero F., unpublished data). We identified nonindigenous species (NIS), examining records from the 19th century to 2010, to trace the origin, date, method of introduction, current distribution and establishment status, and global distribution of NIS. With few exceptions, we named taxa according to Bouillon et al. (2006).

NIS have been classified into five categories according to their establishment status, following Zenetos et al. (2010): casual (found just a few times), established (widely recorded in the basin), invasive (established NIS that are able to rapidly or largely disperse away from the area of initial introduction, having a noticeable impact on the recipient community), and questionable (of doubtful taxonomic status). We decided to discard the category cryptogenic (species whose probable introduction occurred in ‘early times’ and has not been witnessed, e.g. prior to the XIX century) because no species was ascribed to it in this analysis.

The date and location of the first observation of each NIS in the Mediterranean Sea were extracted from the literature. Whenever possible, the actual date of first collection has been reported, along with the publication date of the paper first recording a NIS, since the two dates only rarely coincide.

The vectors of NIS introduction are the most probable mechanism by which a NIS reached the Mediterranean Sea; however, they are often just presumed and are factually demonstrated in only a very few cases. Mediterranean records of Indo-Pacific species, for instance, are often labelled as the result of Lessepsian immigration through the Suez Canal (Por 1978). In this article, however, Lessepsian immigrants (LI) are only the species with established populations in the Red Sea that have expanded their distribution to the Mediterranean Sea by using the new connection, whereas the species that are recorded from the Indo-Pacific but not from the Red Sea, are labelled as Possible Lessepsian immigrants (PLI). The difference implies that LI crossed the Suez Canal using natural means (i.e. own mobility, drifting, rafting, etc.), whereas PLI entered the Mediterranean via human vectors. The route of entrance of PLI, in fact, might have been the Suez Canal, but an Indo-Pacific species transported by a ship, for instance, would reach the Mediterranean even if the ship were to circumnavigate Africa, and hence the passage through the Canal would be only circumstantial. Our knowledge of the hydrozoan fauna of the Red Sea is unfortunately not as detailed as that of the Mediterranean Sea (see Table 1, for a list of Red Sea NSH). The Indo-Pacific species recorded from the Mediterranean Sea, but not yet from the Red Sea, are therefore just considered as PLI. The type locality, broad geographic distribution (comprising all historical records) and native range are given for each NIS.

To identify species that have not been positively recorded for some time, records of all Mediterranean NSH were entered into a relational database whose entries can be arranged to list the collection localities of each taxon in chronological order. Taxonomic records (i.e. records of each taxon, in any kind of report) are reported on a time scale from the original description to the last citation in the literature. Mediterranean species were ranked according to the date of their last record in the basin: species not recorded for 41 years or more, species not recorded for 31–40 years, species not recorded for 21–30 years, species not recorded for 11–20 years, and species recorded in the last 10 years.

Results

How many species?

Picard's (1958) list of Mediterranean Antho- and Leptomedusae includes 191 species, Boero & Bouillon (1993) report 346 NSH species, and Bouillon et al. (2004) list 396 NSH taxa of 457 hydrozoan species (i.e. including the Siphonophora); this represents about 11% of the 3702 nominal known species of the superclass Hydrozoa globally (Bouillon et al. 2006). New records of Mediterranean hydrozoans are continually improving our knowledge of the fauna (Schuchert 2001, 2006, 2007, 2008a,b, 2009, 2010; Galea 2007; Gravili et al. 2007, 2008; Morri et al. 2009), resulting in the most current estimate of 400 NSH species (Fig. 1). Most of these (68%) occur in the Atlantic Ocean and are subdivided into Mediterranean Atlantic (14%), tropical Atlantic (10%), boreal (12%), and circumtropical (21%), cosmopolitan (11%); 20% are endemic to the Mediterranean, 8% are of Indo-Pacific origin, and 4% are non-classifiable.

How many nonindigenous species?

Recent additions to regional faunal lists (e.g. Diphasia digitalis, Dynamena quadridentata, Sertularia thecocarpa, Campanularia morgansi) are almost invariably represented by nonindigenous species (NIS), whereas description of new species is rarer. Sixty-nine nonindigenous species (NIS) of NSH have been recorded in the Mediterranean (Fig. 2; Table 2; Appendix 1): 30 are casual, 20 established, four invasive, and 15 questionable. No cryptogenic species are recognized. Nonindigenous taxa include Hydroidomedusae (Anthomedusae: 24 species; Leptomedusae: 34 species; Laingiomedusae: two species; Limnomedusae: three species), and Automedusae (Trachymedusae: six species) (Table 2).

Table 2. List of non-siphonophoran Hydrozoa NIS

Subclass	NIS in the Mediterranean Sea, not restricted to Levantine or Alboran Seas	NIS present only in the Levantine waters	NIS present in the Strait of Gibraltar, in the Alborán Sea or near areas (absent in the rest of the Mediterranean Sea)
Anthomedusae	Garveia franciscana (I), Trichydra pudica (Q), Calycopsis simplex (C), Eudendrium carneum (E), Eudendrium merulum (E), Amphinema rubrum (C), Octotiara russelli (Q), Moerisia inkermanica (C), Corymorpha annulata (Q), Coryne eximia (E)	Bougainvillia aurantiaca (Q), Bougainvillia niobe (Q), Nubiella mitra (Q), Cytaeis vulgaris (Q), Paracytaeis octona (C), Halitiara inflexa (C), Moerisia carine (E), Sphaerocoryne bedoti (C), Euphysa flammea (Q), Corymorpha bigelowi (Q), Ectopleura minerva (Q), Plotocnide borealis (Q)	Russellia mirabilis (C), Tubularia ceratogyne (Q)
Leptomedusae	Gymnangium montagui (E), Cirrholovenia tetranema (E), Eutima mira (E), Filellum serratum (E), Eucheilota paradoxica (E), Campalecium medusiferum (E), Monotheca pulchella (E), Diphasia rosacea (C), Clytia hummelincki (I), Clytia linearis (I), Clytia mccrady (E)	Aequorea conica (C), Macrorhynchia philippina (I), Laodicea fijiana (Q), Eucheilota ventricularis (E), Diphasia digitalis (C), Dynamena quadridentata (E), Sertularia thecocarpa (E), Sertularia marginata (E), Campanularia morgansi (C), Obelia fimbriata (C)	Aglaophenia parvula (C), Cladocarpus multiseptatus (C), Cladocarpus pectiniferus (C), Cladocarpus sinuosus (C), Halecium sibogae (C), Pseudoplumaria marocana (C), Kirchenpaueria bonnevieae (C), Cryptolaria pectinata (C), Diphasia attenuata (C), Diphasia delagei (C), Diphasia margareta (C), Hydrallmania falcata (C), Sertularella robusta (Q)
Laingiomedusae		Fabienna oligonema (C), Kantiella enigmatica (C)
Limnomedusae	Gonionemus vertens (E), Scolionema suvaense (Q)	Olindias singularis (E)
Trachymedusae	Haliscera bigelowi (E), Haliscera racovitzae (C), Amphogona pusilla (C), Arctapodema australis (C)	Halitrephes maasi (C), Tetrorchis erythrogaster (E)

• Establishment success (C = Casual; E = Established; I = Invasive; Q = Questionable).

NIS were grouped into three main categories according to their distribution: recorded throughout the basin (27), recorded in the Levantine basin (27), recorded in the Strait of Gibraltar and/or Alboran Sea (15) (Table 2, Fig. 2). At the sub-regional scale, more NIS are found in the eastern basin (42 species), with a peak of 38 in the Levant Sea, followed by the Gulf of Lions and the Ligurian Sea (21 species), the Ionian and Adriatic Seas (19 species), the Strait of Gibraltar and nearby areas (22 species), the Tyrrhenian Sea (nine species), Algeria and North Tunisia (three species). Some NIS have been recorded in more than one of these areas.

Fifteen NIS listed in Table 2 are considered as ‘questionable’, mainly due to uncertainties about their taxonomic status and distributional patterns. Twenty NIS are established, with self-maintaining and self-perpetuating populations unsupported by and independent of humans in at least one sector of the Mediterranean Sea (See Appendix 1), four (6%) are invasive species having overcome biotic and abiotic barriers to develop large populations in the Mediterranean waters, and for 30 species (43%) casual records have been reported from one or a few locations only. Modes of introduction are unknown for many NIS (36%). Nine per cent have been introduced by ships. Other probable ways of introduction are the expansion of natural ranges through the Strait of Gibraltar (22%) and the Suez Canal (19% of Mediterranean NIS as Possible Lessepsian immigrants, and about 14% as Lessepsian immigrants).

As an example of the expansion of a NSH-NIS in the Mediterranean we chose Clytia linearis (Thornely, 1900), a successful Lessepsian immigrant that, in the last decades, has been recorded to be abundant at almost every Mediterranean location surveyed, occurring from shallow waters to shaded rocky bottoms (Fig. 3). This species probably colonised the Mediterranean from the Suez Canal (the first Mediterranean record is by Billard in 1926), and then, from Gibraltar, it expanded its distribution to the Atlantic coast of Spain (Altuna Prados 1995; Boero et al. 2005).

How many species could we possibly have lost?

Fewer than 160 species of NSH have been positively recorded in the last 10 years. Conversely, there are no positive records of 67 species in the last 40 years or longer. Four decades of absence might be long enough to suggest the possibility that regional extinction (or at least range contraction) may be considered. This period precedes the first signs of the impact of global warming on Mediterranean biota. However, species might have remained unrecorded simply because their descriptions were not sufficient to allow subsequent positive identification, or due to scant expertise on the taxon, with little sampling efforts. As an example, Tricyclusa singularis Schulze 1876, is case of a species that has been ‘absent’ for a very long time. It is a species of boreal affinity, after its original description from Trieste (the northernmost part of the Northern Adriatic and the coldest part of the Mediterranean); it has never been recorded again from the Mediterranean Sea (Fig. 4). Instead, it was recorded several times from Roscoff (Atlantic coast of France) (e.g. Bedot 1911; Teissier 1965), resulting in a disjunct distribution. The disappearance of T. singularis represents not only the putative local extinction of a species but also the disappearance from the Mediterranean of the family Tricyclusidae that comprises just this single species and genus (Boero & Bonsdorff 2007).

Discussion

Changes in the size and composition of regional species pools can provide important insights about environmental change. In general, species lists for an area are updated by adding new species as soon as they enter the taxonomic record (i.e. records of taxa from any source). These updates disregard the species that were in the previous lists and that have not been recorded anymore from the area for a reasonably long period, for instance more than 40 years, before the appreciation of the impacts of global change, as proposed here. An alternative approach, including putative loss of species, may indicate endangered or even extinct species, but these are usually charismatic and popular species, whereas inconspicuous taxa (the bulk of biodiversity) are usually disregarded. The Mediterranean Sea today represents a model basin for all oceans, being subjected to a period of temperature increase that is radically changing its biota (Bianchi 2007; Boero & Bonsdorff 2007; Lejeusne et al. 2010). Our analysis of records over time can provide ‘early warning signals’ of species that may face higher probabilities of local or regional extinction.

The increased number of tropical species recently recorded from the Mediterranean Sea has been called tropicalisation and also occurs in other taxa (Bianchi 2007). Twenty-eight of the 45 species with tropical affinity have been cited first from the Levant basin and most of them are of Indo-Pacific biogeographic affinity. The rest (17) are tropical Atlantic species. All the Indo-Pacific species have been usually labelled as Lessepsian immigrants and, thus, as having entered through the Suez Canal by the expansion of their natural range. This assumption, however, is correct only if the species in question have been recorded from the vicinities of the Suez Canal or, at least, from the Red Sea. Only 11 of 34 NIS species of Indo-Pacific origin have been recorded from the Red Sea and only these, thus, can be considered Lessepsian immigrants with some confidence. Thirty-eight species recorded from the Red Sea are typically Mediterranean (Table 1). They might represent misidentifications, but might also be anti-Lessepsian migrants or cryptogenic hydrozoan NIS for the Mediterranean Sea.

Twenty-four of 69 NSH-NIS are of Atlantic affinity; eight are of these are restricted to the Gibraltar region or, at least, to the Alboran Sea and 16 expanded further into the Western Mediterranean Basin (see Appendix 1). These species might not even be NIS, having been present, albeit unnoticed, at these locations for a long time. These species might be removed from the list but, as literature reports are very scant for this part of the Mediterranean Sea, their status remains a matter of pure speculation.

The number of NIS (69) matches closely the number of species that have not been recorded for more than 40 years (67). If the latter are removed from the species pool, introduction of new species would not result in a net increase. This is also suggested by data about a hydroid fauna that has been studied over the long term by Puce et al. (2009): the number of species tends to remain stable but the composition of the species assemblage changes. In spite of wide speculation about the impact of global warming on marine biodiversity, the work by Puce et al. (2009) is cited as the only report on the effects of global warming on populations of marine invertebrates in a recent review on this topic (Burrows et al. 2011). The cumulative species list thus shows an increase in biodiversity but this may not correspond to the number of species actually present at any given time. Our data suggest that the regional species pools tend to remain stable, and that the tropical NIS colonising the Mediterranean Sea are probably filling the gaps of species that are becoming rarer than before (thus not being recorded) or are even locally extinct. It is debatable whether lack of records is due to scant sampling effort (which, however, has led to recording of the NIS), or to impact of either abiotic (i.e. rising temperatures) or biotic (i.e. competition or predation) factors, or of a blend of the three possible causes. The processes of species extinction operate at different paces and involve different mechanisms at different spatial scales (Carlton et al. 1999). Red lists of endangered species almost invariably deal with charismatic species (a negligible minority) and disregard the bulk of marine biodiversity, i.e. inconspicuous species. We propose here the lack of records of a species for several decades as an operational tool to recognize putative extinctions. Obviously, the fact that a species is unrecorded from a given area does not mean that it is extinct in that area. Decades of absence of records, however, might allow one to raise hypotheses of putative extinctions that need to be tested with focussed samplings, increasing the bearing of biodiversity estimates in terms of species pools.

The Hydrozoa are well studied along the northern coast of the Western Mediterranean Sea and the Adriatic Sea, but records are scant for the rest of the basin. Hence, the present figures might not accurately reflect biodiversity for the entire Mediterranean. It is probable that more hydrozoan invaders will arrive from tropical regions, as the Mediterranean basin is going through a period of tropicalisation (Bianchi 2007). Moreover, the species with the ability to become dormant under adverse conditions will be favoured (Boero 2002), as well as those with the ability to reverse their life cycle, surviving long journeys in ballast water (Miglietta & Lessios 2009); such ontogeny reversal is more widespread across cnidarians than previously thought (Piraino et al. 2004).

The origins of Mediterranean hydrozoan NIS have not been investigated with molecular tools so far. Molecular approaches will, hopefully, improve our understanding of the origin of NIS and their modes of dispersal (e.g. Miglietta & Lessios 2009). The number of NIS in the Mediterranean Sea is probably underestimated. Especially for the Hydrozoa, the key areas for species introductions (i.e. the eastern basin and the Gibraltar Strait) are still under-studied. It is probable that, at least for shallow waters, the additions to the Mediterranean hydrozoan fauna will mostly consist of NIS, as demonstrated for fish assemblages of the Levantine shallow marine ecosystems (Goren & Galil 2005). The results shown here suggest that species lists are dynamic, requiring continual updating that considers both additions of introduced species and putative subtractions of species with historical but not contemporary records. Without these subtractions, the species lists will never show possible biodiversity crises at the level of species pools, as biodiversity is always on the rise due to the arrival of NIS.