Feeding Habits of the Invasive Atlantic Blue Crab Callinectes sapidus in Different Habitats of the SE Iberian Peninsula, Spain (Western Mediterranean)

Abstract

The blue crab Callinectes sapidus Rathbun, 1896 is native to the western coast of the Atlantic Ocean. Although its arrival to the Mediterranean was probably due to ballast water, this species has several characteristics that have enabled it to successfully invade countless localities in the Mediterranean and the Black Sea. Little is known about its feeding habits and ecosystem impacts in the Mediterranean basin. This study aimed to provide information on the natural diet of C. sapidus by comparing the stomach contents of specimens caught in different seasons and habitats of the SE Iberian Peninsula (hypersaline waters in Mar Menor Lagoon and brackish waters in Guardamar Bay). This study also tested whether gender influences prey selection and if ovigerous females exhibit limited feeding activity. Regarding the frequency of occurrence, the results indicated that in Mar Menor Lagoon the most frequently consumed prey were Crustacea (60%), followed by fish (57%) and Mollusca (29%), whilst in Guardamar Bay, Mollusca (40%), sediment (32%), algae (24%) and Crustacea (24%) were dominant. It has been determined that this species predates heavily on Mediterranean shrimp Penaeus kerathurus, an economically important shrimp species in the lagoon area. Analysis using a generalised linear model indicated that sex, season and size class were factors that significantly influenced the stomach content weight. Furthermore, non-ovigerous females had significantly fuller stomachs than ovigerous individuals. Since the population of Callinectes sapidus tends to increase in the Mediterranean basin, monitoring of its feeding ecology is recommended to determine its impact on the ecosystem.

1. Introduction

Brachyuran crabs can show very different foraging strategies, including filter feeding, scavenging, sand/mud cleaning, plant feeding and predation [1,2]. Some environmental factors (e.g., water temperature, salinity, temporal changes in prey abundance species, in-site nutrient composition) influence their food intake and feeding behaviour [3,4,5]. Therefore, seasonal variations influence the feeding habits and diet of crabs. To this, we must add variability due to ontogeny as body size increases and competitive abilities improve. Moreover, in some crab species ovigerous females do not feed [6,7]. On the other hand, fishing-induced perturbations can also influence the food availability and intake by crabs in different ways: crabs can become very vulnerable after losing appendices upon contact with bottom-towed fishing gears [8], or they can benefit from increased food availability from discarded catch thrown overboard by fishers [9].

The blue crab Callinectes sapidus Rathbun, 1896 is native to the western coast of the Atlantic Ocean, and due to its invasive character and high tolerance of a wide range of key environmental factors such as salinity, its distribution has spread over a wide area covering the Mediterranean and the Black Sea [10,11,12]. C. sapidus has a relatively short lifespan, with a life expectancy of up to 3 years, but it can reach a maximum size of ~23 cm carapace width (CW) in males [13,14]. Thus, the Atlantic blue crab’s maximum size, morphology, omnivorism and resistance to variable environmental conditions make this species a very able competitor and coloniser of new locations. Hines [15] conducted an extensive review of stomach contents and other feeding observations and indicated that the Atlantic blue crab’s diet might include at least 99 species from several phyla, especially molluscs (20–40%), arthropods (10–26%), chordates (fishes; 5–12%) and annelids (polychaetes; 1–7%). Furthermore, the diet often includes significant contributions of plant material (1–20%) and algae (3–30%), as well as sediments (up to >50%) when prey might become scarce [16]. Yet, the Atlantic blue crab’s diet also changes during ontogeny, with juveniles inhabiting shallower waters feeding on a variety of small epibiota and infauna and large adults targeting larger and less diverse prey [17,18]. Miller et al. [19] also indicated that the consumption of bivalves was highest (39%) for subadult (60–119 mm) and adult sizes (≥120 mm), whereas recruits (≤59 mm) ingested significantly higher proportions of plant matter (10–12%). Sex might be another source of dietary variability, since microhabitat partitioning is often reported in estuarine mating environments, with large males being captured in shallower or lower salinity areas than mature females [20,21]. Cannibalism occurs frequently [22]. This adaptability to different environments and variable food disposal have contributed to the successful expansion of Atlantic blue crab in the Mediterranean basin [10,23,24,25]. However, the impacts of this invasive species on the Mediterranean ecosystem and its food web are still largely unknown and more research is needed [26].

Invasive alien species can have devastating effects on the ecosystem and seriously alter the food chain [27,28]. Numerous examples involving trophic cascades have been reported [29,30], such as when the ctenophore Mnemiopsis leidyi invaded the Black Sea ecosystem causing a rapid decline in economically important fish species abundance due to their predation on zooplankton, fish eggs and larvae [31,32]. Some studies specifically focussed on the impacts of invasive crab species such as Carcinus maenas and Paralithodes camtschaticus, explaining how they catastrophically altered food webs [33,34]. Even though the direct impact of C. sapidus in Mediterranean food webs is not well known yet, the study by Clavero et al. [35] noted that after its expansion in the Ebro Delta (Spain), the relative abundances of several species (green crab, eel, sandsmelt, toothcarp and grey mullets) showed a meaningful declining trend. Not only its predation but also its competition with other brachyuran crabs makes it inevitable that some balances in the ecosystem may change [23,26].

Some studies have investigated the stomach contents of C. sapidus in different geographical regions of their native area distribution: estuaries in Florida [17], tidal marsh creeks [36], subtropical coastal lagoons of the Gulf of Mexico [37], New Jersey [38] and Chesapeake Bay [39]. Recent studies on feeding ecology are based in the introduced areas around the Atlantic Ocean, such as the southern New England tidal rivers [40] and the Guadalquivir Estuary, under the influence of the Atlantic Ocean [41]. Yet, little information is available about the diet of the invasive blue crab in the Mediterranean, which can adapt to environments as diverse as coastal, estuarine and inland waters [26,42], although some papers evaluated its trophic position for this geographical region [43,44,45,46]. Here, we aimed to provide information on the natural diet of C. sapidus in Spanish waters, as well as to test whether it has different diets in different environments (i.e., lagoon and marine) and seasons and whether gender has an effect on prey selection. Lastly, we tested the hypothesis that ovigerous females exhibit limited feeding activity.

2. Materials and Methods

2.1. Study Area and Data Collection

Sampling was carried out in two locations: the hypersaline Mar Menor coastal lagoon and in marine brackish waters in Guardamar, Spain (Figure 1). The lagoon has a surface area of 135 km², and the maximum and mean depths are 6 m and 3.6 m, respectively. The lagoon is highly influenced by anthropogenic activities (e.g., fisheries, tourism and agriculture) [47]. Of late, there have been algal blooms owing to excessive nutrient loads, nitrogen and phosphorus [48]. The first record of C. sapidus in Mar Menor dates from 2004 [49]. However, its population did not show any increase until 2017, after which its abundance grew exponentially. The second sampling area, Guardamar Bay, is near a port with small-scale fisheries, recreational fisheries and aquaculture activities. This area is under the influence of organic contamination because the Segura River flows into Guardamar Bay [50]. In Guardamar, C. sapidus is not a target species in the small-scale fisheries, but some fishers occasionally catch these crabs. They constitute a very small part of landings, but sale prices oscillated between 0.5–12.6 EUR/kg in 2022 [51]. In this study, crab samples (n = 258) were collected using trammel nets and unbaited fyke nets. In Mar Menor, crab samples (n = 186) were collected in winter 2022 and spring, summer and autumn 2023, whilst in Guardamar, the samples (n = 72) were obtained in summer and winter 2023. The crabs were placed in an isothermal box with ice immediately after collection to stop digestion and quickly brought to the laboratory where they were stored in freezers at −20 °C until dissection [52,53].

2.2. Laboratory Procedure

In the laboratory, the following data were recorded: carapace width with spine (CW), carapace length (CL), right chela length (RChL), which are related to feeding activity [54], and gender. For the measurement of allometric characteristics, a calliper (precision 0.01 mm) was used, while weight data were determined using a precision electronic balance (to the nearest 0.001 g). In addition, the functional maturity status of females (presence of eggs or not) was noted. The stomachs of crabs were extracted after dissection, and the food components of gut contents were separated in Petri dishes containing 70% alcohol and evaluated under a binocular-stereo microscope. Plant and animal remains were identified at the lowest possible taxonomic level. Other items including sediment, fishing nets fibres, inorganic materials and non-identified organic materials were also weighted. Crabs were classified into three size groups: 50–100, >100–150 and >150–200 mm CW.

Stomach fullness was visually determined and classified at 5 levels (Level 1 = empty stomach, Level 2 = 0.1–25%, Level 3 ≥ 25–50%, Level 4 ≥ 50–75%, Level 5 ≥ 75–100%) [55,56].

2.3. Data Analysis

Concerning the diet analysis, the frequency of occurrence (FO) and the percentage points (PPs) were calculated based on [52,53,57], as follows:

$$\mathrm{F}\mathrm{r}\mathrm{e}\mathrm{q}\mathrm{u}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y} \mathrm{o}\mathrm{f} \mathrm{o}\mathrm{c}\mathrm{c}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}-fi=\left({\displaystyle \frac{bi}{n}}\right)\times 100$$

(1)

In this equation, fi is identified as the frequency of occurrence of the food item i, bi is the number of specimens that had the food item i, and n is the total number of crabs investigated:

$$\mathrm{P}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{g}\mathrm{e} \mathrm{p}\mathrm{o}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{s}-pi={\displaystyle \frac{\sum _{j=1}^n{a}_{ij}}{\mathrm{A}}}\times 100$$

(2)

where pi is the percentage points for the food item i, n is total number of crabs investigated, a_ij is the score of the item i in the stomach of the crab j, and A is the total score of all food items for all crabs investigated.

The overall content weight of the stomachs and the weight of the different prey categories were compared between both locations (Mar Menor and Guardamar) and sexes with a Mann–Whitney U test. The Spearman’s rank correlation was used to (i) investigate the correlation between stomach content weight and chela size and (ii) to test for differences in chela size between sexes. The Kruskal–Wallis test was used to define differences in the stomach content weight between the three size groups of crabs considered. Two types of generalised linear model (GLM) were used. In the first model, main effects were taken into account, and the model was used to investigate the impacts of factors such as sex, season and size group on stomach content weight in both sampling locations. The formula of the GLM [58] is the following:

g(μ_i) = β₀ + β₁sex_i+ β₂season_i+ β₃size_group_i

(3)

where:

g(μi) is the link function executed to the expected value of the stomach content weight.

μ_i = E(Yi) is the expected stomach content weight for the ith observation.

β₀ is the intercept.

β₁, β₂, β₃ are the coefficients for the respective factors.

In the second model, the following interactions were taken into account:

g(μ_i) = β₀ + β₁sex_i + β₂season_i + β₃size_group_i + β₄(sex_i × season_i × size_group_i)

(4)

β₁, β₂, β₃ are the coefficients for the respective factors, and β₄ is their interactions.

Since Mar Menor was sampled for four seasons and Guardamar was sampled for two seasons, models were designed separately for each region.

Multivariate analyses were applied to detect differences in the diet composition, where the variables were FO of diet components (fish, crustaceans, molluscs, sediment, vegetable, etc.). Multidimensional scaling (MDS) was performed to visualize the level of similarity of the data set. Analyses of similarity (ANOSIM) were performed to identify significant differences (p < 0.05) among the factors area, season, size classes and sex. Data were previously standardised using square-root transformation. Similarity percentage (SIMPER) analysis was used to determine which major groups contributed the most to the dissimilarity in diet composition between different factors. R software (Ver. 4.3.2) [59], SPSS (Ver. 20) and PRIMER (Ver. 6) [60] were used for the statistical analyses.

3. Results

3.1. Demographic Characteristics

The mean CW (±SD) of crabs in Mar Menor was 134 ± 23 mm, whilst the mean value in Guardamar was 148 ± 22 mm (Figure 2). The distribution of size classes showed significant differences between locations (Mann–Whitney U test: U = 4298, p < 0.001). The size distribution was not significantly different among sampling seasons in Guardamar (Mann–Whitney U test: U = 438, p = 0.130), whereas there was a significant difference in the size distribution of crabs during sampling seasons in Mar Menor (Kruskal–Wallis test, x² = 25.06, p < 0.001). During the study period, the sex ratio (F/M) in Mar Menor was 1.48, while the sex ratio was 4.54 in Guardamar. In Mar Menor, the sex ratio was 3.3 in summer, 0.2 in autumn, 2.6 in winter and 2.2 in spring. In Guardamar, the sex ratio was 3.5 in winter and 6.2 in summer.

3.2. Stomach Fullness

A total of 43 crabs (17%) had empty stomachs (Level 1). A further 12 stomachs (5%) only had sediment or man-made material (MMM). Hence, a total of 203 stomachs (79% of sampled crabs) contained at least one plant or animal item. In this study, 111, 64, 21 and 19 individuals were classified as Level 2 (0.1–25%), Level 3 (>25–50%), Level 4 (>50–75%) and Level 5 (>75–100%), respectively. In addition, a total of 33 stomachs from ovigerous females were analysed, of which 15 contained at least one plant or animal item. Figure 3 shows the proportion of fullness levels of stomachs in the different sampling locations. In Mar Menor, the proportion of crabs with higher levels of stomach fullness (Levels 3, 4 and 5) was significantly higher than in Guardamar (Mann–Whitney U test: U = 4358, p < 0.001).

3.3. Factors That Influenced the Stomach Content Weights

There was a statistically significant difference in the stomach content weight depending on the sampling location (Mann–Whitney U test: U = 3881, p < 0.001; mean values ± SD = 0.45 ± 0.60 g in lagoon, 0.14 ± 0.31 g in the bay, Figure 4a) and sex, males having higher content weight than females (Mann–Whitney U test: U = 4282, p < 0.001; mean values ± SD = 0.21 ± 0.32 g for females and 0.66 ± 0.76 g for males, Figure 4b). There was also a significant difference between the stomach content weight of berried and not berried females (Mann–Whitney U test: U = 1610, p = 0.010, the mean values (±SD): 0.13 ± 0.33 g for ovigerous females and 0.22 ± 0.32 g for non-ovigerous females, Figure 4c).

With regard to chela size, males had significantly larger chelipeds (93 ± 17 mm) than females (62 ± 8 mm) (Mann–Whitney U test: U = 274, p < 0.001). The stomach content weight of crabs was not significantly correlated with cheliped size (Spearman’s rank correlation: r = 0.029, p = 0.759).

Concerning crab size, there were significant relationships between CW and stomach content weight (linear regression, r = 0.204, df = 250, p = 0.001). Moreover, the total stomach content weight of crabs significantly increased depending on size groups (Kruskal–Wallis test: x² = 11.771, df = 2, p = 0.003). The mean values (±SD) were 0.03 ± 0.05 g for small crabs, 0.36 ± 0.49 g for medium-sized crabs and 0.47 ± 0.73 g for large crabs.

Seasonality also had an effect on stomach content weight in the Mar Menor Lagoon (Kruskal–Wallis test: x² = 60.640, df = 3, p < 0.001,

Table 1. The frequency of occurrence (FO) and mean content weight in different seasons in Mar Menor.

Prey	Summer		Autumn		Winter		Spring
	FO (%)	Content (Mean ± SE) (g)	FO (%)	Content (Mean ± SE) (g)	FO (%)	Content (Mean ± SE) (g)	FO (%)	Content (Mean ± SE) (g)
PLANTS/ALGAE (VEG)	22	<0.01 ± <0.01	7	<0.01 ± <0.01	28	<0.01 ± <0.01	43	0.02 ± 0.01
POLYCHAETA (ANN)	14	<0.01 ± <0.01	15	0.01 ± <0.01	3	<0.01 ± <0.01	9	<0.01 ± <0.01
CRUSTACEA (CRU)	51	0.13 ± 0.03	74	0.57 ± 0.11	28	0.01 ± 0.01	89	0.17 ± 0.03
Amphipoda	7	<0.01 ± <0.01	2	<0.01 ± <0.01	0	-	40	<0.01 ± <0.01
Isopoda	1	<0.01 ± <0.01	2	<0.01 ± <0.01	0	-	9	<0.01 ± <0.01
Cirripedia	4	<0.01 ± <0.01	4	<0.01 ± <0.01	0	-	0	-
Brachyura	18	0.01 ± <0.01	17	0.02 ± 0.01	19	0.01 ± <0.01	31	0.01 ± <0.01
Paguridae	0	-	0	-	0	-	0	-
Panaeus kerathurus	38	0.11 ± 0.03	59	0.40 ± 0.09	9	<0.01 ± <0.01	66	0.14 ± 0.03
Palaemon sp.	8	0.01 ± <0.01	4	0.04 ± 0.04	3	<0.01 ± <0.01	14	0.01 ± <0.01
Other Natantia	7	0.01 ± 0.01	26	0.11 ± 0.04	3	<0.01 ± <0.01	17	<0.01 ± <0.01
MOLLUSCA (MOL)	27	0.02 ± 0.01	22	0.07 ± 0.03	22	0.03 ± 0.03	46	0.02 ± 0.01
Bivalvia	25	0.02 ± 0.01	20	0.06 ± 0.03	19	0.03 ± 0.03	46	0.02 ± 0.01
Gastropoda	6	<0.01 ± <0.01	2	<0.01 ± <0.01	3	<0.01 ± <0.01	0	-
Cephalopoda	3	<0.01 ± <0.01	4	<0.01 ± <0.01	0	-	0	-
ECHINODERMATA (ECH)	4	<0.01 ± <0.01	0	-	0	-	3	<0.01 ± <0.01
Asteridae	4	<0.01 ± <0.01	0	-	0	-	0	-
Echinoidea	0	-	0	-	0	-	3	<0.01 ± <0.01
FORAMINIFERA (FOR)	0	-	4	<0.01 ± <0.01	0	-	0	-
PISCES (FIS)	49	0.15 ± 0.04	65	0.29 ± 0.06	47	0.02 ± 0.01	71	0.12 ± 0.04
AVES (AVE)	0	-	0	-	0	-	6	0.01 ± 0.01
UNIDENDIFIED ORG. (UOM)	4	<0.01 ± <0.01	0	-	0	-	6	<0.01 ± <0.01
SEDIMENT (SED)	22	<0.01 ± <0.01	28	0.03 ± 0.01	9	<0.01 ± <0.01	26	0.01 ± 0.01
MAN-MADE-MATERIAL (MMM)	7	<0.01 ± <0.01	11	0.02 ± 0.01	9	0.02 ± 0.01	11	<0.01 ± <0.01

Note: cells related to content including ‘0’ values.

). The highest content weight was recorded in autumn, and the lowest value was found in winter. There were no seasonal significant differences in the samples from Guardamar Bay (Mann–Whitney U test: U = 420, p = 0.077,

Table 2. The frequency of occurrence (FO) and mean content weight in different seasons in Guardamar.

Prey	Summer		Winter
	FO (%)	Content (Mean ± SE) (g)	FO (%)	Content (Mean ± SE) (g)
PLANTS/ALGAE (VEG)	25	0.02 ± 0.01	22	0.01 ± <0.01
POLYCHAETA (ANN)	8	<0.01 ± <0.01	11	<0.01 ± <0.01
CRUSTACEA (CRU)	25	0.04 ± 0.02	22	0.01 ± <0.01
Amphipoda	0	-	3	<0.01 ± <0.01
Isopoda	0	-	8	<0.01 ± <0.01
Cirripedia	0	-	0	-
Brachyura	6	0.01 ± <0.01	11	<0.01 ± <0.01
Paguridae	6	<0.01 ± <0.01	3	<0.01 ± <0.01
Panaeus kerathurus	17	0.02 ± 0.01	8	<0.01 ± <0.01
Palaemon sp.	6	<0.01 ± <0.01	0	-
Other Natantia	6	0.01 ± 0.01	6	<0.01 ± <0.01
MOLLUSCA (MOL)	42	0.02 ± 0.01	39	0.02 ± 0.01
Bivalvia	36	0.02 ± 0.01	39	0.02 ± 0.01
Gastropoda	3	<0.01 ± <0.01	6	<0.01 ± <0.01
Cephalopoda	3	<0.01 ± <0.01	0	-
ECHINODERMATA (ECH)	0	-	3	<0.01 ± <0.01
Asteridae	0	-	0	-
Echinoidea	0	-	3	<0.01 ± <0.01
FORAMINIFERA (FOR)	0	-	0	-
PISCES (FIS)	25	0.07 ± 0.05	17	0.01 ± 0.01
AVES (AVE)	0	-	0	-
UNIDENDIFIED ORG. (UOM)	6	<0.01 ± <0.01	3	<0.01 ± <0.01
SEDIMENT (SED)	36	0.01 ± <0.01	25	<0.01 ± <0.01
MAN-MADE-MATERIAL (MMM)	14	0.04 ± 0.03	14	0.04 ± 0.03

Note: cells related to content including ‘0’ values.

).

Considering the main effects, the generalised linear model (GLM) showed that stomach content weight in Mar Menor crabs was significantly affected by season and size group (Appendix A;

Table A1. Summary of GLM (Model 1) with tests of main effects in Mar Menor.

Source	Type III
	Wald Chi-Square	df	p Value
(Intercept)	71.092	1	<0.001
Sex	3.617	1	0.057
Season	35.549	3	<0.001
Size group	7.488	2	0.024

Note(s): dependent variable: content amount. Model: (Intercept), sex, season, size group.

and

Table A2. Summary of GLM (Model 1) with parameter estimators in Mar Menor.

Parameter	B	Std. Error	95% Wald Confidence Interval		Hypothesis Test
			Lower	Upper	Wald Chi-Square	df	p Value
(Intercept)	0.628	0.1123	0.407	0.848	31.214	1	<0.001
[sex = female]	−0.174	0.0916	−0.354	0.005	3.617	1	0.057
[sex = male]	0 a	.	.	.	.	.	.
[season = summer]	0.016	0.1021	−0.185	0.216	0.024	1	0.878
[season = autumn]	0.563	0.1216	0.325	0.802	21.485	1	<0.001
[season = winter]	−0.190	0.1345	−0.454	0.074	1.991	1	0.158
[season = spring]	0 a	.	.	.	.	.	.
[size_group = 1]	−0.346	0.1619	−0.664	−0.029	4.574	1	0.032
[size_group = 2]	−0.226	0.0971	−0.416	−0.035	5.407	1	0.020
[size_group = 3]	0 a	.	.	.	.	.	.
(Scale)	0.243 b	0.0253	0.198	0.298

Notes: dependent variable: content amount. Model: (Intercept), sex, season, size_group. a = Set to zero because this parameter is redundant, b = Maximum likelihood estimate.

), while in Guardamar, the main factors were season and sex (Appendix A;

Table A4. Summary of GLM (Model 1) with tests of main effects in Guardamar.

Source	Type III
	Wald Chi-Square	df	p Value
(Intercept)	19.609	1	<0.001
Sex	4.185	1	0.041
Season	5.466	1	0.019
Size group	0.683	1	0.409

Notes: dependent variable: content amount. Model: (Intercept), size_group, sex, season.

and

Table A5. Summary of GLM (Model 1) with parameter estimators in Guardamar.

Parameter	B	Std. Error	95% Wald Confidence Interval		Hypothesis Test
			Lower	Upper	Wald Chi-Square	df	p Value
(Intercept)	0.284	0.0951	0.097	0.470	8.888	1	0.003
[sex = female]	−0.218	0.1066	−0.427	−0.009	4.185	1	0.041
[sex = male]	0 a	.	.	.	.	.	.
[season = summer]	0.174	0.0744	0.028	0.320	5.466	1	0.019
[season=winter]	0 a	.	.	.	.	.	.
[size_group=2]	−0.062	0.0752	−0.210	0.085	0.683	1	0.409
[size_group=3]	0 a	.	.	.	.	.	.
(Scale)	0.087 b	0.0150	0.062	0.121

Notes: dependent variable: content amount. Model: (Intercept), sex, season, size_group. a = Set to zero because this parameter is redundant, b = Maximum likelihood estimate.

). However, when interactions were considered, the combined effect of sex, season and size group had a statistically significant effect on stomach content weight in Mar Menor but no significant effect in Guardamar (Appendix A;

Table A3. Summary of GLM (Model 2) with tests of interactions in Mar Menor.

Source	Type III
	Wald Chi-Square	df	p Value
(Intercept)	55.491	1	<0.001
Sex	0.497	1	0.481
Season	36.683	3	<0.001
Size group	16.004	2	<0.001
Sex x season x size group	28.577	11	0.003

Notes: dependent variable: content amount. Model: (Intercept), size group, sex, season, sex x season x size group.

and

Table A6. Summary of GLM (Model 2) with tests of interactions in Guardamar.

Source	Type III
	Wald Chi-Square	df	p Value
(Intercept)	13.789	1	<0.001
Sex	2.476	1	0.116
Season	4.521	1	0.033
Size group	1.670	1	0.196
Sex x season x size group	1.195	3	0.754

Notes: dependent variable: content amount. Model: (Intercept), size_group, sex, season, sex x season x size group.

).

Mar Menor (MM) stomachs contained significantly more fish and crustaceans than Guardamar (GU) stomachs, while there were no significant spatial differences for other prey groups (

Table 3. Mean consumption of prey groups by crabs of different locations and sexes and summaries of Mann–Whitney U test.

Prey	Mean Value (g) (±SE) at Locations		U Value	p Value
	Mar Menor (MM)	Guardamar (GU)
Fish	0.16 ± 0.03	0.04 ± 0.03	4079	<0.001
Crustacea	0.23 ± 0.03	0.02 ± 0.01	3914	<0.001
Mollusca	0.03 ± 0.01	0.02 ± 0.01	6129	0.201
Polychaeta	0.003 ± 0.001	0.001 ± 0.001	6594	0.724
Plants/Algae	0.01 ± 0.002	0.01 ± 0.004	6519	0.656
Sediment	0.01±0.003	0.01 ± 0.001	6174	0.197
Man Made Material	0.01 ± 0.01	0.04 ± 0.02	6353	0.229
Prey	Mean Value (g) (±SE) at Sexes		Uvalue	p Value
	♂	♀
Fish	0.23 ± 0.05	0.07 ± 0.02	5588	<0.001
Crustacea	0.34 ± 0.06	0.08 ± 0.01	5329	<0.001
Mollusca	0.04 ± 0.01	0.03 ± 0.01	7331	0.750
Polychaeta	0.004 ± 0.002	0.002 ± 0.001	7259	0.470
Plants/Algae	0.003 ± 0.001	0.01 ± 0.002	6641	0.046
Sediment	0.02 ± 0.01	0.01 ± 0.002	6533	0.027
Man Made Material	0.02 ± 0.02	0.01 ± 0.01	7374	0.725

). Furthermore, the stomachs of males included significantly more fish, crustaceans, plants/algae and sediment than female stomachs (Table 3).

3.4. The Frequency of Occurrence (FO) and the Percentage Points (PPs) of Prey Groups

Regarding the frequency of occurrence (FO), the most frequently preferred prey group in the Mar Menor Lagoon was Crustacea (60%), followed by fish species (57%) and Mollusca (29%). While in Guardamar, the most frequent prey was Mollusca (40%), followed by sediment (32%), algae (24%) and Crustacea (24%) (Figure 5a). The economically important shrimp species Penaeus kerathurus (Forskål, 1775) was observed in 44% and 13% of the stomachs in Mar Menor and Guardamar, respectively (Table 1 and Table 2). Concerning the percentage points (PPs), the dominant group was Crustacea with 51% dominance in Mar Menor, followed by fish (35%) and Mollusca (8%) (Figure 5b). While in Guardamar, the dominant group was fish (31%) followed by man-made material (e.g., micro-plastics, nets, ropes) (27%) and Crustacea (18%) (Figure 5b). Echinodermata, foraminifera, birds (the individual birds were presumably already dead prior to interaction, and the crabs exhibited scavenging behaviour) and unidentified organic materials were rarely observed in stomachs (Figure 5a,b).

Concerning the FO, the results showed that 42% of female stomachs included fish, 41% of stomachs included Crustace and 33% included Mollusca. Crustacea were found in 65% of the stomachs of male crabs, fish in 59% and sediment in 31% (Figure 6a). Similarly, concerning the prey weight in terms of PPs, the heaviest items were Crustacea and fish for both females (40% and 33%, respectively) and males (52% and 35%, respectively) (Figure 6b).

Furthermore, our study showed differences in the diet between size groups. Plant/algae and Crustacea were the most frequently observed group in small-sized crabs (50–100 mm CW); 38% of stomachs included these items. The highest FO value in medium-sized crabs (100.1–150 mm CW) was for Crustacea (55%), and the highest FO values in the largest crabs (150.1–200) corresponded to fish (49%).

Multivariate analysis of FO corroborates these results. The ANOSIM test found slight significant differences for the interaction between location and season (R = 0.336; p = 0.001), which it can show in the MDS graph (Figure 7). These results are due to the differences between the samples of Mar Menor and Guardamar in summer (R = 0.343; p = 0.001) and between Mar Menor–summer and Guardamar–winter (R = 0.349; p = 0.001). The analysis did not show differences between size classes and sex, nor did it when it was compared as an orthogonal factor related to location–season (

Table 4. Results of ANOSIM analysis for diet components (%FO) between the different factors considered in the test. Location = MM (Mar Menor) and GU (Guardamar); season = + (hot) and * (cold); sex = F (female) and M (male); size group = 1 (<50 cm CW), 2 (50–100 cm CW) and 3 (>100 cm).

Source of Variation	R	p
Location	0.336	0.001
Season	0.186	0.001
Location x Season	0.308	0.001
Pairwise Test MM *, MM+	0.271	0.001
Pairwise Test MM *, GU *	0.124	0.002
Pairwise Test MM *, GU+	0.098	0.005
Pairwise Test MM +, GU *	0.349	0.001
Pairwise Test MM +, GU +	0.343	0.001
Pairwise Test GU *, GU +	−0.04	0.917
Sex	0.009	0.249
Size Group	0.103	0.001
Pairwise Tests 1–3	0.165	0.007
Pairwise Tests 1–2	0.265	0.003
Pairwise Tests 3–2	0.056	0.029
Size Group (Location x Season)	0.029	0.389
Sex (Location x Season)	0.25	0.114

). SIMPER analysis resulted in 87.33% dissimilarity between summer MM and GU samples, due to the different presences of crustaceans and fish. The dissimilarity between MM–summer and GU–winter was 89.54%, and the diet components contributing most to these differences were fish, molluscs and crustaceans (

Table 5. Results of SIMPER analysis. Location = MM (Mar Menor) and GU (Guardamar); Season = + (hot) and * (cold).

	Contribution of Main Component Diet (%)
Groups (% Dissimilarity)	Crust	Fish	Mollusc	Inorg	Algae	Sedim
MM * and MM+ (87.8%)	40.2	36.6	11.3	5.4		4.5
MM * and GU * (87.3%)	17.8	21.1	24.9	14.6	10.2	8.1
MM+ and GU * (89.9%)	39.1	29.7	12.8	6.6	5.0
MM * and GU+ (89.5%)	22.2	25.3		15.1	12.5	9.9
MM+ and GU+ (87.3%)	38.7	31.8	8.6	7.8	6.9
GU * and GU+ (86.6%)	21.9	17.7	18.9	17.1	13.9	7

).

4. Discussion

The present study showed that more than ninety percent of the C. sapidus diet consisted of crustaceans, fishes and molluscs in Mar Menor Lagoon. In Guardamar Bay, the dominant groups were fish, man-made material and crustaceans. These differences are thought to be primarily influenced by variations in habitat characteristics, water parameters and nutrient levels across the sampling sites. Studies published in recent years highlight a marked increase in the population of C. sapidus in the Mediterranean Sea, accompanied by a decline in the abundance of certain native fish and crustacean species [35]. Another study on the trophic ecology of C. sapidus from a different region of the Mediterranean (Bardawil Lagoon, northern Sinai coast, Egypt) affected by blue crab invasion reported that the most common prey consisted of molluscs, crustaceans and fish species, similar to our findings [61]. Kampouris et al. [26] collected data on the feeding habits of C. sapidus in the eastern Mediterranean and reported a total of 13 prey species (10 molluscs, 2 crustaceans and 1 fish) in the Thermaikos Gulf, Aegean Sea, whilst they found 35 species (16 fish, 9 crustaceans and 8 molluscs) in Papapouli Lagoon. The aforementioned study also claimed that C. sapidus mainly feeds on economically important species (e.g., Mytilus galloprovincialis, Ostrea edulis, Penaeus kerathurus, Dicentrarchus labrax, Mugil cephalus and Mullus surmuletus). Ortega-Jiménez et al. [41] investigated the feeding habits of C. sapidus in Guadalquivir Estuary (east Atlantic coast of Spain) and noted that the main identified food items were fish (49.9%), molluscs (44.4%) and crabs (32.3%). The findings from the Guadalquivir Estuary are consistent with the results we obtained in the Mar Menor Lagoon. On the other hand, the main diet of C. sapidus in the Apalachicola Estuary, Florida, United States consisted of bivalves (35.7%), fishes (11.9%), xanthic crabs (11.4%) and blue crabs (9%), as well as shrimp. Gastropods and plant matter were the less preferred prey [17]. Another study carried out in Alabama reported the cumulative index of relative importance (IRI) consisted of four main groups (fish, bivalves, brachyurans and digested animal tissue) with 91% dominance [62]. This crab species is considered a scavenger and consumes dead fish and crustaceans like crabs and shrimps, as well as plant species in Chesapeake Bay [63].

The present study confirmed the findings in Vivas et al. [42] regarding the Atlantic blue crab’s preference for the most important commercial species in the Mar Menor Lagoon, Panaeus kerathurus. This species was found in fifty percent of the crabs in the lagoon whose stomachs were not empty. Moreover, in terms of PPs, our study highlighted that P. kerathurus comprised more than a third of the diet. However, Escudero-Lozano et al. [64] showed that environmental variables best explain the trends of Atlantic blue crab and P. kerathurus and that more data are needed to assess the potential risk posed by C. sapidus. In addition to predation, C. sapidus can influence the feeding habits of other crustaceans (e.g., green crab Carcinus aestuarii Nardo 1847) due to competition [26].

Numerous studies noticed that crabs exhibit different feeding habits in different seasons [60,65,66], in agreement with our results from the Mar Menor Lagoon. The highest fullness score of stomachs was found in autumn, and the lowest value was determined in winter in the lagoon. It has been reported that the highest water temperatures were recorded in August and the lowest in February, while peak salinity levels in the lagoon were observed during autumn [67]. Furthermore, our observations from Guardamar Bay indicated that the stomach fullness level was higher in summer than in winter. This can be related to the metabolic activity of crabs. For example, Hines et al. [68] noted that crabs are more active in summer months. However, there are also exceptions in the literature; another study carried out in Florida (St. Johns’s River) showed that there were no seasonal and spatial differences in the diet of C. sapidus [69]. Moreover, food diversity and prey preference may also show seasonal changes. For example, Laughlin [17] noticed that bivalves were dominant in spring and winter in the stomach contents of C. sapidus in estuaries of Florida. Similarly, the present study indicated FO values of algae, crustaceans, molluscans and fishes were higher in spring compared with other seasons in Mar Menor Lagoon.

The results of the present study showed that the small crabs mainly fed on plants/algae and crustaceans, medium-sized crabs mainly consumed crustaceans and lastly the fish species were the most preferred group for larger crabs. Similarly, Taylor et al. [40] underlined that although large individuals of C. sapidus mainly fed on bivalves, crabs and fishes, small-sized specimens (≤49 mm CW) of this species commonly prefer amphipods, shrimps and other small-sized crustaceans. Another study related to the feeding habits of juvenile specimens of C. sapidus (4–40 mm CW) declared that the gut contents of individuals consisted of clams, amphipods, polychaetes, small crustaceans, plant matter and detritus in Chesapeake Bay [39]. Another key factor related to variations in prey is the sampling locations. For example, our study showed that fish and Crustacea were consumed in the lagoon with a significant difference compared to the bay. Laughlin [17] also noted that the diet variability between stations showed differences, and it decreased with increasing crab size. The same study highlighted that there was a difference in diet in deeper and shallower sampling locations.

Our study rarely found bird remains, as it was considered that the crabs probably fed upon dead individuals. Similarly, Kampouris et al. [26] observed the feeding activity of C. sapidus on dead seagulls and snakes in Thermaikos Gulf, eastern Mediterranean. In addition, our study found that sediment and MMM were represented with relatively high ratios in the contents, especially in Guardamar. Similarly, concerning the existence of litter in the stomachs of C. sapidus, a study by Renzi et al. [70] found metal fragments (13%) and plastics (13% PE; 6.7% PET) in the stomach contents in Lesina Lagoon, Mediterranean.

The present study highlighted that ovigerous females rarely fed, and their stomach fullness levels were lower than non-ovigerous individuals. A similar situation was reported for other crab species. For instance, according to Safaie [71], the percentage of empty stomachs in egg-carrying individuals was higher than non-ovigerous females of the blue swimming crab Portunus segnis. Howard [6] drew attention to the empty stomachs of ovigerous females in the European edible crab Cancer pagurus.

In conclusion, understanding the diets and feeding strategies of animals contributes to a deeper comprehension of symbiotic relationships, resource competition and the overall completeness and functionality of ecosystems [53]. The present study provided novel findings and confirmed the results of previous studies on the feeding habits of C. sapidus in the western Mediterranean. C. sapidus mainly preferred Crustacea, Pisces and Mollusca in our study area. Furthermore, the results emphasized that season, size class and sex were important factors that influenced the stomach contents of C. sapidus.

Since there is a risk of the population of this invasive species increasing, it is recommended that the feeding ecology be followed with studies to be carried out at certain periods in the future in order to better understand its impact on the ecosystem. Stomach content analysis is important for quantitatively assessing the impacts of invasive species, as it provides valuable data on their trophic interactions. For instance, this line of research has proven highly valuable in evaluating the invasive Indo-Pacific lionfish (Pterois volitans), which exerts pressure on native fish species in the Caribbean. In regions with high lionfish abundance, they have been reported to cause significant declines in the recruitment of Atlantic reef fishes [72]. Furthermore, monitoring studies on stomach contents have been emphasized as essential for understanding temporal changes in diet composition at the local scale [73]. Such analyses can aid in the development of effective management strategies that support the sustainable use of resources and enhance efforts to control invasive species.

As C. sapidus is an invasive species consumed in some Mediterranean countries and even exported to Asian and Northern European markets [24], the development of incentive-based policies to promote its fishery is recommended as a means of controlling its population growth. Additionally, integrating the species into local gastronomic culture and implementing the selective culling of egg-bearing females as part of a pilot project are proposed as complementary management strategies. Lastly, comprehensive studies aimed at identifying not only the prey but also the predators of the blue crab in the Mediterranean can be important for developing fisheries’ management tools designed to protect predator species and for contributing to the control of blue crab populations.