Morphology of the larval stages of a Mediterranean population of the allochthonous Say ’ s mud crab , Dyspanopeus sayi ( Decapoda : Brachyura : Panopeidae )

In this study the complete larval development (including four zoeal stages and a megalopa stage) of Dyspanopeus sayi is described and illustrated in detail based on larvae reared in the laboratory. Ovigerous females were collected in Alfacs Bay, Ebro Delta (NW Mediterranean) where the species was recently introduced. These larval stages were compared with others collected in the plankton from the same area, the larval stages described previously for this species and the larval stages of the two other Panopeidae that inhabit the Iberian Peninsula: Panopeus africanus and Rhithropanopeus harrisii. Differences were found in both zoeal and megalopal larval characters between D. sayi and the other two species, which could raise doubt about their placement in the same family.


INTRODUCTION
The native distribution of Say's mud crab, Dyspanopeus sayi (Smith 1869), encompasses the Atlantic coast of North America from Florida to Canada (Nizinski 2003).It is considered an invasive species in other parts of the world as a result of human activities and has probably been accidentally transported in ballast water like many other marine invertebrates (Davidson and Simkanin 2012).Currently, outside of its native range, D. sayi has been recorded from southwest England, Queens Dock, Swansea (Wales) (Ingle 1980, Clark 1986), on the French and Dutch coasts of the North Sea (Vaz et al. 2007), and more recently in the Black Sea (Micu et al. 2010).It is also present in the Mediterranean Sea, where it has been collected in Venice, the Marano and Varano lagoons, and in the Po River Delta (western Adriatic Sea) (Froglia and Speranza 1993, Mizzan 1995, Florio et al. 2008).The most recent record is from the western Mediterranean, which constituted the first record for the coast of the Iberian Peninsula (Schubart et al. 2012).Figure 1 shows the current distribution of Dyspanopeus sayi worldwide.
Dyspanopeus sayi is a euryhaline and eurythermic species that inhabits estuaries and shallow coastal marine waters (see Schubart et al. 2012).This species is the second Panopeidae that has established large populations in estuarine habitats of the Iberian Peninsula.The first species was Rhithropanopeus harrisii (Gould 1841), reported for the first time in the Guadalquivir River (south Atlantic coast of Spain) (Cuesta et al. 1991).Another panopeid inhabiting the Iberian Peninsula is Panopeus africanus A. Milne-Edwards, 1867, which is an endemic crab with a wide distribution from Angola to Portugal (Rodríguez and Paula 1993).
In the present study the complete larval development (four zoeal stages and a megalopa) of Dyspanopeus sayi is described and illustrated in detail based on larvae reared in the laboratory.We compared these larval stages with those collected in the plankton, and the larval stages described previously for the other two Panopeidae that inhabit the Iberian Peninsula, namely P. africanus and R. harrisii.

MATERIALS AND METHODS
Three ovigerous females of Dyspanopeus sayi of 16 to 18 mm cephalothorax length were collected by beam trawls from sandy-muddy bottoms of Alfacs Bay, Ebro Delta (40°40'N, 0°40'E) covered by the seagrass Cymodocea nodosa and the alga Caulerpa prolifera (see Pérez andCamp 1986, Fusté 1988), in August 2011.Specimens were transported to the laboratories of the IRTA (Institut de Recerca i Tecnologia Agroalimentàries) in Sant Carles de la Ràpita.The females were kept in 40-L aquariums.After hatching, actively swimming zoea I were transferred to 500-mL beakers (n=5).A total of 20 larvae were placed in each beaker.The subsequent rearing cultures were conducted at a constant salinity of 34, a temperature of 18±1°C and a natural photoperiod of ca. 12 h light per day (early spring condition).After the water had been changed, the Chlorophyceae Tetraselmis chuii and Artemia sp.naupii were provided daily as feed.
A qualitative plankton survey was carried out on 24 and 25 September 2012 in the Alfacs Bay, Ebro Delta using a plankton net with a mouth opening of 0.25 m 2 and mesh size of 250 µm.Samples were taken during daytime, fixed in ethanol (96%), and later sorted in the laboratory under a Wild MZ6 compound microscope.
For easier observation of larvae structures and setation under microscope, a digestion-stain procedure (Marco-Herrero et al. 2012) was carried out.Initially, entire specimens were placed for 10 minutes in a watch glass with 2 ml of heated lactic acid.Immediately after, 3 drops of Clorazol Black stain (0.4 g Clorazol Black powder dissolved in 75 ml 70% ethanol) were added to the heated solution.After 5-10 minutes, the specimen was removed from the solution and placed on a slide with lactic acid, in order to proceed with the dissection of the appendages.Drawings and measurements were made using a Wild MZ6 and Zeiss compound microscope with Nomarski interference, both equipped with a camera lucida.All measurements were made by an ocular micrometer.Descriptions and measurements of different larval stages were based on at least 10 specimens of each stage from culture larvae.In Figure 7 first and second maxilliped of zoea I are drawn without exopod, and plumose natatory setae of the maxillipeds exopods of zoeae II-IV are drawn truncated.Description and figures are arranged according to the standards proposed by Clark et al. (1998).
Measurements taken for the zoeal stages were: rostrodorsal length (RDL), distance from the tip of the rostral spine to the tip of the dorsal spine; cephalothorax length (CL) between eyes (base of the rostrum) to the posterolateral carapace margin; cephalothorax width (CW) as the distance between the tips of lateral spines; rostral spine length (RL) from base of eye to tip of rostral spine; and antennal length (AL) from base of eye to tip of the protopod.Measurements for the megalopa included: CL measured from the frontal to posterior margin of carapace; CW as the carapace maximum width.
A larval series has been deposited at the Biological Collections of Reference of the Institut de Ciències del Mar (ICM-CSIC) in Barcelona under accession numbers ICMD13022501-3.

RESULTS
In the plankton samples a total of 9 zoeae I, 2 zoeae II and 1 zoea IV of Dyspanopeus sayi were identified and used for morphological and meristical comparison with those reared in the laboratory.
The larval development of Dyspanopeus sayi consists of four zoeal stages and a megalopa.At 18±1°C and 34 salinity the zoeal development was completed in a minimum of 15 days (appearance of the megalopa).The duration and survival of each larval stage is shown in Figure 2. The first zoeal stage is redescribed (see Schubart et al. 2012) and only the main differences in subsequent stages are described.
Cephalothorax (Fig. 3A, G): dorsal spine straight and well developed with small tubercles over the surface; rostral spine straight and slightly longer than dorsal spine; ventral caparace margin without setae, 1 pair of posterodorsal simple setae; eyes sessile.
Antenna (Fig. 5E): protopod long, equal in length to rostral spine, with rounded tip and without spines; exopod reduced to a minute bud with 1 small terminal simple seta; endopod absent.
Pleopods: all present as undifferentiated buds.

DISCUSSION
There is a well-established population of Dyspanopeus sayi in the area of Alfacs Bay, Ebro Delta, NW Mediterranean, as evidenced by the collection of mature males and ovigerous females since 2005 (see Schubart et al. 2012;Guerao pers. obs.) and by the occurrence of larval stages in the plankton of the bay in September 2012.The presence of zoea I, zoea II and zoea IV in these samples from the inner part of the bay suggests that this species carries out its complete larval development in the bay, independently of the possibility of larval dispersal offshore.The scattered populations of this species in the Mediterranean compared with the continuous distribution long the French and Dutch coastline (see Fig. 1) is due to the low number and location of appropriate habitats in the Mediterranean, as there are no estuaries with a gradient of mixed salinities.The most similar habitats are lagoons, deltas, and closer bays, which are ecosystems with brackish waters; however, these habitats are not abundant in the Mediterranean basin.This scarcity of available habitats also explains the low number of populations of the other panopeid Rhithropanopeus harrisii introduced into the Mediterranean, in contrast with its wider distribution on the European Atlantic coasts (Projecto-García et al. 2010).
The zoeal stages collected in the Alfacs Bay plankton were measured and dissected.No significant differences were found in size, morphology or setation with respect to the cultured material.This is important information for examining morphological descriptions of larvae because meristic data obtained from specimens reared in the laboratory can be considered as valid, regardless of possible anomalies due to culture conditions (González-Gordillo and Rodríguez 2000, Wehrtmann and Albornoz 2003, Marco-Herrero et al. 2012).
According to previously published data, the duration of zoeal development of D. sayi ranges from 14 days at 21°C to 27 days at 14°C (Chamberlain 1957).Kurata (1970) reported complete zoeal development in a minimum of 15 days (April culture) and 14 days (May culture) but did not give data on temperature.Data obtained in the present study, 15 days at 18°C, fall within the range mentioned above and corroborate the well-known relationship between temperature and duration of decapod larval development (Anger 2001).
Previous descriptions of larval stages of D. sayi, except the zoea I described by Clark (2007) and Schubart et al. (2012), are incomplete, brief or inaccurate.Illustrations are also incomplete and in some cases are of low quality.The present study provides, for the first time, data on cephalothorax setation for zoea II-IV, in addition to information on the right setation pattern of mouthparts.Setation is described for the megalopa stage and illustrations of the sternum are provided for the first time, in addition to illustrations of the ischial spine of the cheliped, spinulation and the dactyli of the pereiopods.The following are some of the noteworthy differences: a fourth pair of inner serrulated setae was observed on the telson margin in zoea IV in the present study rather than in zoea III as reported by Chamberlain (1961); and the short "feeler" on the dactylus of pereio-pod 5 described by Kurata (1970) was not observed in the present study.Other differences in the setation pattern (fewer setae) could be due to miscounts.
Panopeidae in its present composition (see Ng et al. 2008) is a complex family with high heterogeneity between larval forms (Martin 1988).In some cases, this variability in larval morphology has been used as evidence to support taxonomic changes in the family, such as the recent establishment of the genus Acantholobulus Felder and Martin 2003, for some species of Panopeus H. Milne-Edwards, 1843.The larval stages of D. sayi fit in Group I of the classification of the Xanthid larvae by Martin (1984), which also includes R. harrisii and Panopeus africanus; however, it should be pointed out that other panopeid species have been placed in other groups of this classification based on larval morphology.This may suggest that this family needs to be studied further, taking into account other evidence based on adult morphology, and using techniques such as DNA analysis combined with larval morphology.Although they are all in Martin's Group I, the zoeal stages of the three Iberian panopeids show clear differences that allow easy identification.They also show differences in the megalopa stage, a larval phase which is not considered in the classification by Martin (1984).The main differences observed in the zoeal stages are observed in the ratio between the dorsal and rostral spine lengths, as well as in the antennal and pleonal morphology (see Table 1).Zoeal stages of Panopeus africanus have well-developed dorsal and rostral spines that are similar in size in zoea I; however, the dorsal spine is longer than the rostral spine in the other three zoeal stages.In R. harrisii and D. sayi, both cephalothoracic spines are well developed, but the rostral spine is longer than the dorsal spine, and in R. harrisii it is more than twice as long as the dorsal spine.The antenna of the zoeae of the three species are characterized by a reduced exopod, although it is less reduced in Panopeus africanus, which even has a fourth zoeal stage that shows two terminal setae instead of one, as in the rest of the zoeal stages and in the other two species.In all cases the protopod is as long as the rostral spine, but the tip is acute in P. africanus and R. harrisii and rounded in D. sayi with no spinulation.Panopeus africanus, however, has strong spines increasing in size towards the tip in zoea I and with fewer spinules in subsequent stages; in R. harrisii there are only minute spinules in the distal part.Differences were also observed in the pleon morphology, such as the number of dorsolateral processes, which were only present in pleonite 2 in R. harrisii and in pleonites 2 and 3 in P. africanus and D. sayi.This is one of the features characterizing Group I; therefore, R. harrisii must be considered an exception with respect to this character.Other differences in pleonal features are shown in Table 1.
The megalopa stage also shows clear differences between the three species (see Table 1).The rostrum is similar in P. africanus and D. sayi, although the spines at the basal angles, called "horns" in some papers, are more developed and acute in P. africanus.The megalopa of R. harrisii does not have these spines on the rostrum.The chelipeds are also similar in P. africanus and D. sayi but the strongly curved ischial spine is not present in R. harrisii (in some cases there is a small spine but it is never curved).The main difference in the megalopa stage is the number of segments of the antennal flagellum: there are eight in P. africanus and D. sayi, and six in R. harrisii.Other differences are seen in the mandibular palp, sternum, the dactyli of pereiopods, and uropods, and are shown in Table 1.These larval differences allow larvae of the three panopeids inhabiting Iberian Peninsula waters to be accurately identified; however, the fact that species be-table 1. -Main morphological and meristic differences between larval stages of Dyspanopeus sayi, Panopeus africanus and Rhithropanopeus harrisii.Abbreviations: dsl, cephalothorax dorsal spine length; rsl, cephalothorax rostral spine length; sp., spines; plp, posterolateral process; seg., number of segment; s., setation; sbls, subterminal long setae.