Does the river blenny Salaria fluviatilis (Asso, 1801) (Actinopterygii: Perciformes) still survive on the Mediterranean island of Cyprus?

The river blenny (Salaria fluviatilis) is rarely mentioned in the natural history literature of Cyprus. This has led to doubts about its existence there. Mistakes with mismatched fish collections have occurred in the past, so it is worth asking whether the recorded specimens are actually from Cyprus and whether the species survives there today. Archived correspondence from the original specimen collector confirms that the record of S. fluviatilis on Cyprus is genuine, yet our extensive surveys there during the last 4 years have failed to locate the fish. A review of the species’ Mediterranean island distribution shows that it exists on 10 islands that are either close to the mainland or are relatively very large in areal extent. Although the mystery of S. fluviatilis survival on Cyprus lingers, efforts must be made to survey all potential habitats and consider planning a re-introduction project if the population is confirmed as extinct.


Introduction
The river blenny (Salaria fluviatilis Asso, 1801) is very rarely mentioned in the natural history literature of Cyprus. This has led to doubts about the species' presence on the island. Mistakes with misplaced and mismatched fish collections have occurred in the past, so it is worth confirming that the species really exists in Cyprus (Zogaris et al. 2012a).
The uncertain status of S. fluviatilis on Cyprus is an obstacle for conservation. Even during a policy-relevant review of taxa of importance for conservation when Cyprus entered the European Union in 2004, the existence of the species was not presented in official species lists and therefore no special conservation measures were granted to this taxon or its habitat (Hadjisterikotis pers. comm.). Some researchers do not trust that the species was ever found on the island and the species is not given in inland-water fish lists or literature reviews (Charalambidou and Gücel 2009). Also during a recent phylogenetic survey of the species across the Mediterranean, the Cyprus population was not referred to (Almada et al. 2009); although other insular populations, and especially populations on the edge of the species' distribution, which are of increased phylogenetic interest, have been actively investigated (Perdices et al. 2000;Kottelat 2004). Cyprus lies near the eastern edge of the species' potential distribution. Here we review the status of the species' occurrence on Cyprus and, with respect to other Mediterranean islands, we consider local conservation requirements.
Salaria fluviatilis has a near circum-Mediterranean distribution, but is known from just 10 Mediterranean islands. In the Eastern Mediterranean it is found on the islands of Ikaria, Lesvos, Kerkyra, Crete, Rhodes and Euboea (Bianco et al. 1996 ;Koutsikos et al. 2012) and it is also recorded from Cyprus (Bath 2003). In the western Mediterranean it is found on Sicily, Sardinia and Corsica but is absent from many islands where perennial streams are not well developed, such as in Dalmatia or the smaller western Mediterranean islands (Kottelat and Freyhof 2007;Froese and Pauly 2012). Salaria fluviatilis is mainly a riverine species but can also be found in brackish river mouths, lakes and reservoirs (Changeux and Pont 1995;Neat et al. 2003); it will seasonally tolerate brackish conditions but has fully adapted to survive and reproduce in freshwaters (Plaut 1998). It prefers cobble and gravel substrate usually in streams with moderate to high current velocity; it is often found in the deeper parts of streams (Freeman et al. 1990) and also quite high and far from the river mouths, sometimes even above 500 m elevation (Alp and Kara 2007). The males make a nest under large stones and the larvae are pelagic, so areas with still waters in the streams are also required (Vinyoles and de Sostoa 2007). This small fish is sometimes difficult to catch or to observe; consequently it is often difficult to know its status, and new distributional records have come to light (e.g. El-Absy and Mir 1986;Crivelli 2006;Alp and Kara 2007). Salaria fluviatilis is of considerable biogeographical interest because it is the only freshwater representative the Blenniidae to reproduce in freshwaters in the Mediterranean (Cote et al. 1999) and it is often one of the few fish occurring yearround in very small Mediterranean island and peninsular streams. This species is of high conservation interest because its marginal populations in the Eastern Mediterranean have been shown to have significant genetic divergence from the main lineage within this species (Almada et al. 2009). Some S. fluviatilis populations may represent cryptic species, as was recently found in Lake Trichonis, Greece (Kottelat 2004). Populations of S. fluviatilis are generally highly localized (Cote et al. 1999) and there is evidence of a decline in response to certain anthropogenic pressures (Vinyoles and de Sostoa 2007). Mediterranean islands may hold evolutionarily significant populations of this species and their lotic habitats are particularly susceptible to long droughts, anthropogenic river diversions engineered barriers to dispersal and pollution, so these island populations need careful monitoring (Crivelli 2006).

Material and methods
We reviewed the existence of the species in Cyprus in the literature and in the collection of The Natural History Museum, London, UK. We also reviewed records of an extensive fish survey recently carried out on Cyprus (partially published in Zogaris et al. 2012a). The survey explored streams and wetlands using two sets of battery-powered backpack electrofishing equipment (SR-L24 and a Bulgarian custom-made unit), a fry net, and interviews during the spring-summer periods of 2009, 2010, 2011 and 2012. Within a total field work period of 64 field days, 170 sites were sampled (155 were electrofished) in 31 independent river basins primarily on the western and southern parts of the island. Seventy-eight sites (46%) had no fish ( Figure 1) and native fishes were remarkably scarce (Zogaris et al. 2012a). Using a simple two-page questionnaire, interviews referring to the presence of fishes (especially the eel Anguilla anguilla) in particular reaches were undertaken with 149 local inhabitants and experts on Cyprus. This questionnaire was administered primarily to natural resource experts, villagers and naturalists, including knowledgeable government officials and wildlife resource users primarily in the southern part of the island; these data will soon be published in a review, but are partially available in the grey literature (Zogaris et al. 2012b). Finally, data on the presence of S. fluviatilis on other Mediterranean islands were collected, where knowledge exists of their complete freshwater ichthyofaunas (as reviewed in Crivelli 2006 andKottelat andFreyhof 2007). A principal components analysis (PCA) ordination was employed to test whether the presence of S. fluviatilis on Mediterranean islands can be interpreted by key abiotic descriptors, such as an island's area, the shortest distance from the mainland, the highest peak of the island, and the mean annual precipitation. Given that the species is fairly widespread in freshwater streams along the Mediterranean coastline for validation of the abiotic descriptors, two more factors were employed, the longitude and the latitude. All data were log-transformed and analysis used the PRIMER 6β Statistical Package.

Sampling and interviews
The sampling and questionnaire survey failed to find S. fluviatilis on Cyprus. Most of the researched sites were in the western and southern parts of the island so there may be potential for the species to exist in isolated areas of freshwater in the east and north of Cyprus, however perennial lotic reaches are extremely scarce in that part of Cyprus. Of course there is also a possibility that the species is still present even in some inland waters that were checked using the methods employed in the present research. This may be due to: (1) the weakness of the particular sampling technique, (2) seasonal variability of species presence and habitat conditions, and/or (3) incomplete survey coverage (i.e. the entire volume/surface of the habitat could not be checked so the scarce species escaped detection). Of the 31 river basins explored; a representative sample of areas was surveyed presumably hosting adequate habitat conditions for fish survival (particularly in spring). Unfortunately, several sites with deeper water, or high-conductivity waters were not adequately surveyed and it is very difficult to report on the percentage of available appropriate habitat that was sampled. The last 3 years were climatically favourable with rather strong flows from winter to late spring throughout most of the surveyed basins.
The interviews also gave no concrete evidence concerning the species or its habitats; not a single individual who lived on the island was aware of the species, including scientists in relevant fisheries and wildlife conservation positions.

The museum specimens
To our knowledge the only preserved specimens of S. fluviatilis from Cyprus were collected by Roland L.N. Michell a British national, over a hundred years ago. R. L.N. Michell was not a naturalist or ichthyologist but we know he lived on Cyprus for a period and has written on Cypriot ethnography (Michell 1908). Michell's three existing samples are labelled Blennius varus (junior synonym of Salaria fluviatilis) and arrived at the museum in 1909. A letter produced by the collector of the three specimens of S. fluviatilis deposited in the Natural History Museum, London, shows that they were in fact collected in streams in the District of Limmasol, Cyprus (BMNH 1909.2.25.2-4). In his letter to Dr G.A. Boulenger, (28 January 1909) Michell states: "It [S. fluviatilis] is found in 2 or 3 of the torrents of this District [Limmasol]. Nearly all of these run dry (or very nearly dry) during the hottest months of summer. I have been trying for 4-5 years to obtain specimens without success, and have only recently succeeded in getting a few". The archived letter of Michell shows that the species was found in the southern part of Cyprus and that locals were probably aware of its presence, because locals finally brought the fish to him. In fact, our questionnaire showed that many of the older respondents frequently collected food items from streams (especially eel and freshwater crabs), so it should have been possible to know whether other fish species were present. Since this archived letter was found, there can be little doubt that these specimens are from Cyprus (see Appendix for a transcript of the full letter).
The three museum samples were re-checked and all morphologically match S. fluviatilis, and the diagnostic characteristics are evident. The best preserved specimen, an adult male (Figure 2) shows all the major distinguishing features as given in modern identification keys (Kottelat and Freyhof 2007). These specimens are preserved in 70% alcohol solution, they are partially shrivelled and meristic measurements need to be taken with care. However, DNA extraction may be possible for future phylogenetic investigations.
The Mediterranean island distribution of S. fluviatilis It may be instructive to explore key environmental attributes of other Mediterranean islands that host S. fluviatilis. After data were gathered from an assortment of 12 islands that do not have S. fluviatilis and 10 that do, a PCA analysis was performed for four axes for the 22 islands (Table 1, Figure 3). The cumulative variation of the first two axes was 84.4% (Table 1), so it can be inferred that a plot based on these axes represents a sound ordination of the samples (islands). According to the results of the PCA, geographical location given by longitude and latitude factors is not important for the interpretation of the variation in the first two axes. The measure of areal extent that an island occupies is the most important descriptor for the first axis and the shortest distance from the mainland is the most important descriptor for the second axis (Table 1). Most of the islands inhabited by the S. fluviatilis are fairly close to the mainland and most are fairly large islands; only the largest islands away from the mainland host S. fluviatilis.
Why is S. fluviatilis absent from Cyprus? Cyprus, being the third largest Mediterranean island, has adequate stream habitats with perennial flow and many irrigation reservoirs where S. fluviatilis could survive.  In fact, approximately 341 km of streams are considered as flowing perennially according to preliminary water body maps by the Water Development Department (unpublished data, G. Dörflinger pers. comm.). Although most of these are highgradient mountain streams unsuitable for S. fluviatilis, we estimate that at least 20% of the perennial or partially perennial streams could host the species. Furthermore many perennial streams are connected to older dams that store water for irrigation, often below 500-m elevation; therefore, the species could have survived in at least a few of the older reservoirs. It is our opinion that extinction caused solely by natural forces (such as stream desiccation due to drought) is unlikely to have caused a total population collapse with respect to this species, on one of the Mediterranean's largest islands. The apparent scarcity of S. fluviatilis on Cyprus is unusual and may point to a recent anthropogenic extinction. Many of the local respondents to interviews point to the widespread use of DDT for an intensive government anti-malarial campaign, which began in 1946 and ended in 1978. Questionnaire respondents frequently mentioned a "remarkable absence of freshwater crabs, semi-aquatic grass snakes, amphibians and fish", which they attributed to the prolonged and widespread DDT poisoning before the late 1970s (Zogaris et al. 2012b). During the DDT poisoning some aquatic species may have become remarkably range-restricted and/or have gone extinct from the island; the spraying coincides with the time of the extinction of the dipper (Cinclus cinclus), an obligate river bird (Flint and Stewart 1983). Other anthropogenic pressures that may have Figure 3. Principal components analysis ordination on the abiotic descriptors of 22 Mediterranean islands, including the 10 islands where Salaria fluviatilis is present (P) with black dots. The axes refer to island areal size and distance to mainland; and the numbers refer to the island numbers as in Table 2. affected or synergistically precipitated the presumed S. fluviatilis decline/extinction are the widespread water abstraction points that have proliferated in Cyprus, especially since the 1960s. Cyprus currently has 108 dams, and many of the streams of the island are artificially intermittent or ephemeral in their flow below the major dams (Zogaris et al. 2012a). Salaria fluviatilis is vulnerable to habitat degradation and connectivity changes (Cote et al. 1999;Benejam et al. 2010) so the increased water exploitation has destroyed much suitable habitat. One may implicate alien species invasion as a problem for the species on Cyprus, as the number of carnivorous fish has increased in reservoirs and associated perennial streams (i.e. catfish, big-mouth bass, etc). However, in Spain, where S. fluviatilis was found in reservoirs almost as often as in rivers, the species showed an ability to coexist with invasive species both in lotic (e.g. Blanco-Garrido et al. 2009) and lentic (e.g. Garcıa-Berthou and Moreno-Amich 2000) environments. This coexistence has been related to the cryptic coloration and benthic habits of the species that would allow a lower predation pressure than that suffered by native cyprinids (Garcıa-Berthou and Moreno-Amich 2000). The species is tolerant of brackish or seawater salinity levels (Plaut 1998), which means that it could survive in isolated river-mouth pools or other pooling water during periods of extreme drought. Further research on the effects of widespread pesticide spraying may be the key to the presumed demise of this and other aquatic species, and this requires meticulous historical research. From a conservation standpoint, the freshwaters of Cyprus are uniquethey have a high degree of isolation from neighbouring Levantine and Anatolian lands (Por 1975;Abell et al. 2008;Plötner et al. 2012). There are very few ways in which Cyprus could be re-colonized by a freshwater fish if a species is lost. The island has been almost completely isolated from the mainland save for a relatively short periodduring the Messinian Salinity Crisis (approx. 5.59-5.33 million years ago) (Baier et al. 2009). The ancestor of S. fluviatilis, a relative of the marine peacock blenny, Salaria pavo, is said to have been a euryhaline species, perhaps allowing incursions into freshwater and subsequent dispersal via the sea (Almada et al. 2009). The biogeographically confined and range-restricted Mediterranean island populations of S. fluviatilis point to the possibility that the species' distribution may have its origins in the different sea levels and salinity conditions following important geological events (such as glacial period sea-level fluctuations or older salinity fluctuations) and not in long pelagic migrations through current seawater conditions. Today, S. fluviatilis does not usually enter the sea and it reproduces solely in freshwater, but the almost circum-Mediterranean distribution of this species shows that it once somehow reached offshore islands; and this explains its existence in several very small river basins on islands and peninsulas around the Mediterranean (Plaut 1998).
Risk of extinction of freshwater species due to long-term anthropogenic pressures is highly increased by natural vulnerabilities and biogeographical pressures that raise extinction rates of freshwater species on islands. Obviously the extinction pressures on seasonally semi-arid Mediterranean islands are heightened within restricted aquatic habitats, especially where small isolated river basins exist, such as on Cyprus. One would expect certain poor-dispersers, such as freshwater fish, to be particularly vulnerable to extinction on Cyprus. Island biogeographic constraints as well as increased climatic aridity have been shown to increase extinction rates of northern temperate reptile species on several Mediterranean islands (Foufopoulos et al. 2011). Cyprus's freshwater fauna is susceptible to a combination of climatic and long-term anthropogenic pressures (Davis et al. 1998;Hadjisterikotis et al. 2000).

Conservation implications
Current knowledge of Cyprus's inland waters suggests that the island's S. fluviatilis population may be extinct. Verification of local conservation status needs to be reached as soon as possible because this population is potentially of outstanding evolutionary interest, within a species that is known to harbour endemic cryptic taxa at the geographic margins of its range (Kottelat 2004;Almada et al. 2009). In the Greek Islands and along the coast of the Hellenic Peninsula it should be noted that S. fluviatilis was found in extremely short reaches of insular and peninsular river basins and some were partially intermittent; so the species can survive in restricted water bodies. It is imperative that a full exploration of all Cyprus's water bodies be executed before the species' local extinction is confirmed. The rarity or loss of S. fluviatilis from Cyprus is testament to the fact that Cyprus has witnessed a widespread degradation of large areas of unique and critical natural inland waters. However, adequate habitat for S. fluviatilis exists today on both upland areas and near reservoirs (and within some reservoirs), so the real problem is re-colonization. If the island's populations are extinct, human-assisted recolonization through a scientific stocking programme should be implemented. A carefully planned re-introduction scheme may help fill a vacant niche, to restore the only freshwater-reproducing fish in Cyprus's streams.