A Burning Issue: Interactions of the Bearded Fireworm Hermodice carunculata with Artisanal Fisheries—A Case Study from Malta (Central Mediterranean)

Abstract

The bearded fireworm Hermodice carunculata (Polychaeta) has become increasingly problematic in Mediterranean artisanal fisheries, yet remains understudied. This study provides a detailed analysis of interactions between H. carunculata and artisanal fishers in Marsaxlokk, Malta’s largest fishing village. Combining fisher interviews (local ecological knowledge) and field data, the study reveals that fishing practices inadvertently sustain and amplify local fireworm populations by discarding worms and by-catch nearshore, thereby providing abundant food sources. The highest fisher activity correlated with significantly denser fireworm populations and smaller-sized individuals, indicating population growth driven by fisher practices. Fireworm predation significantly impacts fishers economically, causing an estimated direct loss of 52.5% of the expected profit across the five onboard sessions, due to damaged catch, along with additional indirect losses from reduced gear efficiency and increased labor. The worm’s painful sting adds further nuisance and discomfort for fishers who frequently handle infested gear. Despite awareness of fireworm behavior, fishers largely did not recognize their practices as exacerbating the issue, creating a feedback loop perpetuating the problem. Behavioral experiments suggested that modifying fishing practices and gear design might mitigate fireworm impacts. Addressing this socio-ecological challenge requires further targeted research, education, and policy support to break the cycle that benefits fireworm proliferation in the region to the detriment of fishers.

1. Introduction

The amphinomid Hermodice carunculata (Pallas, 1766) is a large thermophilic polychaete that can reach a length of 77 cm [1], distributed in the tropical and subtropical waters of the western Atlantic Ocean, including the Caribbean Sea and the Gulf of Mexico, as well as the Red Sea and the Mediterranean Sea [2,3]. The worm can occur down to depths of 300 m in the tropics and subtropics [4], but in the Mediterranean Sea, it is most abundant in shallow water, with Righi et al. [5] reporting that in Italy, fireworms are mostly found in the top 20 m of the infralittoral.

H. carunculata is generally considered to be a generalist predator and opportunistic omnivorous scavenger, feeding on carrion as well as injured/immobilized organisms [2,5,6,7], although in the Caribbean, the species is a facultative corallivore [8,9,10]. Righi et al. [5] and Simonini et al. [7] provide an extensive list of the worm’s prey species in the Mediterranean, which includes a broad range of invertebrates and fish. Because of its polyphagous feeding habits, H. carunculata is a nuisance species for artisanal fishers, attacking and consuming the fish caught in benthic fishing gear, specifically bottom longlines [2], trammel and gill nets [11,12], and traps [12], rendering the fisher’s catch not suitable for sale [2,11,12,13].

As with other amphinomids, H. carunculata bears tufts of harpoon-shaped notochaetae which are erected if the animal is threatened (hence the common name of ‘bearded fireworm’). These serve as deterrents, as the fragile tufts can readily detach from the worm’s body and come into contact with nearby susceptible organisms including fish and also humans [14]. In humans, the notochaetae deliver a neurotoxic and inflammatory effect, with victims experiencing a stinging and burning sensation which can lead to edema, erythema, and tissue irritation [15]. This further adds to the nuisance caused by this species, as fishers have to deal with the pain when handling certain gear with entangled fireworms or leftover notochaetal tufts detached from the worms, prolonging potential exposure to the stinging. There is debate on whether the irritation is caused by the presence of a toxin in the notochaetae, or is entirely mechanical [16,17], or is a synergistic effect between both mechanisms [18]. Despite the fact that no toxin-secreting glands have been found, the presence of potential toxins (named carunculines) has been confirmed [14].

Its wide ecological tolerances and plasticity in behavior allow this species to thrive in disturbed environments, both natural and anthropogenic [19,20]. Additionally, in the Mediterranean Sea, no effective predators of H. carunculata have been identified [4], rendering the species essentially free from predator control.

Due to rising seawater temperatures, H. carunculata is undergoing thermophilic expansion in the Mediterranean, moving northwards, such that some authors regard it as an emerging ‘native invader’ with potential economic impacts and public health risks on sea-users such as fishers, SCUBA divers, and beachgoers [2,5,21,22,23,24]. This species has already become a significant economic problem that is painful to deal with for artisanal fishers using demersal fishing gear, as are the bottom longlines [2], trammel and gill nets [11,12], and traps [12], yet surprisingly little research has been conducted on its effects on artisanal fisheries, with currently one study by Simonini et al. [25] focusing on an efficient fireworm-attracting device to collect and monitor H. carunculata. However, considering the predicted further increases in sea temperature [26,27], more research is urgently required to understand the fisher-fireworm interactions and the implications of this species’ expanding distribution.

Artisanal fisheries are widespread throughout the Mediterranean and are economically important for local communities. In the last few decades, this industry has seen a decline in the number of fishers, as a result of diminishing fish stocks, conflicts with the recreational sector and with large-scale fisheries, and a lack of interest among youth in taking up a profession in this sector [28,29,30]. Any significant impact of fireworms on artisanal fisheries continues to stress an already strained sector.

Maltese artisanal fisheries are not immune to the aforementioned problems associated with the increasing prevalence of H. carunculata. Here we present a case study from Malta, centered on the fishing village of Marsaxlokk, the largest on the islands. Apart from having the highest concentration of fishers, Marsaxlokk fishers have been complaining that there are large numbers of H. carunculata in shallow water close to the shore, which are impacting their activities. Our approach was to assess the impacts of the fireworm on artisanal fishers based at Marsaxlokk by drawing on their local ecological knowledge (LEK), conducting onboard observations of their actual fishing operations, and carrying out both in-situ and ex-situ behavioral experiments.

2. Materials and Methods

2.1. LEK Survey

A total of 15 artisanal small-scale fishers from the fishing village of Marsaxlokk were interviewed. Fishers were opportunistically selected on the basis of availability and willingness to participate. The questions, which were asked verbally, are summarized in

Table 1. Edited version of the questions asked to fishers, translated from the original Maltese.

Survey Questions for Artisanal Fishers at Marsaxlokk

Are you a part-time or full-time fisher? Which fishing methods do you use? When did you first start observing fireworms in your catch? Has the population of fireworms changed with time? How? What is the extent of damage caused by the fireworm presently? At what level of fireworm-induced damage does your catch become unsaleable? Is there a time of year (season) where the fireworm increases significantly? With which fishing method is the fireworm most caught? In your experience, where are fireworms evidently present and abundant in Marsaxlokk and along the southern coast of Malta? With reference to Question 9, what do you think causes this fireworm abundance? Do you also find fireworms at depths greater than 20 m? Does their size vary with depth/location, and if yes, how? Do fireworms have any preference in what they eat? From your experience, do you think the fireworm can swim small distances toward prey, and if yes, why do you think so? Has damaged catch due to fireworms affected your profit? Is handling fireworms a nuisance due to stinging? Does fireworm abundance in your gear affect your work effort? What do you do with the fireworms when cleaning out the fishing nets? Have you been doing something to mitigate the problem of fireworms, and if yes, what are you doing? What do you think can be done/should be done by the authorities? Why do you think the population of fireworms has increased over the years?

. Responses were written down on the spot by the interviewer, and no personal data were collected. The 15 surveys represent around 30% of the fireworm-affected fishing community in Marsaxlokk. Fishers deploying trammel nets further than 25 nautical miles from the coast were not included, as when approached, they indicated that they have no interactions with the fireworm and thus could not provide direct LEK, but they would only offer hearsay from their peers, which was not the information we sought.

2.2. Field Experiments

2.2.1. Field Sampling

Four sampling sites for fireworms were established: three sites along the seafront of Marsaxlokk village, where fishers carry out most of their activities, and one site at Kalanka Bay, along the Delimara coast, where there are no activities associated with artisanal fishing. Three replicates were set up at each of the four sites (Figure 1).

Within Marsaxlokk Harbour, the three sites (North, West, and South Seafront) represented varying intensities of artisanal fishery activity with Site 2 = Low-Intensity (LI) Fishing, Site 3 = High-Intensity (HI) Fishing, and Site 4 = Moderate-Intensity (MI) Fishing (see

Table 2. Details of the four sites within the two localities studied.

Locality	Site Number	Geographical Coordinates of the Three Replicates	Depth (m)	Bottom Type	Intensity of Fishing- Related Activities *
Kalanka Bay	1	35°49′26.0″ N 14°33′40.2″ E 35°49′27.1″ N 14°33′36.4″ E 35°49′25.4″ N 14°33′36.0″ E	4.2 m to 5.7 m	Rocky coastline and sandy bottom	None
Marsaxlokk Harbour	2	35°50′25.1″ N 14°32′51.3″ E 35°50′26.1″ N 14°32′48.8″ E 35°50′27.3″ N 14°32′45.3″ E	1.5 m to 3.5 m	Artificial shore with sandy bottom with substantial quantities of benthic litter	Low-Intensity (LI)
	3	35°50′24.8″ N 14°32′41.5″ E 35°50′22.5″ N 14°32′39.7″ E 35°50′20.1″ N 14°32′38.5″ E			High-Intensity (HI)
	4	35°50′15.2″ N 14°32′41.4″ E 35°50′12.3″ N 14°32′40.8″ E 35°50′09.1″ N 14°32′46.6″ E			Moderate-intensity (MI)

* The intensity of artisanal fishing activity was quantified based on the frequency of trammel net operations within each site, which in turn resulted in the on-site discarding of fireworms and by-catch. On a weekly basis, the estimated number of fishing trips was: Site 1; none, Site 2; 1–5, Site 3; 20–25, and Site 4; 5–10. These estimates were derived from field observations and personal communications with local fishers.

). The aim of sampling fireworms was to see whether the intensity of artisanal fishery activities affects the population density, weight, and length of the worms. Site 1 (Kalanka Bay) served as a Reference Site as there are no fishing activities at this site. The substratum at all four sites was carbonate sediment, ranging from medium sand in Kalanka Bay, where benthic litter was almost absent, to fine–medium sand in Marsaxlokk Harbour, where seabed litter contributed substantial solid inclusions.

Marsaxlokk Harbour is subject to non-point sources of pollution and to disturbance from boating, fishery activities, the Malta Freeport, and the Delimara Power Station (Figure 1). Kalanka Bay is much less subject to pollution and disturbance, since it is only visited by beachgoers during summer, and by occasional pleasure craft.

To assess the fireworm population at each site, two different sets of weighted benthic traps were deployed: a funnel trap and a mesh trap (Figure 2A,B). The mesh trap simulates the mesh of a trammel net (but with a smaller mesh size), and worms were able to enter and leave the trap at will. The funnel trap had only one opening, and fireworms that entered this were unable to leave. Around three Trachurus mediterraneus (Mediterranean horse mackerel), weighing 80 ± 5 g, were placed in each trap as bait.

On Day 1, a funnel trap was deployed at each site at 21:00 and retrieved at 5:00 (8 h later), before sunrise. During preliminary observations, fireworms had eaten the bait, but most had escaped prior to trap retrieval. Those that remained were highly active, with many of these escaping during the retrieval process. For this reason, retrieval was done before sunrise. On Day 2, the same procedure was repeated using mesh traps. This protocol was carried out 5 times between August and September 2023. On retrieval, all fireworms in the traps were immediately relaxed in a refrigerator at 5 °C for 30 min, and then their length (accuracy ± 0.1 cm) and weight (accuracy ± 1 g) were measured after blotting with absorbent paper to remove excess water.

2.2.2. Onboard Field Observation of Trammel Net Fishing

Five onboard observation trips, accompanying a fisher using trammel nets in the southeast of Malta (Figure 3), were made between September 2023 and March 2024. When the nets were retrieved, H. carunculata individuals were collected by the observer (A.S.) and transported in seawater to the laboratory. The worms were relaxed, measured, and weighed as already described. Catch data was recorded as well as which fish species in the nets were found attacked by fireworms. Details of the five onboard observation trips are given in

Table 3. Details of fishing and environmental parameters for the field observations (S1–S5) on trammel net (TN) fishing.

	S1	S2	S3	S4	S5
	Sept 2023	Oct 2023	Dec 2023	Dec 2023	Mar 2024
	Fishing data
Time at which TN started to be deployed	21:00	21:45	18:00	20:55	22:00
Length of time TN was left at sea (hours)	11	9	13	10	8.5
Number of TN	4	4	10	10	16
Mesh size (mm)	33–45	33–45	28–45	28–45	24–45
	Environmental data
Location	Refer to Figure 3
Depth (m)	13–15	2–4	2–4	3–15	3–15
Bottom type	Sandy	Sandy	Rocky/Sandy	Rocky/Sandy	Rocky/Sandy
Weather conditions	Sunny	Sunny	Cloudy with sparse showers	Sunny with sparse clouds	Cloudy
Current *	Moderate	Moderate	Slow	Very slow	Slow

* Based on fisher’s observations and whether trammel net moved from point of deployment.

.

The fisher’s catch was visually assessed for damage caused by fireworms using the following categories: minor, moderate, significant, and severe (see

Table 4. Damage category and relative description of the damage caused by Hermodice carunculata observed during the onboard field observation of trammel net fishing.

Category	Description of Damage	Supporting Photograph
Minor	Less than 10% of total body mass damaged, with one or two skin perforations, intact internal organs and tissue, and no visible bones.
Moderate	Up to 50% of total body mass damaged, with internal organs and tissue eaten by the worm up to this percentage, and bones moderately visible.
Significant	Up to 80% of total body mass damaged, with internal organs and tissue eaten by the worm up to this percentage, and bones significantly visible.
Severe	More than 80% of total body mass damaged, with only bones visible, and very little to no skin, internal organs and tissue remaining.

Note: Assessment was made visually while fisher was hauling the trammel net out of the water and while fisher was cleaning the nets. All photographs were taken by A.S. during the onboard sessions.

for details).

2.3. Behavioral Experiments

H. carunculata collected during the onboard observations were kept in aerated tanks and starved for 33 days (based on [7,31]). Three ex-situ and one in-situ experiments were carried out to understand the worms’ behavior and how they interact with the trammel net fishing gear. Considering the nocturnal behavior of fireworms, the ex-situ experiments were carried out in minimal light (0.23 cd) just sufficient for visual observations.

2.3.1. Experiment 1: Drop Test

This experiment was designed to see if fireworms are able to swim. It was inspired by some fishers’ belief, as recorded in answers to the LEK survey, that fireworms are able to swim up from the bottom to attack fish in the trammel nets.

Twenty-five fireworms were each dropped three times into a large tank containing water to a depth of 40 cm, and their behavior was observed until 5 s post-landing on the bottom of the tank. Specimens were chosen within the length and weight range: 11–23 cm, 7–31 g, to test different sizes of worms.

2.3.2. Experiment 2: Effect of Trammel Net Mesh Size

This experiment was designed to find the critical mesh size that allows adult fireworms (length: 13–19 cm; weight: 12–24 g) to pass through.

Three cylindrical traps with mesh of different sizes (1.8 cm × 1.8 cm; 1.0 cm × 1.0 cm; 0.7 cm × 0.7 cm) at each open end were constructed (Figure 2C). Five fireworms were placed in each trap, and the traps were placed on the bottom of a large tank. The worms were observed for 30 min, and the number of worms showing the following behaviors was recorded: unresponsive (u), successful in exiting the trap through a mesh end (s), and failed to exit the trap (f). This experiment was repeated three times such that 15 fireworms were tested for each mesh size.

2.3.3. Experiment 3: ‘Body Lifting’

This experiment was designed to test how high fireworms can lift their bodies off the bottom to reach up to prey suspended in the water, simulating fish trapped in a trammel net.

Fifteen starved fireworms (length: 13–19 cm; weight: 12–24 g) were placed individually at the center of a large tank. Each worm was induced to raise its body by suspending a piece of T. mediterraneus (the caudal fin and approximately 3 cm of the peduncle) above the bottom, tied to a line at a height slightly greater than the worm’s total length. The behavior of the worms was observed for 30 min, and the maximum height to which a worm raised its body off the bottom was recorded.

2.3.4. Experiment 4: ‘Climbing of a Vertical Nylon Rope’

This experiment was designed to see if fireworms could still access fish in trammel nets if a vertical nylon rope separated the netting from the weighted footrope which is in contact with the seabed.

Six traps were constructed, consisting of a plastic water bottle with the open neck pointing down and attached to a weight by a 5 mm thick, 15 cm long nylon rope (Figure 2D). These dimensions represent a thickness equivalent to that of the footrope, and a length which prevents the bottle from touching the seabed and fireworms from accessing the bait in the bottle, by raising their body (as observed in the above behavioral experiments) instead of crawling up the rope. The bottleneck had a diameter of 25 mm which would allow most fireworms to pass through it to get to the bait. The traps were deployed along the Marsaxlokk seafront. To ensure that the bottles do not come in contact with the seabed, the ropes were fully suspended in the water without excess slack. The traps were baited with whole T. mediterraneus (total weight: 80 ± 5 g), left overnight for 8 h, and retrieved at dawn. The number of fireworms in each trap was recorded.

2.4. Statistical Analysis

The IBM SPSS Statistics for Windows, Version 29.0 (IBM Corp., Armonk, NY, USA, 2022) statistical package was used for the statistical analysis.

Generalized Linear Models (GLMs) were employed since both datasets (fireworm abundance, and fireworm length and weight data) did not satisfy the assumptions of normality and homogeneity of variance. A Poisson GLM with a log-link function was used to test whether fireworm abundance varied significantly between the traps and across sites, since the dataset was count data, right-skewed, and contained several zero counts. Poisson regression was chosen over alternatives such as negative binomial or zero-inflated models due to the small sample size, the biological nature of the zeros, and hence the greater robustness of the Poisson model under these conditions. To test whether there was a significant difference between the length and weight of trapped fireworms between trap types and across the sites, a Gamma GLM with a log-link function was used.

In both GLMs, Site 1 was set as the reference category to which Sites 2 to 4 were compared.

3. Results

3.1. LEK Survey

All fishers interviewed were based at Marsaxlokk, with more than three-quarters being over 60 years old, with fishing experience spanning their entire lifetime.

When asked about the fireworm and its interaction with their fishing practices, fishers were generally well-informed about its habits, but stated some beliefs which helped inform subsequent behavioral experiments to test them.

All fishers interviewed agreed that fireworms are mostly observed in trammel nets (100% of responses), followed by pots (60% of responses) and demersal set longlines (40% of responses)—all of which are fishing gears deployed on, or close to, the seabed. For this reason, and the fact that trammel nets were the most used fishing gear by fishers in Marsaxlokk (Figure 4), the onboard observations were carried out during trammel net fishing expeditions.

All interviewed fishers reported that H. carunculata has always been present and observed in their fishing gear, with 66.7% saying that the population of fireworms had increased over their years of fishing. Moreover, most fishers reported that fireworm sightings are highest during summer (80.0%), with 13.3% suggesting that there is no change in the number of sightings throughout the seasons. The remainder (6.7%) stated that sightings are higher in winter. Fishers also reported an increase in fireworm catches at increasing depths over the years, suggesting that the species is locally gradually expanding its depth range.

The interviewed fishers attributed the abundance of the fireworms along the southern Maltese coast, mostly to the availability of food, as well as to the presence of Posidonia oceanica beds. Only one fisher directly linked the high coastal fireworm population density to the fishers inadvertently transporting the worms to the port during their fishing activities. When asked about the reason for the increase in population of H. carunculata, fishers pointed out the increasing sea temperature (40.0%) and the lack of biological or anthropogenic control of the worm (26.7%) as the main reasons.

Three-quarters of the fishers interviewed stated that their catch can only be sold when the damage by fireworms is minor or none at all, rendering most of the affected catch unsaleable (Figure 5). Moderate to severe damage results in the caught fish having little remaining edible flesh, and such fish are normally disposed of by throwing back into the sea, usually when fishers clean their nets once docked inside the port.

When questioned on the worms’ diet, 80.0% of fishers confirmed their opportunistic scavenging behavior, stating that the worm feeds upon anything that becomes available to it—even on fishes that are still alive, immobilized in fishing gear.

26.7% of the fishers believe that the fireworm can swim small distances toward fish caught in the fishing gear. This prompted the behavioral experiments on swimming.

33.3% of fishers suggested that the size and color of the fireworm vary with depth and habitat, with one fisher stating that “the shallower the sea [within ports], the smaller the fireworms”.

The main impact of the fireworm on fishers’ work was reported to be the burning pain caused by handling the species due to its chaetae (53.3%), followed by reduced income from damaged catch (46.7%) and increased work effort in clearing the worms from the catch and the gear (40.0%). The remaining 13.3% stated they are not impacted by the fireworm. While all interviewed fishers reported discarding fireworms at sea when cleaning out the fishing nets, 13.3% mentioned occasionally leaving the fireworms to dry out on land, still entangled in the fishing nets. To mitigate the negative impacts of fireworms, some 20.0% of the interviewed fishers have opted to change their fishing practices and fish in areas thought to have fewer fireworms (through hearsay or through personal experience) or to fish in deeper waters. The remaining 80.0% expressed helplessness in addressing the issue. In response to Question 20, two fishers suggested developing specialized nets to catch fireworms as a way to mitigate the problems the species is causing them, while 60.0% expressed that nothing can be done.

3.2. Field Experiments

3.2.1. Field Sampling

A total of 89 fireworms were caught using the funnel traps and 153 fireworms using the mesh traps, based on 60 deployments with each trap type. The mesh trap consistently yielded more fireworms across sites and sampling sessions, with the HI Fishing Site showing the highest abundance of H. carunculata (Figure 6).

This was further confirmed by Poisson GLM (

Table 5. Summary of Poisson GLM results for fireworm abundance across sites for each trap type.

Trap Type	Comparison	Regression Coefficient (β)	Standard Error (SE)	p-Value
Funnel Trap	Site 3 vs. Site 1	1.653	0.364	<0.001
Mesh Trap	Site 2 vs. Site 1	−0.827	0.262	0.002

) which showed statistical significance between the HI Fishing Site and the Reference Site using the funnel trap (p < 0.001), and the LI Fishing Site and the Reference Site using the mesh trap (p = 0.002). With the funnel trap, the HI Fishing Site had an expected fireworm count of about 5.2 times higher than the Reference Site, whereas with the mesh trap, the LI Fishing Site had an expected fireworm count of about 0.44 times that of the Reference Site (Figure 7). The mesh trap also caught larger and heavier individuals than the funnel trap (Figure 8).

The Gamma GLM analysis highlighted further significant differences across the sites in length and weight and trap types as indicated in

Table 6. Summary of Gamma GLM results for fireworm length and weight across sites and trap types.

Variable	Trap Type	Site	Regression Coefficient (β)	Standard Error (SE)	p-Value	Direction of Effect
Length	Funnel	Site 3	0.028	0.0094	0.003	Smaller than Site 1
	Funnel	Site 4	0.024	0.0115	0.035
	Mesh	Site 4	0.052	0.0068	<0.001
Weight	Funnel	Site 3	0.071	0.0209	<0.001	Lighter than Site 1
	Funnel	Site 4	0.071	0.0280	0.011
	Mesh	Site 3	0.018	0.0069	0.008
	Mesh	Site 4	0.096	0.0142	<0.001

and Figure 9 below.

3.2.2. Onboard Field Observations of Trammel Net Fishing

Onboard observations of trammel net fishing were made as this was the fishing gear most used by the interviewed fishers and therefore most impacted by fireworms.

Of all species caught, including invertebrate by-catch (

Table 7. Number of individuals attacked by fireworms from the total catch during onboard observations of trammel net fishing. For attacked species, the total number of individuals of the species caught in a session is also given. ‘0’ denotes that the species was caught but not attacked; ‘-’ denotes species that were not caught during the particular onboard session.

		Onboard Session
		1	2	3	4	5
		Sept 2023	Oct 2023	Dec 2023	Dec 2023	Mar 2024
Class	Species caught
Teleostei	Chelon sp.	-	0	0	-	-
	Dentex sp.	-	-	-	0	0
	Diplodus sargus	-	-	-	-	0
	Diplodus vulgaris	0	-	0	-	0
	Echiichthys vipera	-	-	0	-	-
	Lichia amia	-	-	0	-	-
	Mullus surmuletus	6	-	7	5	3
	Total no. of individuals caught	17	-	13	8	6
	Muraena helena	0	-	0	0	-
	Sarpa salpa	-	-	-	0	-
	Scorpaena sp.	0	-	0	0	1
	Total no. of individuals caught	9	-	5	11	7
	Serranus scriba	0	-	-	-	-
	Sparisoma cretense	48	-	-	0	-
	Total no. of individuals caught	223	-	-	4	-
	Sparus aurata	-	-	0	0	-
	Symphodus sp.	-	-	-	0	0
	Synodus saurus	0	-	-	-	-
Elasmobranchii	Raja sp.	-	-	-	-	0
Cephalopoda	Octopus vulgaris	-	-	-	-	0
	Sepia officinalis	-	0	0	0	0
Gastropoda	Aplysia sp.	-	-	-	-	0
	Charonia tritonis	0	-	-	-	-
Bivalvia	Pinna rudis	-	-	-	0	-
Malacostraca	Erugosquilla massavensis	-	-	-	-	0
	Portunus segnis	-	0	0	-	-
	Maja squinado	0	-	-	-	-
Asteroidea	Echinaster sepositus	0	-	-	-	-
Total no. of fireworms caught		315	0	8	90	26

), H. carunculata was the most frequently caught species in three of the five onboard observation sessions (a total of 439 individuals). Moreover, when summing up all catches from the five onboard sessions, this species was the most abundant. Onboard session 2, which specifically targeted the blue crab Portunus segnis (21 individuals caught), yielded no fireworms feeding on the catch or entangled in the net.

The total percentage of attacked individuals of M. surmuletus, S. cretense, and Scorpaena sp., amounted to 48%, 22%, and 3%, respectively. Considering only these three attacked species, the total percentage of attacked catch for the five onboard observation sessions collectively was 23%. Based on Marsaxlokk street fish-market prices in 2023, this would have amounted to an estimated loss of €210 out of a potential profit of €400 (52.5%).

When considering numbers rather than percentages, S. cretense was attacked the most, followed by M. surmuletus, likely because these were the most commonly caught species during the onboard sessions.

Damage to the catch ranged from ‘slight’ to ‘severe’, with the ‘severe’ cases resulting in only the skeleton and patches of skin remaining. This was infrequent, but ‘significant’ damage (see Table 4) was very common in S. cretense and M. surmuletus. Only one individual of Scorpaena sp. showed signs of fireworm attack (one fireworm was found attached to skin), contradicting a belief held by some fishers (various personal communications to A.S., 2023) that fireworms do not attack scorpionfish. Fireworm aggregations were also observed entangled in fishing nets, where catch was absent.

During the onboard sessions, the mechanical hauling of nets was occasionally observed to scatter fireworms on the deck (see Figure 10), with some landing on the fisher’s bare skin.

3.3. Behavioral Experiments

3.3.1. Experiment 1—Drop Test

In no trial was swimming observed. Moreover, the sinking worms were not even able to twist their bodies from ventral side up to dorsal side up and only righted themselves once they landed on the bottom of the tank. This shows that in this particular experiment, fireworms show no swimming behavior.

3.3.2. Experiment 2—Effect of Mesh Size

The critical mesh size for the selected range of fireworms (length: 13–19 cm; weight: 12–24 g) was 0.7 cm × 0.7 cm with an 80% success rate of passing through the mesh. The smaller mesh size (0.5 cm × 0.5 cm) resulted in a 100% failure rate, and the larger mesh sizes (1.0 cm × 1.0 cm; 1.8 cm × 1.8 cm) resulted in 100% success rates.

3.3.3. Experiment 3—‘Body Lifting’

None of the starved fireworms were observed to lift the anterior part of their body to reach up for the food. However, 15 fireworms from Experiment 1 (the drop test), were recorded to lift the anterior part of their body up to almost their full length by leaning against the corners of the smooth, plastic tank in which the test was held. However, all fireworms failed to gain traction on the smooth surface and eventually fell backward on their dorsal side.

In the holding tanks, some fireworms were also observed spontaneously raising their bodies, up to around half their body length, in the absence of a vertical surface to lean against.

3.3.4. Experiment 4—‘Climbing of a Vertical Nylon Rope’

Traps were set up in an area where a large population of fireworms had been previously observed and was present. No individuals were found in any trap, even though worms could be observed crawling on the seabed. This indicates that, under the tested in-situ conditions, H. carunculata did not crawl up a 15 cm long and 5 mm thick nylon rope to access T. mediterraneus bait.

4. Discussion

This study provides one of the first detailed assessments of the ecological and socio-economic interactions between H. carunculata and artisanal fisheries in the Mediterranean. Owing to the limited scientific literature on this issue, fishers’ LEK, supported here by empirical data, offers a valuable perspective, particularly in communities with long-standing fishing traditions and experience such as those at Marsaxlokk.

LEK surveys showed that although fishers possess sound knowledge of fireworm habits and behavior, they generally do not recognize how their own practices contribute to the species’ proliferation. The fact that only one fisher recognized the connection between their own practices and the increase in fireworm population highlights a gap in understanding. This gap reinforces a feedback loop, whereby fishing activities both sustain and are negatively affected by growing fireworm populations. Fishers’ belief that fireworm abundance is increasing with warming sea temperatures is supported by García-Monteiro et al. [32], who reported a warming trend in the Mediterranean Sea of 0.040 ± 0.001 °C per year between 2003 and 2019.

Field sampling corroborated the LEK findings. Fireworms were consistently smaller and lighter within Marsaxlokk Harbour, with the HI Fishing Site showing the highest abundance and markedly smaller individuals compared with the Reference Site. The LI and MI Fishing Sites had lower expected counts, but still reflected a clear positive relationship between fishing intensity and fireworm numbers. These patterns indicate a rapidly growing population, supported both by transport into Marsaxlokk Harbour via benthic fishing gear and by constant food input from fishery discards. These results build on the work of Cenni et al. [33] and Simonini et al. [25] and are, to our knowledge, the first to quantify the fishers’ impact on fireworm abundance. Observations of fireworms curling up and entangling in nets, and being unintentionally or intentionally cut, further suggest that fragmentation, combined with the species’ strong regenerative capacity [3,31,34], may unintentionally enhance population increase, underscoring the need for better fisher awareness.

Onboard observations highlighted the economic implications of these interactions. Across five onboard sessions, fireworm attacks led to an estimated 52.5% loss in expected profit. Because most surveys and observations occurred in winter when fishers reported fireworm numbers to be low, summer losses are likely to be even higher as shown by Rescio et al. [12]. The proportion of attacked fish (23%) aligns with studies from nearby regions, such as Lampedusa for trammel nets (20% attacked) [11] and on the southeastern Sicilian coast for bottom longlines (38% attacked) [2], emphasizing the wider Mediterranean relevance of this problem. Continued monitoring across the basin is therefore needed to determine the geographic extent and severity of fireworm impacts.

Because of high consumer expectations, fireworm damage reduces the marketability of otherwise good products. Beyond direct damage, fireworms reduce gear efficiency and increase labor requirements, further diminishing profitability. According to Tiralongo et al. [2], the indirect negative impact of fireworms due to financial loss is more detrimental than direct damage. Entangled worms also expose fishers to painful stings, particularly when mechanical hauling fragments individuals and scatters them onto decks as seen during this study’s onboards. Although fishers tended to downplay the discomfort due to fireworm stings, it remains a significant occupational nuisance. Apart from fishers, fireworms also pose risks to other sea-users as starkly illustrated by a reported case of envenomation by the bearded fireworm, in which pruritus persisted for three weeks [35].

The fireworm’s broad dietary plasticity, opportunistic scavenging, avoidance of energetically unprofitable prey, and ability to survive prolonged starvation contribute to its resilience and expansion and make it difficult to control. In the onboard observations, fireworms ignored the presumably less profitable prey items, such as most of the invertebrate by-catch species. This is contrary to earlier reports which showed fireworms to feed frequently on starfish [5] and less frequently on mollusks (including octopus, bivalves, and sea slugs) [5,13]. Moreover, fish species previously documented as targets elsewhere (Diplodus sp., M. helena, [2]) remained unaffected in the presence of the highly attacked S. cretense (48%) and M. surmuletus (22%). Cuttlefish (S. officinalis), was also not attacked, for unknown reasons. This suggests that the fireworm opportunistically switches from active predation to cost-effective scavenging when moribund or immobilized fish are present, as also speculated by Toso et al. [23], likely to maximize energetically efficient foraging. Therefore, the lack of predators and the long list of prey species, on top of its ability to survive prolonged fasting—starved individuals in captivity, remained alive beyond a year (A.S. personal observations, 2023)—allows H. carunculata to establish new populations and to increase in numbers. The absence of fireworm attacks during the second onboard session where P. segnis were targeted suggests that the crab’s defensive morphology may deter attacks. Scorpion fishes (Scorpaena spp.), although they can be attacked, the attached fireworm was unable to penetrate the fish’s skin, which indicates that the tough skin may deter fireworms.

Although high fireworm numbers may help process discards on the seabed, their negative economic consequences outweigh any ecological “clean-up” value from the fishers’ perspective. Understanding how fireworms interact with demersal fishing gear is therefore crucial for developing effective mitigation measures. The behavioral experiments suggest that fireworms likely enter nets by lifting the anterior portion of their body (as previously also documented by Toso et al. [23]) and crawling across collapsed sections of netting resting on the seabed when currents and wave action bring the nets closer to the bottom. Strong currents can also occasionally suspend fireworms into the water column (A.S. personal observation, 2025), facilitating their entry into the nets, where they attach to prey. Although fishers often described fireworms as capable of swimming short distances, our results align with existing scientific knowledge that the worms are unable to swim [36], highlighting the need to validate LEK through controlled observations. Furthermore, Experiment 4 suggests that inserting a short vertical nylon connector rope between the netting and footrope to reduce net-seabed contact could limit worm access. Additional modifications, such as larger buoys to keep nets more elevated and taut, may further reduce fireworm interactions. These modifications require field testing, as well as outreach and education, since some fishers expressed skepticism during personal communication to A.S. (2023), citing concerns over reduced catch efficiency for high-value demersal species.

During the interviews, fishers’ expressed interest in using fireworm traps, similar to the fireworm-attracting devices developed by Simonini et al. [25], offering a potential strategy to mitigate fireworm surpluses in fishing areas. While effective at trapping fireworms, this measure would require large-scale institutional support and funding to produce a visible and meaningful reduction, as it would need to be implemented frequently and consistently across fishing ports to offset the continual influx of fireworms introduced by fishing activities and sustained by fishery discards—as also pointed out by fishers during personal communications to A.S. (2023). Once trapped and collected, these fireworms could be repurposed in sustainable applications, for example, to clean discarded shellfish and recover calcium-rich shells, or incorporated into integrated aquaculture systems as a bioremediator, creating a practical use for the species while simultaneously helping to control its abundance [37].

Other strategies, such as disposing of discards farther offshore, could reduce nearshore food availability for fireworms, though this may be logistically challenging for many fishers as cleaning nets at sea instead of at the port carries a risk of catch spoilage, particularly during summer due to the time required. While some fishers are now fishing in deeper waters or presumed fireworm-free areas to reduce encounters, future climate-driven expansion of this thermophilic species as sea temperatures rise [26,27] may limit the long-term effectiveness of such strategies, especially if H. carunculata spreads into deeper water, as reported by some interviewed fishers.

Overall, mitigating the “fireworm problem” remains complex. Fishers face a difficult trade-off between altering long-standing fishing practices and reducing fireworm-related losses, a challenge compounded by the limited availability of mitigation strategies that are both practical and consistently effective under real fishing conditions. Measures developed without direct engagement with fishers risk losing effectiveness due to limited feasibility within daily fishing operations, underscoring the need for approaches grounded in the operational realities and practical constraints of daily fishing practices.

In our study, many interviewed fishers over 60 reported feelings of helplessness, whereas Rescio et al. [12] similarly observed greater reluctance among fishers over 50 but found that younger fishers were more collaborative and willing to participate in mitigation initiatives. This generational difference highlights an opportunity to address the fireworm problem through targeted research, education, and coordinated management developed in close collaboration with fishers, researchers, and other stakeholders from the outset.

By co-designing and demonstrating mitigation measures in partnership with willing fishers, practical solutions that align with artisanal fishing practices can be developed and tested, increasing the likelihood of broader adoption as measurable benefits become evident. Such inclusive, collaborative approaches are essential for implementing effective gear and practice modifications and for breaking the self-perpetuating cycle in which fishing activities inadvertently sustain H. carunculata to the detriment of both fishers and coastal ecosystems.