The first lionfish ever reported in the Mediterranean was caught by a trawler off the coast of Israel in 1991 and identified as P. miles (Golani and Sonin 1992). From that moment, no more lionfish were reported until 2012, when two specimens were captured in Lebanon (Bariche et al. 2013). Soon after, lionfish were reported in Turkey, Cyprus, Greece and Italy (Turan et al. 2014; Crocetta et al. 2015; Iglésias and Frotté 2015; Oray et al. 2015; Turan and Öztürk 2015; Azzurro et al. 2017). Lionfish were first considered invasive in the Mediterranean in 2016, when they were reported in large groups and numbers in Cyprus (Kletou et al. 2016). Lionfish have now established and successfully spread through a large part of the eastern Mediterranean (Gökoğlu et al. 2017; Turan et al. 2017; Dimitriadis et al. 2020; Ulman et al. 2020; Vavasis et al. 2020) and continue to expand their range westwards (Azzurro et al. 2017; Phillips and Kotrschal 2021). Today, invasive lionfish populations are confined to the eastern part of the Mediterranean (Dimitriadis et al. 2020; Phillips and Kotrschal 2021), with only sporadic sightings elsewhere. The northernmost report of lionfish is that of a single individual found near the island of Vis, in Croatia (Dragičević et al. 2021) while the westernmost lionfish was also a single individual sighted in the Alboran Sea, Spain (Fortič et al. 2023). Since no established populations are present at these locations, the individuals in Croatia and Spain may be the result of isolated aquarium releases. The northernmost part of the Aegean Sea has also remained free from lionfish, probably due to the colder waters (Dimitriadis et al. 2020; Phillips and Kotrschal 2021).

Genetic studies revealed that lionfish found in the Mediterranean originate from the Red Sea and that they most likely entered their new range during multiple invasion events through the Suez Canal (Bariche et al. 2017). The origin of Mediterranean lionfish is corroborated by the absence of established populations of P. volitans in the Basin; while both P. miles and P. volitans are often sold together in the aquarium trade (Kimball et al. 2004) and, consequently, are found in the invaded western Atlantic, genetic studies showed that only P. miles is present in the Red Sea (Hamner et al. 2007; Kulbicki et al. 2012; Wilcox et al. 2018). Thus, the lionfish population of the Mediterranean is considered the result of P. miles entering through the Suez Canal and the reports of P. volitans in this sea (e.g. Gürlek et al. (2016); Gökoğlu et al. (2017); Ayas et al. (2018)) are most likely the result of misidentifications or descriptions of individuals that came from isolated aquarium releases.

The northern Red Sea is inhabited by another lionfish species that is biologically and ecologically similar to P. miles; Pterois radiata. P. miles and P. radiata often occur together on the coral reefs of the northern Red Sea and in comparable abundances (Gavriel and Belmaker 2021). Interestingly, P. radiata has never established in the Mediterranean (Kulbicki et al. 2012; Gavriel and Belmaker 2021). It was hypothesised that P. radiata may be less invasive than P. miles due to its smaller size and slightly higher degree of habitat and diet specialisation (Kulbicki et al. 2012; Gavriel and Belmaker 2021). Comparative studies analysing behavioural, physiological and reproductive traits in a controlled environment may help elucidate what aspects are preventing P. radiata from becoming invasive in the Mediterranean Sea.

P. miles entered the Mediterranean from one of its easternmost locations and continue to expand westwards and northwards (Bariche et al. 2017; Phillips and Kotrschal 2021), calling for continuous updates to pinpoint the location of their current invasion front. Citizen science, defined as the involvement of lay people in data collection, is an effective tool to track the expansion of invasive species (López-Gómez et al. 2014; Larson et al. 2020; Hermoso et al. 2021). This is especially true for species such as P. miles; they are appreciated by divers for their attractive morphology and colouration, increasing the chances of lay people spotting and recognising them. P. miles are also difficult to misidentify, especially in the Mediterranean, where closely-related species (i.e. native scorpionfishes) have a markedly different appearance. Finally, the awareness amongst lay people and stakeholders on the invasiveness of lionfish is high (Kleitou et al. 2021), making them attentive and willing to collaborate with scientists. Citizen science is, therefore, particularly suited to monitor the invasive range of lionfish in the Mediterranean, a sea where the diving industry is well established and dive centres are numerous (Phillips and Kotrschal 2021).

As a follow-up to Phillips and Kotrschal (2021), we contacted dive centres on the Mediterranean coast to ask whether they see lionfish during their dives and if they remember the first year that they saw them. We used a list of dive centres on the Mediterranean coast compiled in 2021 (Phillips and Kotrschal 2021). From this list, we contacted all the dive centres that were still open and reachable via email in April 2023. In most countries, we sent emails in two languages: the first language spoken in the country and English. Translations into local languages were provided by native speakers. We sent emails in two languages to make our survey accessible to those who do not speak English fluently and to foreigners running dive centres in countries of which they do not speak the local language. Dive centres in Egypt, Albania, Montenegro, Malta and Israel were contacted only in English. We sent a reminder to every dive centre that did not respond within a week and we recorded responses for four weeks after the reminder. We used the GPS coordinates of the location of dive centres as an estimation of the point where lionfish are seen as most dives are done in the waters close to a dive centre. Any response that we received in a language different from English were translated through Google translate. When a dive centre reported a range of years as an answer to the date of the first sighting (e.g. 2020–2021), we considered the most recent year in the range as year of first sighting. Data were analysed in R 3.6.2 (R Development Core Team 2019). Maps were produced with the package ‘leaflet’ (version 2.0.4.1, Cheng et al. (2021)).

Contacting 996 dive centres yielded 326 responses (Fig. 1A). Sightings were reported by 82 dive centres, mostly in the eastern Mediterranean (Fig. 1B). Lionfish were seen by almost every dive centre that responded from Israel, Cyprus, Turkey, Greece and Albania. The lionfish reported at the furthest locations from the Suez Canal were reported in Croatia (42.6513°N, 18.0608°E), Malta (35.9500°N, 14.4063°E) and the Italian islands of Sicily (36.7330°N, 15.1205°E) and Sardinia (40.5699°N, 8.2430°E). When compared with the results by Phillips and Kotrschal (2021) (Fig. 1C), our data show that, in just two years, lionfish have expanded their invasive range in the Mediterranean at two fronts: the northern Aegean Sea and the southern Adriatic Sea. While most of the dive centres reported no lionfish in 2021 in the northern part of the Aegean, they almost all did in 2023; the only two dive centres reporting no lionfish in the northern Aegean were also the ones with the northernmost coordinates. A limited expansion can be seen also in the southern Adriatic, where two dive centres reported lionfish sightings in 2023, while none did in 2021.

Figure 1.

Maps of respondents and lionfish sightings. Panel A shows the respondents to our survey in 2023. Each dot represents a dive centre that we contacted, with orange dots representing dive centres that responded and black dots representing dive centres that did not. Panel B shows the responses to our survey in 2023. Each dot represents a dive centre that responded to our survey in 2023 with orange dots representing dive centres that reported lionfish sightings and black dots representing dive centres that reported no sightings. Panel C shows the responses to the survey in 2021 (Phillips and Kotrschal 2021). Each dot represents a dive centre that responded to the survey in 2021 with orange dots representing dive centres that reported lionfish sightings and black dots representing dive centres that reported no sightings.

The years and locations where lionfish were first seen (Fig. 2) corroborate an expansion of the lionfish invasive range in the Mediterranean. Lionfish were first seen in the northern Aegean, Ionian Sea and southern Adriatic between 2020 and 2022. Individuals in Sicily, Sardinia, Croatia and Malta were also seen only in the most recent year range. This suggests that lionfish found at these locations are probably not just the results of aquarium releases; if that was the case, we could have also expected reports in the past. More likely, these individuals have been transported by strong currents from the eastern Mediterranean, either as larvae or eggs. It is important to note that none of the dive centres reporting lionfish in Malta, Italy and Croatia provided pictures and, therefore, misidentification is still a possibility for these sightings. The dive centres reporting sightings at these locations confirmed that there are no established lionfish populations there.

Figure 2.

Map of years of first sighting. Each dot represents a dive centre that reported lionfish sightings, either in 2021 or 2023 and included in their response the year when lionfish were first sighted. The darkness of dots shows the year range when lionfish were first sighted.

Our results show that the invasive range of P. miles continues to expand rapidly in the Mediterranean. Similar to most coral reef fishes, lionfish eggs hatch into pelagic larvae (Ahrenholz and Morris 2010; Vásquez-Yeomans et al. 2011). Larvae are the life stage with the highest dispersal potential in coral reef fishes (Shanks 2009) and are arguably the main contributor to the dispersal of lionfish, which are highly site-attached as adults (McCallister et al. 2018; Gavriel et al. 2021; Phillips et al. 2024). The Mediterranean invasion is following a similar course to that of other Lessepsian species, which typically expand in the Mediterranean starting from the Levantine Sea and gradually spread westwards and northwards towards the Aegean and Ionian Sea (Azzurro et al. 2013).

Many Lessepsian species remain confined to the eastern Mediterranean and are rarely found in high numbers elsewhere (Azzurro et al. 2013; Galil et al. 2017). A modelling study (Johnston and Purkis 2014) predicted that lionfish were unlikely to become invasive in the Mediterranean. However, our and others’ (Azzurro et al. 2017; Phillips and Kotrschal 2021) empirical evidence suggests that the lionfish population of the Mediterranean is well-established and keeps expanding westwards. One of the reasons for the unpredicted success of lionfish could be that their invasion is developing under a strong effect of climate change. When climate change is accounted for (Loya-Cancino et al. 2023), lionfish are predicted to find suitable conditions in the Mediterranean. Another reason could be that Johnston and Purkis (2014) considered that lionfish only spread as larvae under the action of natural currents and, because the Mediterranean is less connected by internal currents than the Atlantic, they concluded that lionfish are unlikely to spread across the whole Mediterranean. Although larvae are probably the phase at which lionfish move over long distances, adults can enter new areas as happened recently in Brazil, where lionfish managed to cross the Amazon-Orinoco plume, most likely by adults moving along deeper mesophotic reefs (Soares et al. 2022, 2023). Additionally, the Mediterranean is highly trafficked and this could allow lionfish larvae or eggs to cross large stretches of sea independently from currents if they enter the ballast water. It will be challenging to disentangle the effects of the factors contributing to the unpredicted success of lionfish in the Mediterranean, as multiple phenomena are at play without possibilities to manipulate them.

Remarkably, several P. miles sightings were reported in areas that were considered to have winter surface temperatures that are too cold for this species (< 15 °C) such as the northern Aegean and southern Adriatic (Kimball et al. 2004; Johnston and Purkis 2014; Dimitriadis et al. 2020). Although it is too early to conclude that P. miles will establish at these locations, climate change has been predicted to facilitate the expansion of tropical invasive species’ ranges in the Mediterranean and other ecosystems (D’Amen and Azzurro 2020). Climate change is resulting in a gradual process of ‘tropicalisation’ of the Mediterranean; the biological community composition is shifting in favour of Lessepsian species at the cost of native ones (Giorgi 2006; Galil et al. 2017). This has already resulted in Lessepsian fishes outweighing and outnumbering native ones in marine protected areas of the eastern Mediterranean (Giakoumi et al. 2019). Another major difference between the Mediterranean and the Red Sea is the stronger seasonality of the former. It remains unknown how the seasonality of Mediterranean waters affects the dynamics and distribution of lionfish and other invasive species of tropical origin.

Our study shows that citizen science is a fruitful approach to monitor lionfish populations at the large scale in the Mediterranean, where the dive industry is strong and awareness towards lionfish is high. Different approaches are needed to monitor the state of the invasion on the southern coasts of the Mediterranean, where data are lacking and the number of dive centres is extremely low (Fig. 1A). When we contacted members of a Libyan spearfishing association through social media, they reported seeing lionfish relatively frequently on the (eastern) Libyan coast. Moreover, lionfish were reported at several locations on the southern coast of the Mediterranean in the past, including Tunisia (Dailianis et al. 2016; Al Mabruk and Rizgalla 2019). This suggests that, as expected, the lionfish invasion and its expansion are not limited to the northern coast of the Mediterranean.

Lionfish have 18 venomous spines; one on each of the first 13 rays of their dorsal fin, one on each of their pelvic fins and three on their anal fin (Aktaş and Mirasoğlu 2017). They show high site fidelity and often return to the same hiding place over the course of several weeks (McCallister et al. 2018; Gavriel et al. 2021; Phillips et al. 2024), although this can vary significantly at the individual level (Tamburello and Côté 2015; Gavriel et al. 2021). Lionfish are often found, either individually or in small groups, hiding in caves and crevices during the day and swim in the open only at dawn and dusk to hunt for prey (Cure et al. 2012; McCallister et al. 2018; D’Agostino et al. 2020; Gavriel et al. 2021). The eastern Mediterranean offers a markedly different habitat from the coral reefs of the Indo-Pacific and the tropical Atlantic (Kulbicki et al. 2012; Côté and Smith 2018). The Mediterranean is a sub-tropical environment dominated by rocky reefs, seagrass meadows and sandy patches (Bussotti and Guidetti 2011; La Mesa et al. 2011; Kleitou et al. 2021). Despite these habitat differences, P. miles have established well in the Mediterranean and have already reached higher population densities than in their native range (Kulbicki et al. 2012; Phillips et al. 2024). It is perhaps not surprising that P. miles are thriving in the eastern Mediterranean as, in the western Atlantic, lionfish have been reported in habitats that are novel for this species, including mangrove forests, river estuaries and seagrass beds (Barbour et al. 2010; Jud et al. 2011; Claydon et al. 2012). Analyses of the population structure and dissections of females indicate that the Mediterranean population of P. miles is reproducing and will remain a stable presence (Savva et al. 2020; Mouchlianitis et al. 2022).

Fishes make up most of the lionfish diet (Barbour et al. 2010; Harms-Tuohy et al. 2016; Zannaki et al. 2019), although they have been reported to also feed on invertebrates (Valdez-Moreno et al. 2012). Lionfish are stalking, gape-limited predators: they slowly follow their prey, sometimes for several minutes, with flared pectoral fins before striking and swallowing them whole (Green et al. 2011; Green and Côté 2014). They tend to prefer small, shallow-bodied benthic and demersal fishes in the Caribbean (Green and Côté 2014; Ritger et al. 2020) and show a similar prey preference in the Mediterranean, where they also adopt the same hunting strategy (Zannaki et al. 2019; D’Agostino et al. 2020). In their native range and the invaded Atlantic, lionfish are a widespread component of the community of coral reef predators (Lesser and Slattery 2011; Cure et al. 2012; Kulbicki et al. 2012; Côté and Smith 2018). In the Atlantic, P. volitans can have strike success rates as high as 85%, the highest reported in animals in the wild (Vermeij 1982; Green et al. 2011).

The high predation effectiveness in their invaded range has been attributed, at least in part, to prey naïveté (Côté and Smith 2018) (but see Cure et al. (2012)). The ‘naïve prey hypothesis’ (or ‘prey naïveté hypothesis’) posits that prey that are exposed to an exotic predator are not prepared to recognise or effectively react to it due to a lack of co-evolutionary history (Sih et al. 2010). Numerous studies support the relevance of prey naïveté in the lionfish invasion in the Atlantic. For instance, several prey species do not react to lionfish with the same readiness as they do with native predators (Anton et al. 2016; McCormick and Allan 2016; Haines and Côté 2019, but see Marsh-Hunkin et al. 2013). In the eastern Mediterranean, exotic prey species from the Red Sea, which co-evolved with P. miles, are abundant and occur together with Mediterranean prey. Exotic prey show a markedly higher flight initiation distance when a lionfish is approaching them compared to Mediterranean species, supporting the hypothesis that prey naïveté is also relevant in the Mediterranean invasion (D’Agostino et al. 2020).

Experiments on prey naïveté in the context of lionfish invasions raise the question of whether the selection pressure posed by this new predator will result in adaptations in local prey. It follows from the definition of prey naïveté that it can be counteracted by evolutionary adaptation: after several generations of co-existence with a novel predator, prey should evolve innate responses (Anton et al. 2020). However, how rapidly can such evolutionary adaptations evolve in prey? This is an unresolved question: some estimates based on data on multiple taxa suggest that hundreds of generations are needed (Anton et al. 2020), while there is evidence showing that 10–30 generations can be enough for predators to drive evolutionary changes in prey (O’Steen et al. 2002; Nunes et al. 2014; Melotto et al. 2020). The great variability in the number of generations needed for local prey to evolve an innate response to predators is probably explained by factors such as the pressure posed by predators and the genetic variability of prey populations (Nunes et al. 2014). The potential for prey to adapt to a new predator such as lionfish is of high scientific relevance, but also has practical implications because it will determine the long-term effects on the local prey communities of the Mediterranean and western Atlantic. In an experiment in the western Atlantic (Kindinger 2015), the antipredator response of damselfish (Stegastes planifrons) to P. volitans was measured and compared to that displayed against a control, native predator. Damselfish were generally naïve to P. volitans, including individuals from populations that had co-existed with lionfish for three and seven years (Kindinger 2015). Local adaptation by prey to P. miles has never been tested in the Mediterranean.

Prey naïveté interferes with innate predator recognition in animals (Sih et al. 2010; Anton et al. 2020). However, this is not the only mechanism resulting in prey reacting to a predator. Individual fishes can learn which species can pose a threat to their survival through associative learning (Kelley and Magurran 2003). Predator recognition can be learned either directly, when a fish escapes an attack from a predator or indirectly when an individual observes predation events or associates the presence of a predator with the presence of danger-related cues (e.g. blood, stress pheromones) from other fishes (Brown 2003). Learned predator recognition is pervasive in fishes (Brown 2003; Kelley and Magurran 2003; Mitchell et al. 2011); prey fishes can learn to associate danger cues with the presence of a predator during a single conditioning event and retain a behavioural response to that predator for extended periods of time (Chivers and Smith 1994, 1995; Mitchell et al. 2011). Could native prey fishes compensate for their lack of innate responses to lionfish through learned predator recognition? This is an open question for both the Mediterranean and the Atlantic invasion. Specific work on how well prey species learn that lionfish pose a threat to their survival is limited to one study on a species from the native range of lionfish. This study suggests that even prey that co-evolved with lionfish seem to have difficulties associating them with danger, while other predatory fishes can be learned more readily (McCormick and Allan 2016). This has led to the hypothesis that lionfish circumvent learned predator recognition mechanisms in prey (Côté and Smith 2018). Whether Mediterranean or western Atlantic prey can learn to recognise lionfish as predators is currently unknown and more research is needed to test for the relevance of circumvention of learned predator recognition in lionfish prey.

Predator recognition allows prey to mount an appropriate behavioural response to a predator. Therefore, invasive lionfish are predicted to select on traits that make prey better able to recognise them, either innately or through learning. However, predators can also select on prey behavioural traits that make them less likely to be preyed on due to processes that are not related to predator recognition (Blake and Gabor 2014; Belgrad and Griffen 2016). For example, boldness affects the susceptibility of mud crabs (Panopeus herbstii) to be preyed on, creating the potential for predators to select for boldness in this species. Interestingly, boldness has a different effect on susceptibility to predation depending on the predator: toadfish (Opsanus tau) consumed more frequently shy mud crabs, while blue crabs (Callinectes sapidus) consumed more frequently bold mud crabs (Belgrad and Griffen 2016). This is due to major differences in the hunting strategies of the two predators. In the context of lionfish invasions, individuals of small, benthic fishes that are bolder or simply more active at dusk or dawn can be predicted to be at a higher risk of predation. This is because lionfish hunt at twilight and show a preference for benthic and demersal prey (Cure et al. 2012; Green and Côté 2014; McCallister et al. 2018; D’Agostino et al. 2020; Ritger et al. 2020). We can, therefore, expect lionfish to select for prey individuals whose activity peaks do not coincide with peaks in lionfish hunting and that are, overall, less active or hide more in their hiding spots. This could lead to changes in the behaviour of populations of native fishes in the invasive ranges of lionfish. Whether such selection pressure is at play in the context of lionfish invasions and any consequences on prey populations has never been investigated.

The ‘enemy release hypothesis’ posits that exotic organisms benefit from reduced top-down control due to a paucity of natural enemies in their newly-invaded ranges (Colautti et al. 2004). The success of lionfish as invaders has been attributed to a lack of natural predators in the areas that they invade (Côté and Smith 2018). However, the natural enemies and source of mortality of lionfish in their native range remain unknown. It seems unlikely that any predator feeds consistently on the venomous and spinous adult lionfish and events of predation remain sporadic and anecdotal, both in their native and invaded ranges (Côté and Smith 2018). The cornetfish Fistularia commersonii and the groupers Epinephelus striatus and Mycteroperca tigris have been reported to feed on lionfish (Bernadsky and Goulet 1991; Maljković et al. 2008). There is also indication that large groupers may act as biological control agents in the Caribbean (Mumby et al. 2011), although large-scale studies suggest that lionfish density does not correlate with that of groupers (Hackerott et al. 2013). In the Mediterranean, the only convincing example of predation is that of an octopus (Octopus vulgaris) filmed while catching and carrying a lionfish in Cyprus (Crocetta et al. 2021). The scarce knowledge on lionfish predators limits any conclusions on the importance of relaxed predation as an explanation for the high invasiveness of lionfish.

There are other factors than reduced predation on adults that could explain the large population sizes that lionfish reach in their invaded ranges. First, parasites, rather than predators, could be limiting the fitness of adults in their native range (Tuttle et al. 2017). This is supported by data from studies that found relatively low numbers of parasites on invasive lionfish in the Atlantic compared to conspecifics in the native Indo-Pacific (Loerch et al. 2015; Sellers et al. 2015; Tuttle et al. 2017). Such comparisons have not yet included Mediterranean lionfish. Second, a main source of mortality for coral reef fishes is predation at or soon after settlement (Carr and Hixon 1995; Webster 2002; Almany and Webster 2006). Predation on larvae and recruits could, therefore, be the main source of mortality for lionfish (Phillips and Kotrschal 2021). Lionfish larvae are pelagic and probably less defended than the adults (Kitchens et al. 2017) and could be prey of plankton feeders before settlement on the reef and small demersal predators at settlement (Phillips and Kotrschal 2021). Relaxed predation on larvae and recruits could explain the lionfish population increase in their invaded ranges if there were a lower abundance of plankton feeders and predators than in their native range. However, there are no studies comparing the mortality of lionfish recruits between their invasive and native ranges and hypotheses involving lionfish at these stages are difficult to test since spawning in P. miles and P. volitans has never been described (Côté and Smith 2018).

The high effectiveness of lionfish as predators implies that they are a potential threat to the native fish community of the areas that they are invading. P. volitans are, indeed, having a profound impact on the fish community of the western Atlantic, where they prey heavily on numerous species of very high conservation value (Rocha et al. 2015; Ingeman 2016; Côté and Smith 2018). Invasive P. volitans can dramatically reduce the biomass of local species in the Atlantic (Albins and Hixon 2008; Green et al. 2012, 2014), with hypothesised effects on the stability of coral reef ecosystems (Lesser and Slattery 2011). The impact of P. miles on the Mediterranean biodiversity has received little consideration. Preliminary assessments suggest that lionfish are reducing the abundance of certain native species and are, therefore, altering the community composition of the Mediterranean (Turan and Doğdu 2022). However, experiments directly linking P. miles density with the densities of Mediterranean species are currently lacking.

The authors have declared that no competing interests exist.

No ethical statement was reported.

This study was supported by a WIAS (Wageningen Institute of Animal Sciences) Graduate Programme fellowship awarded to D.B.

D.B. and A.K. conceived the study. D.B. collected the data. D.B. analysed the data and wrote the first version of the manuscript. All authors (D.B., B.A.J.P., R.N., P.A.J., M.N. and A.K.) provided feedback on earlier versions of the manuscript and contributed to its final version.

Davide Bottacini https://orcid.org/0009-0003-2902-1067

Bart J. A. Pollux https://orcid.org/0000-0001-7242-2630

Reindert Nijland https://orcid.org/0000-0003-0049-3768

Patrick A. Jansen https://orcid.org/0000-0002-4660-0314

Marc Naguib https://orcid.org/0000-0003-0494-4888

Alexander Kotrschal https://orcid.org/0000-0003-3473-1402

All of the data that support the findings of this study are available in the main text.