A new Cryptorchestia species, Cryptorchestia ruffoi Latella & Vonk, sp. n. from the island of Rhodes in south-eastern Greece, can be distinguished on the basis of morphological and phylogenetic data. Morphological analysis and DNA sequencing of mitochondrial and nuclear protein-coding genes indicated that this species is related to C. cavimana (Cyprus) and C. garbinii (Mediterranean regions, with a recent northward expansion). Results supported a genetic separation between the Cryptorchestia species of the east Mediterranean regions and those of the northeast Atlantic volcanic islands examined in this study (C. canariensis, C. gomeri, C. guancha, and C. stocki from the Canary islands, C. monticola from Madeira, and C. chevreuxi from the Azores). The Mediterranean and Atlantic Cryptorchestia species appear to be also morphologically distinct. Cryptorchestia ruffoi sp. n., C. cavimana, C. garbinii, and C. kosswigi (Turkish coast) clearly have a small lobe on the male gnathopod 1 merus. This character was the main diagnostic difference between Cryptorchestia (sensu Lowry, 2013) and Orchestia. However, among the six northeast Atlantic island Cryptorchestia species only C. stocki has a small lobe on the merus of gnathopod 1. Reduction or loss of the lobe in the Atlantic Island species cannot be ruled out; however, molecular phylogenetic analysis leads us to presume that this lobe independently evolved between the east Mediterranean Cryptorchestia species and C. stocki from Gran Canaria.

biogeography, freshwater, Greek islands, molecular phylogeny, taxonomy

The genus Cryptorchestia Lowry & Fanini, 2013 is partitioned from Orchestia Leach, 1814, and is associated with freshwater-soaked leaf litter (Lowry and Fanini 2013). Cryptorchestia species from the East Mediterranean region were found in riparian habitats (Ruffo et al. 2014; present study), while the species from the Canary Islands, Madeira, and Azores (North East Atlantic area) live in humid, evergreen broadleaf laurel forest (laurisilva) (Stock 1989; Stock and Boxshall 1989; Ruffo 1990; Stock and Abreu 1992; Villacorta et al. 2008).

Here Cryptorchestia ruffoi sp. n. is described from the island of Rhodes, Greece. Specimens were first collected during the military occupation of Rhodes by the Italian army in 1928: the geologist Angelo Pasa of the Museo Civico di Storia Naturale of Verona found two talitrid specimens in a spring on Monte Smith (Fig. 1). Sandro Ruffo subsequently identified the specimens as belonging to a probable new taxon with morphological resemblances to Cryptorchestia cavimana (Heller 1865) and stored them in the Amphipoda collection of the Verona Museum. In 2010, one of us (L.L.) organised an expedition to Rhodes to find more talitrid amphipods from freshwater springs and succeeded in collecting more than 30 specimens.

Figure 1. 

Occurrence of Cryptorchestia ruffoi sp. n. on Rhodes, Greece (MSNVR). In the north, near the town of Rhodes, the locality of 1928 on Monte Smith. The other place represents small streams in the vicinity of Psintos, Epta Piges spring and stream.

A morphological characterisation forms part of this study. In order to gain insight into its position in relation to other Cryptorchestia species of the Mediterranean and North East Atlantic areas, a phylogenetic analysis was performed on DNA sequences of both mitochondrial (mt) and nuclear gene fragments, cytochrome oxidase I (COI), and histone H3 (H3), respectively. In particular, focus was on Cryptorchestia ruffoi sp. n. from Rhodes, C. cavimana (Heller, 1865) from Cyprus, C. garbinii Ruffo, Tarocco and Latella, 2014, from mainland Europe, C. canariensis (Dahl, 1950), C. gomeri (Stock, 1989), C. guancha (Stock & Boxshall, 1989), and C. stocki (Ruffo, 1990) from the Canary Islands, C. monticola (Stock & Abreu, 1992) from Madeira, and C. chevreuxi (De Guerne, 1887) from Terceira, Azores.

Morphology. The specimens analysed were collected in two localities from Rhodes (Fig. 1) and stem from different years with a gap of 82 years between them. A total of 37 specimens was examined for the description and measurements of the new species. Thirty one specimens were preserved in 70% ethanol, two mounted in glass slides in Faure's medium, two mounted on stubs for scanning electronic microscope (SEM) photography, and two used for molecular analyses. SEM photographs were obtained with a Zeiss EVO 40 XVP Scanning Electronic Microscope at the MUSE-Science Museum of Trento. The photo of a male paratype (Fig. 2) was obtained with a stereo microscope Leica M 165c, mounted with a Leica DFC450 camera at the Museo Civico di Storia Naturale of Verona. Type material is deposited in the Museo Civico di Storia Naturale of Verona (MSNVR), Verona, Italy, and the Naturalis Biodiversity Center (RMNH), Leiden, The Netherlands.

Figure 2. 

Cryptorchestia ruffoi sp. n., paratype, male (MSNVR). Lateral view.

PCR amplification and DNA sequencing. Specimens of Cryptorchestia ruffoi sp. n. from Rhodes, C. cavimana from Cyprus, C. garbinii from Europe and Macedonia, C. canariensis, C. gomeri, C. guancha, and C. stocki from the Canary Islands, C. monticola from Madeira, and C. chevreuxi from Terceira, Azores were stored in ethanol (Table 1). Platorchestia platensis from Capri island, Italy, was included in this study as the outgroup species (Table 1).

Genomic DNA was extracted from pereopods or whole organisms using QIAamp DNA Mini kit (QIAGEN). A PCR product of ca. 400 base pairs (bp) was amplified from the gene encoding the mt COI (some of our samples were old museum specimens in which mitochondrial DNA was degraded and consequently produced only short DNA sequences). The PCR-mediated reaction was performed using the primers BI-COI and SUBIR cited in Davolos and Maclean (2005). The PCR amplification conditions were 2 min at 95 °C, followed by 35 cycles, each consisting of 10–15s at 95 °C, 15s at 48–50 °C and 10s at 72 °C; the final PCR extension step lasted 10 min at 72 °C. The amplified fragments were checked by electrophoresis in 1% agarose gels and then used as templates for cycle sequencing reactions (BigDye chemistry) followed by DNA sequencing (ABI Prism 3130 capillary sequencer) using BI-COI and SUBIR primers. In addition, a fragment of ca. 350 bp of the gene encoding the nuclear histone H3, was PCR amplified using the primers H3Of and H3Or cited in Davolos and Pietrangeli (2014). The PCR cycling parameters were 2 min at 95 °C, followed by 35 cycles, each consisting of 10s at 95 °C, 10s at 48–52 °C and 5s at 72 °C; the final extension step lasted 10 min at 72 °C. The PCR products were verified and then sequenced using H3Of and H3Or primers, as above described.

Bioinformatic analysis. The nucleotides obtained in this study and the amino acid residues inferred were compared with sequence data accessible in the GenBank databases at the National Center for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov) using the BLASTN algorithm. The nucleotide sequence alignments were made in ClustalX (1.8) using the default parameters. Evolutionary analyses for the combined mt COI and nuclear histone H3 gene sequences were conducted in MEGA7 (Kumar et al. 2016). The evolutionary history was inferred by using the Maximum Likelihood method based on the General Time Reversible model (Nei and Kumar 2000). The tree with the highest log likelihood (-2646.1008) was used. Initial trees for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value (Fig. 10). A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 0.1808)). For Maximum Likelihood analysis, bootstrap resampling was performed with 1000 replications. The novel annotated sequences from the COI and the histone H3 genes from Cryptorchestia species of this study have been submitted to the GenBank (NCBI) database (Table 1).

The major objective of this molecular study, based on DNA sequences of the mt COI and nuclear histone H3 gene regions, was to estimate the evolutionary relationships of Cryptorchestia ruffoi sp. n. in relation to other Cryptorchestia species of the Mediterranean and North East Atlantic areas. DNA sequences from the mt region between the COI and COII genes were also analysed; all the species analysed here showed the peculiar rearrangement (data not shown), originally reported in Davolos and Maclean (2005). The phylogenetic scenario based on a Maximum Likelihood method suggested major diversification events, with evolutionary relationships between species generally well supported (Fig. 10). It is possible to recognise a well-supported group including C. ruffoi , C. cavimana, and C. garbinii. This monophyletic group points to a common origin of this Cryptorchestia lineage that currently appears to be limited to the East Mediterranean basin. Probably C. garbinii has recently colonised Europe, Macedonia (present study) and other regions by a northward expansion (Ruffo et al. 2014). The unique genetic Rhodean Cryptorchestia lineage agrees with the species rank of morphological differentiation identified in this study (see taxonomic section for C. ruffoi sp. n.). Another outcome of the analyses presented is the presence of a clade that contained the Cryptorchestia species (sensu Lowry, 2013) of the North East Atlantic area examined in this study: the two closely related species endemic to Gran Canaria: C. canariensis, and C. stocki (a within-island speciation appears the most-parsimonious hypothesis), C. monticola from Madeira, C. gomeri from La Gomera, C. guancha from Tenerife, and C. chevreuxi from Terceira (Fig. 10).

Figure 10. 

Molecular phylogeny by Maximum Likelihood method obtained in a combined analysis using mitochondrial cytochrome oxidase I (COI) gene region (363 bp), and H3 histone (H3) gene fragment (330 bp) sequences (a total of 693 positions in the final dataset) from Cryptorchestia ruffoi sp. n. and other Cryptorchestia species reported in Table 1. Platorchestia platensis was used in this study as outgroup species. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Nodes that have bootstrap values greater than 0.5 are labelled. The GenBank accession numbers of the newly determined sequences from the COI and the histone H3 genes used in this study are reported in Table 1.

Our results support the proposal by Lowry and Fanini (2013) in that the former Orchestia cavimana belongs to a new genus (Cryptorchestia). However, the North East Atlantic terrestrial talitrid species, formerly ascribed to Orchestia, apparently cannot be included within this new genus (Fig. 10). It is important to bear in mind that C. ruffoi sp. n., C. cavimana, C. garbinii, and C. kosswigi from the east Mediterranean regions have a small lobe (probably used in rasping or scrubbing) on the male gnathopod 1 merus as well as on the carpus and propodus (see Ruffo 1949; Ruffo et al. 2014; present study). This character has been proposed to be the main diagnostic difference between Cryptorchestia (the type species being O. cavimana) and Orchestia (in the latter there is a palmate lobe only on male carpus and propodus of gnathopod 1). However, among the northeast Atlantic island terrestrial Cryptorchestia species apparently only C. stocki (endemic to Gran Canaria as well its closely related species C. canariensis) has a small lobe on the merus of gnathopod 1 (Dahl 1950; Stock 1989; Stock and Boxshall 1989; Ruffo 1990; Stock and Abreu 1992). Although previous studies did not reveal that small lobe in C. canariensis, a better study of its growth stages could provide further data. The present observations are in line with our DNA sequencing findings that clearly showed a genetic separation of the North-East Atlantic and the Mediterranean Cryptorchestia species (Fig. 10). An alternative hypothesis that postulates reduction or loss of the small posterior palmate lobe cannot be disproved, however, our analysis identified similarity in this small structure present on the male gnathopod 1 merus as independently convergent within the east Mediterranean Cryptorchestia species and C. stocki. Therefore, the occurrence of the small lobe among Cryptorchestia species seems to have no clear evolutionary information regarding inclusion in the genus. Overall, the diversification process of Cryptorchestia species is of particular interest for future studies. We aim at investigating their evolutionary history by using a larger dataset and multiple calibrations in different parts of a Bayesian inferred phylogeny.

We thank C. Fišer, A. Latella, V. Lencioni, G. Luquet, P. Oromì, and M. Tarocco for their contribution in collecting some of the samples. We would also like to thank K. van Dorp (Naturalis Biodiversity Center, Leiden, The Netherlands) for her assistance in recovering museum samples of C. chevreuxi and C. monticola; R. Salmaso (Museo Civico di Storia Naturale, Verona, Italy) for her help in assembling the pictures and Nicola Angeli (Museo delle Scienze, Trento, Italy) for the SEM photographs. University students M. Severijns, J. Westerveld (Univ. Leiden), and J. Brugman (Univ. of Amsterdam) are thanked for their useful work in participating on a B.Sc. project of characterstate scores of the new species. The costs for the molecular analysis of this work were supported by Sapienza University of Rome, Rome, Italy. We thank L. Hughes and J. Lowry (Australian Museum, Sydney) for their constructive review comments.