Recent expansion of two invasive crabs species Hemigrapsus sanguineus (de Haan, 1835) and H. takanoi Asakura and Watanabe 2005 along the Opal Coast, France

Nowadays, invasions of invertebrate species in coastal ecosystems have become an ineluctable and irreversible phenomenon. The recent introduction of two western pacific crustacean decapods—the Asian shore crab, Hemigrapsus sanguineus, and the brushclawed shore crab, Hemigrapsus takanoi—along the French coast has revealed the problematic effects that invasive species can have on biodiversity and also the competition that invasive species represent for native crab species. This study describes the distribution and abundance of both Hemigrapsus species along the Opal coast on the French side of the Dover Strait in spring 2008. Both species occupy habitats similar to that of the green crab, Carcinus maenas. However, the habitats colonised by the two species are clearly segregated: low hydrodynamic muddy habitats for H. takanoi and high hydrodynamic habitats with fine and medium sands for H. sanguineus. Both species can live in sympatry in harbours. In spring 2008, the maximum density outside Dunkirk harbour was 12 ind.m for H. sanguineus, and inside Dunkirk harbour, the maximum density was 60 ind.m for H. takanoi. In this location, H. takanoi dominated C. maenas significantly. No ovigerous females of either invasive species were found during the spring. Both species have high colonisation potential, ranging from south of the Bay of Biscay to Germany for H. takanoi and from the western part of the English Channel to Germany for H. sanguineus.


Introduction
Among the most recent brachyuran species to have colonised the French Atlantic coastline are two Asian crabs from the genus Hemigrapsus (superfamily Grapsoidea). The brush-clawed shore crab H. takanoi Asakura and Watanabe, 2005 (Figure 1) was initially identified as H. penicillatus (de Haan, 1835) and first recorded in 1994 at La Rochelle on the Atlantic coast of France (Noël et al. 1997). This species was subsequently reported from the harbour at Le Havre in 1999 on the French side of the English Channel (Breton et al. 2002). In the same year (1999), the Asian shore crab H. sanguineus (de Haan, 1835) (Figure 2) was also discovered in Le Havre (Breton et al. 2002). The aims of this study are to present the geographical distribution of both invasive species in the intertidal zone of the Opal coast from the eulittoral zone to evaluate the density of both species in this area, and to estimate the possible competition represented by both nonnative species of Hemigrapsus for the native common shore crab, Carcinus maenas (Linnaeus, 1758). The known distribution of both exotic species in European waters as of July 2008 is also presented.

Field site
The 160 km of the French side of the eastern English Channel from the Authie estuary in the south to the Belgian border in the southern North Sea is know as the Opal Coast ( Figure 3). This coastline is characterised by a succession of sand dunes and rocky shores, extending from Boulogne-sur-mer to Gris Nez Cape and around Blanc Nez Cape. Hard substratum with dykes and boulders are also found at the estuaries of the Authie, Canche and Aa, and in the three main harbours of the region: Boulogne-sur-mer, Calais and Dunkirk. All these potential favourable crab habitats were sampled at the beginning of the spring 2008, from 3 April to 6 May.  Table 1 for the abbreviations of the sampled locations)

Sampling procedure
At each of the 16 sampled sites shown in Figure  3, the crabs collected were found under 3 × 30 boulders in the eulittoral zone (approx. upper, middle and lower) for a total of 270 boulders per site. Although the middle of the eulittoral zone was always sampled at each site, it was not always possible to sample all three parts due to a lack of boulders (Table 1). Depending on the number of boulders at each site, between one and five series of 30 boulders were turned over, with the total number turned over varying between 30 and 270.
In the areas where Hemigrapsus species were present in abundance, an estimation of the species density was made based on a 1 m 2 quadrat, with six replicates being done at each site (except outside Dunkirk harbour, where only three replicates were sampled) ( Table 2). The densities of the crabs were also estimated in subquadrats of 0.25 m 2 in four sites on three square meters at each site, for a total of 12 sub-quadrats (Table 3). Rocks and boulders were turned over to collect the crabs; in some cases, it was also necessary to extract the crabs from the burrows that they had excavated.

Laboratory observations
The collected crabs were identified, counted and sexed. Then the carapace width (CW) was measured between the third antero-lateral teeth (Delaney et al. 2008). The class-size histograms were constructed in 1 mm classes.

Statistical analyses
A variety of tests were used to examine different hypotheses (H0 & H1). A χ 2 test (Scherrer 1984) was used to verify the existence of a distribution pattern: A χ 2 test was used to analyse the difference in density (n.ind.m -2 ) at the five sites where the density was estimated.
-H0: the densities of the three shore crabs were similar in all five sites.
-H1: there was a significant difference between the densities of the five sites.
An equal distribution test (d 2 ) was also performed to test the equi-probability of the samples.
-H0: there was similar chance of sampling all three species.
-H1: one species was dominant. sanguineus and Carcinus maenas were not homogeneous and depended on the sites sampled (χ 2 , p < 0.05). The proportion of H. sanguineus was higher in the sites located to the north of Boulogne-sur-mer (U wmw , p < 0.05), while the proportion of C. maenas was higher in the sites located to the south of Boulogne-sur-mer where the other shore crabs were absent or rare (U wmw , p < 0.05).

Sex-ratio
Among the 2,395 crabs collected during the sampling period (Table 1)

Class size
The frequency distributions of the male and female 1.0 mm carapace width (CW) classes ( Figure 4) indicate a larger extended size class for C. maenas than for the two other species. The smallest C. maenas male was 4 mm, while the smallest female was 3 mm; the largest male was 47.5 mm, and the largest female was 45 mm. Most of the male and female sizes were similar, ranging from 9 to 25 mm (mean size about 16-17 mm) ( Figure 4A). The smallest H. takanoi (male and female) was 7 mm, while the largest female was 20 mm and the largest male was 25 mm.
Most of the males and females were similar in size, ranging from 9 to 19 mm (mean size about 12-13 mm) ( Figure 4B). For the third species, H. sanguineus, the smallest male was 4 mm and the largest, 26 mm; the smallest female was 9 mm and the largest, 27 mm. Most of the males and females were similar in size, ranging from 9 to 20 mm (mean size about 12-13 mm) ( Figure  4C).  Table 1). Statistical tests revealed that the distribution of H. sanguineus was not similar in the three part of the eulittoral (KW, p< 0.01) and that the abundance of the species in the mid-eulittoral was significantly higher than those found in the upper and lower part of the eulittoral (Post Hoc test, p < 0.01).
There was no significant difference between the estimated abundances for the upper and the lower eulittoral. In addition, no specimens of this species were collected in the upper eulittoral at AM, and none were found in the lower eulittoral at LP and GN.

Density of the three shore crabs
The densities of the three species of shore crabs were estimated at the five sites where the abundances of both   Table 3. Number of ind.0.25m -2 in 12 sub-quadrats of the three shore crabs: Carcinus maenas (CM), Hemigrapsus sanguineus (HS) and H. takanoi (HT), estimated at 5 Opal coast sites during spring 2008. SQij: sub-quadrat, where i is the numbering of quadrat and j that of the sub-quadrat. See Figure 1 and Table 2 for the site locations and names For four sites (BHP, WF, GF and DH), 0.25 m 2 sub-quadrats were sampled to identify the respective densities of the three shore crabs (Table 3). For three of the four sites (BHP, WF and GF), there was no significant dominance of the non-indigenous species Hemigrapus over C. maenas. However, at DH, H. takanoi dominated C. maneas significantly (d 2 , p < 0.01).
In fact, of all the sites sampled along the Opal coast in spring 2008, this site had the highest density of H. takanoi (Table 2).

Discussion
Hemigrapsus takanoi and H. sanguineus are presumed to have been introduced along the European coast by larvae released from ballast water (Noël et al. 1997;Gollasch 1999;Breton et al. 2002). By April-May 2008, respectively two and three years after the first observations in the Opal coast, both species are now common in all this area. At the five sites where crab densities were estimated (i.e., BHP, WF, GF, ODH and DH), the sediment particle size distribution was analyzed by dry-sieving the sediment through a stack of Wentworth-grade sieves using the Buchanan technique. Percentages were calculated for each of five sediment types, defined according to grain size: gravel (> 2000 µm), coarse sand (500-2000 µm), medium sand (200-500 µm), fine sand (63-200 µm) and silt/clay (< 63 µm) (Table 4). Both Hemigrapsus habitats showed a bimodal distribution, with medium sand and gravel for H. takanoi and fine sand and gravel for H. sanguineus. The proportion of fine sand was higher at WF than at ODH, and the proportion of silt-clay was higher at GF than at any other site, particularly DH. There is a segregation of the habitats colonised by the two non-indigenous species: H. takanoi tends to occupy low hydrodynamic habitats with mud, while H. sanguineus tends towards high hydrodynamic habitats with fine and medium sands. Usually, both species can live in sympatry in the habitats favoured by H. takanoi, though at Grand Fort Philippe (Aa estuary), no H. sanguineus were sampled with H. takanoi.

Hemigrapsus takanoi: habitat and distribution
Hemigrapsus takanoi was found in low hydrodynamic zones, such as the Aa estuary and the harbours at Boulogne-sur-mer and Dunkirk, where abundant populations existed under the boulders covering the soft-bottom composed of gravel, medium sand and mud. In Calais harbour, only two specimens were found, but the area sampled (lower eulittoral) had few boulders and thus the habitat appeared unfavourable for crabs. Class-sizes in colonised sites were similar, probably indicating the synchronous colonisation of this non-indigenous species in favourable habitats along the Opal coast. The sex-ratio (≈ 1) was similar to the one reported by Noël et al. (1997) for the French Atlantic coast, but the largest specimens found along the Opal coast were smaller than those found on the French Atlantic coast: 25 mm vs. 28 mm for the males and 20 mm vs. 23 mm for the females respectively. In the Netherlands, the biggest males of H. takanoi are found below tide marks, at least in some localities (C. d'Udekem d'Acoz, personal communication).
Since its first recorded sighting at La Rochelle on French Atlantic coast in 1994 (Noël et al. 1997), the species has been reported along the coast of the Bay of Biscay from Laredo in Spain and north of the Bay of Biscay (Loire Atlantique and Morbihan (Noël et al. 1997;Noël and Gruet 2008). Breton et al. (2002) reported its presence in Le Havre harbour during 1999. In August 2008, one adult male was sampled at Saint-Vaast-La-Houge in the Bay of Seine (Dauvin 2009). Apart from the Opal coast, which is the southern limit of its presence in the North Sea ( Figure 5), H. takanoi is now common along the coasts of Belgium (Dumoulin 2004 (Gollasch 1999) and later in Lower Saxony at Norddeich (Obert et al. 2007) ( Figure 5). For the time being, this species has not been reported from the English coast and in France from the south of Brittany to the western part of the Bay of Seine, in spite of favourable habitats in harbours, such as Brest and Saint-Malo, and in many small estuaries around the Brittany. However H. takanoi is present along approximately 1,000 km of coastline, ranging from the western part of the Bay of Seine in France to Lower Saxony in Germany.
Along the Atlantic coast, the maximum density of H. takanoi in the 1990s was 20 ind.m -2 (Noël et al. 1997), and 5 ind.m -2 in the Osterscheldt, Netherlands during September 2000 (Nijland 2000). In spring 2008, only two years after its first observation in Dunkirk harbour its density reached 60 ind.m -2 , five times higher than the density of its congeneric species H. sanguineus (maximum 12 ind.m -2 ). This shows the high colonisation potential of this nonindigenous species, which can produce up to 50,000 eggs three to four times during the spawning season (McDermott 1998a). Since no ovigerous female was found during the spring sampling period, the spawning period must start later in the summer, as with H. sanguineus. Spawning period begins in May in the Bay of Biscay (Noël et al. 1997).

Hemigrapsus sanguineus: habitat and distribution
Hemigrapsus sanguineus was found in abundance on the rocky shores of the Opal coast from the Alprech Cape to the Gris Nez Cape ( Figure 3) under the boulders on soft-bottoms composed mainly of fine sand and gravel. The entire eulittoral zone appeared to be colonised, but during the spring sampling period (April-May), the mid-eulittoral was the preferred habitat of this Asian shore crab. This is in agreement with the observations of C. d'Udekem d'Acoz (personal communication) in the Eastern Scheldt, the Netherlands. In that area, H. sanguineus was found rather high on the shore and not on the lowest part of the shore. On the other hand, H. takanoi is spread on most on the shore and is also found below tide marks. However the high limestone cliffs such as the Cape Blanc Nez with its flint boulders and mobile coarse sand due to high hydrodynamics, was unfavourable for this non-indigenous crabs; only four H. sanguineus specimens were found under the 90 boulders at Cape Blanc Nez. Outside the rocky section of the Opal coast (Figure 3), boulders were only present along the estuary dykes, at Authie, Canche and Aa, and the three main harbours at Boulogne-sur-mer, Calais and Dunkirk. Although such habitats appear to be favourable for H. sanguineus, only one individual was found at Berk on the Authie estuary to the south of Boulogne-sur-mer. None were found at the mouth of the Canche estuary, probably due to the low salinity in the zone sampled.
The spring distribution along the Opal coast, with the highest abundances occurring in the mid-eulittoral, was comparable to those reported from along the North American coast, but the species was also observed in the subtidal zone during winter period (Kraemer et al. 2007). Hemigrapsus sanguineus can survive in various biotopes including artificial structures, mussel beds and oyster reefs; however it is usually found under the shelter of rocks, shells and other debris on tidal flats (McDermott 1991(McDermott , 1998aAhl and Moss 1999;Williams and McDermott 1990). Furthermore, H. sanguineus is able to tolerate a wide range of salinities and temperatures, as well as the damp conditions in the upper intertidal regions (Benson 2005). In Tanaba Bay, the recruitment of juveniles was evident in the upper and mid-eulittoral, with the larger individuals migrating to the lower eulittoral (Lohrer et al. 2000).
Class sizes were similar in the colonised sites of the Opal coast. Nevertheless, the maximal CW observed (26 mm for the males and 27 mm for the females) were smaller than those reported along the Atlantic coast of the United States, where the sizes reached 43 mm for the males and 36 mm for the females (McDermott 1998a), and smaller than those observed in the Cotentin (French side of the English Channel) during the summer 2008 (34 mm for the males and 31 mm for the females) (Dauvin 2009). Figure 6 shows the current known distribution of H. sanguineus for the northeastern European coast. They are present over approximately 1,100 km, including the coasts of the western part of the Channel (Dauvin 2009), the Bay of Seine (Breton et al. 2002;P. Hacquebart personal communication), the eastern part of the Channel (this study and F. Durand, personal communication), Belgium (Breton et al. 2002;Kerckhof et al. 2007), the Netherlands (Nijland and Beckman 2000;d'Udekem d'Acoz and Faasse 2002;Faasse 2004;Campbell and Nijland 2004;Nijland and Faasse 2005;Nuyttens et al. 2006;d'Udekem d'Acoz 2006) and Germany (Obert et al. 2007). This species has been recorded in the Mediterranean Sea only once, in August 2001 in the northern Adriatic (Schubart 2003).
Hemigrapsus sanguineus was observed for the first time in 1988 at Cape May, New Jersey, United States (Williams and McDermott 1990). By 2008, it had colonised the entire temperate zone from North Carolina to Maine (McDermott 1998b(McDermott , 2000Delaney et al. 2008). The colonised habitats are varied littoral zones, estuaries and/or mudflats that have rocks, boulders and shells (McDermott 1991(McDermott , 1998bAhl and Moss 1999;Williams and McDermott 1990;Gerard et al. 1999;Jensen et al. 2002;Brousseau et al. 2003;Kraemer et al. 2007;Delaney et al. 2008). The maximum densities on the United States coast were those reported by McDermott (1998b) for the site 'Townsends and Hereford Inlets' in New Jersey, where the density was 320 ind.m -2 , and by Kraemer et al. (2007) for a site in western Long Island Sound in 2001-2002, where the density was 350 ind.m -2 . Other densities over 100 ind.m -2 have been reported along this coast: 150 ind.m -2 on Long Island (Brousseau et al. 2003) and 190 ind.m -2 at Demerest Lloyd State Park in Massachusetts (Jensen et al. 2002). Still, for 52 sites along a 725 km coastal transect from New Jersey to Maine, Delaney et al. (2008) reported a lower density in August 2005, with a maximum of about 44 ind.m -2 in the south of the transect and a total absence of this species in the northern part of transect as far as 43°48.5'N latitude. It is possible that, after a phase of intense colonisation, the density declined. In the western part of Long Island Sound estuary, Kraemer et al. (2007) recorded a reduction of the mean density from about 120 ind.m -2 in the period 1998-2001 to 80 ind.m -2 in the period 2002-2005. This probably indicates that the Opal coast population, with its maximal density of 12 ind.m -2 , is just beginning its colonisation and that the maximum of density will be reached only in several years. Similar abundances, around 10, were also observed by Dauvin during summer 2008 in northern Cotentin (English Channel) (Dauvin 2009). Like H. takanoi, H. sanguineus shows a high colonisation potential. This species can produce up to 40,000 eggs several times during the spawning season (McDermott 1998a). No ovigerous females were found during the spring period. The spawning period apparently starts later in the summer: 72.5 % of females found in northern Cotentin in the summer of 2008 in temperature conditions similar to the Opal coast were ovigerous. The sex-ratio (1.31) was similar that the one reported by McDermott (1998b) for the Atlantic coast of the United States (1.23), with the number of males higher the number of the females.

Competition with Carcinus maenas
When H. sanguineus occur in high densities, this species can play an important role in restructuring the prey communities in intertidal habitats because they have the potential to affect populations of native species such as crab, fish, and shellfish by disrupting the food web (Brousseau et al. 2001). Specifically, this Asian species occupies habitats similar to the common shore crab Carcinus maenas, but along the American coast, it also competes with larger species, like the blue crab Callinectes sapidus Rathbun, 1896 and the rock crab Cancer irroratus Say, 1817 (Gerard et al. 1999;Tyrell and Hariis 1999;Jensen et al. 2002). Carcinus maenas has been reported to reduce feeding when H. sanguineus is present (Griffen andByers 2006). Nevertheless, MacDonald et al. (2007) has shown that C. maenas is better than H. sanguineus at obtaining food, probably because it is the fastest at finding and consuming the food, as has been shown experimentally by De Graff and Tyrell (2004).
According to Jensen et al. (2002), C. maenas has become uncommon under rocks in some areas from the east coast of North America since the arrival of H. sanguineus. These authors demonstrated that the number of C. maenas juveniles under rocks is drastically reduced in the presence of H. sanguineus, compared to areas when the two species do not overlap. C. d'Udekem d'Acoz (personal communication) has observed a drastic reduction in the number of juvenile C. maenas in some Dutch shore with a very high density of H. takanoi. Only 20% of C. maenas juveniles in the study of Jensen et al. (2002) being found under rocks in areas occupied by Hemigrapsus. Similarly, in their New Jersey to Maine transect, Delaney et al. (2008) observed the absence or low density of C. maenas in the areas colonised by H. sanguineus, and conversely higher abundances of C. maenas where H. sanguineus was absent. A similar pattern was observed for the Opal Coast, with a dominance of C. maenas south of Boulogne-sur-mer in an area weakly colonised by both non-indigenous Hemigrapsus species. For the moment, no competition has been found between the two non-indigenous species. But in the Dunkirk harbour, H. takanoi dominated the native shore crab C. maenas significantly; this site had the highest density of crabs in the spring of 2008 (60 ind.m -2 ).

Future distribution and possible impact of Hemigrapsus species along the European coasts
The reproductive output of H. sanguineus is important: mature females may have 3-4 clutches per breeding season, with a mean clutch size of 15,000 and a maximum of at least 40,000-50,000 eggs per crab, but the number of these eggs which hatched and the larval survival in the filed remain unknown (McDermott 1998b;Gerard et al. 1999). Planktonic larval stages, which last for about a month under optimal temperature and salinity conditions (Epifanio et al. 1998), vary in relation to the sea temperature, from 16 days at 25°C to 55 days at 15°C, thus providing an efficient mechanism for dispersal. This long larval dispersal phase could be favourable to the propagation of the species in the Channel and the North Sea under a megatidal regime that ensures efficient larval dispersal (Ellien et al. 2000).
For both non-indigenous Hemigrapsus species, it is hypothesized that they have probably dispersed from the Netherlands and Belgium populations which were being introduced with ballast water (Gollasch 1999). This road of colonization from the North Sea to the eastern part of the English Channel was well illustrated for the introductive razor clam Ensis directus (Conrad, 1843). Since the first specimens were found in 1979 near the mouth of the river Elbe in the German Bight, its extension in the North-European waters was rapid (Dauvin et al. 2007). Its progression in the eastern Channel appears to move in the opposite direction of the residual tidal transport and its distribution reaches nowadays the eastern Bay of Seine (Dauvin et al. 2007). The late colonization of Berk site which is the southwards and the most distant from Belgium reinforces this hypothesis on North Sea origin of the Hemigrapsus species along the Opal Coast.
In Japanese waters, the growth and maturation of these Hemigrapsus is rapid. Newly settled juveniles have a mean carapace width around 2 mm but reach 20 mm in about two years (Fukui 1988). The crabs become reproductively mature at this age and, although growth is slower in mature crabs, they can reach a maximum carapace width of 40 mm, corresponding to a maximum lifespan of around eight years. This old age was probably exceptional and more of the adult specimen could be life more than 2 years (C. d'Udekemn d'Acoz, personal communication). Ledesma and O'Connor (2001) and McDermott (1998a) have suggested that the length of the reproductive period of H. sanguineus is related to latitude and therefore to water temperature. In southern parts of Japan, the breeding season is 8 months long (Fukui 1988), whereas in northern Japan, it lasts three months (Takahashi et al. 1985in McDermott 1998a. Although McDermott (1998a) reported that breeding occurred through September at Gooseberry in Buzzards Bay, ovigerous females were found only until early August in Sandwich in Cape Cod Bay. In New Jersey, ovigerous females were observed from June to August, with some of them producing at least two broods during the summer. In New Jersey, the smallest female was 12.1 mm and the largest was 35.8 mm. It is probable that this smallest female was only one year old, given that during the July-August 2008 observations in northern Cotentin (the French coast of the English Channel), the smallest female was 13 mm and the largest was 31 mm, whereas the largest male was 34 mm (Dauvin 2009).
In the Western Pacific, the distribution of H. sanguineus ranges from ~20° to 50°N latitude, including the coasts of Hong Kong, Taiwan and Japan and the Pacific coast of China and Korea (McDermott 1998a). On the eastern coast of the United States, this species has been present from North Carolina to Maine ~ 34° to 43°N (McDermott 1998b;Delaney et al. 2008). On the European coast, the latitude ranges today from ~ 49° to 54°N, but the northern limit can extend northwards to ~ 60-65°N due to the warmer temperatures along the north-eastern European coast in relation to the North Atlantic drift. In the future, H. sanguineus may also extend its European distribution southwards to north-eastern Africa, thus invading the entire Mediterranean, where it could be in competition with a common native Grapsinae crab Pachygrapsus marmoratus (J.C. Fabricius 1787).
It is also possible that the second species H. takanoi will increase its latitudinal distribution considerably in the European waters, as H. sanguineus will likely do to the south and northwards ~ 60-65°N in relation to the North Atlantic drift. But the preferred harbour habitats of H. takanoi have poorly diversified benthic communities due to the confined environment and the organic and metallic pollutions that eliminate sensitive species. This could well be the best location to examine its competition with C. maenas, which appears to be abundant in the muddy sediment under boulders.
In the future it may be necessary to survey: i) the geographical extension of both Hemigrapsus species along the European coast, especially in the UK; ii) the increase of the established populations along the Opal Coast, particularly H. sanguineus, which is known to produce chemical cues that promote gregarious settlement and encourage rapid population increases in colonised areas O'Connor 2007); iii) the competition with intertidal crabs, not only with C. maenas but also with those with an intertidal juvenile phase that have commercial value, such as Cancer pagurus and Necora puber (L., 1767); iv) the possible consumption of Hemigrapsus by top level predators; i.e. Kim and O'Connor (2007) found that H. sanguineus megalops can be consumed by the killifish, Fundulus majalis (Walbaum, 1792); and v) the potential effect of predation on wild intertidal Mytilus edulis (L., 1758) populations and the Mytilus edulis and Crassostrea gigas cultivated in shellfish farms. Shellfish production is highly developed along the French Atlantic coast and is essential for the coastal economy. In fact, if H. sanguineus migrated to deeper waters during the winter, it could present a threat to these mussel and oyster farms. But due to the low temperature, its metabolism will slow down and the predation will be reduced.