Sphaerosyllis levantina sp. n. (Annelida) from the eastern Mediterranean, with notes on character variation in Sphaerosyllis hystrix Claparède, 1863

Abstract Examination of polychaete specimens from Haifa Bay (Israel, eastern Mediterranean Sea) revealed several individuals exhibiting morphological characteristics similar to Sphaerosyllis hystrix Claparède, 1863. A detailed morphometrical analysis of the Israeli specimens in comparison to specimens of Sphaerosyllis hystrix and Sphaerosyllis boeroi Musco, Çinar and Giangrande, 2005 supported the description of the former as a new species, Sphaerosyllis levantina sp. n. Individuals of Sphaerosyllis hystrix formed a very heterogeneous group with strong character variations in the analysis and the presumed cosmopolitan distribution of the species is discussed based on literature records.


Keywords
Polychaetes, Syllidae, Exogoninae, Sphaerosyllis, new species, Mediterranean, Cybertaxonomy, Scratchpads introduction The polychaete genus Sphaerosyllis Claparède, 1863 (Annelida) is one of the most species-rich genera of the syllid subfamily Exogoninae. At present, ca. 48 species are considered valid within Sphaerosyllis after the recent split of the group into the three genera Sphaerosyllis, Prosphaerosyllis and Erinaceusyllis (San Martín 2005). Up to date, 18 species of the genus have been recorded from the Mediterranean Sea , one of them described but yet unnamed (San Martín 2003), another one in the process of description (Del Pilar-Ruso and San Martín in press). In the framework of a project focusing on the soft bottom benthos of Haifa Bay (Israel, eastern Mediterranean Sea), a number of individuals of the genus Sphaerosyllis were found to exhibit morphological features which did not entirely correspond to any description of known Sphaerosyllis species, namely falcigers with a strong serration and with a subdistal spine present in all chaetigers. A subdistal spine on the blades of at least some falcigers has been described for the type species of the genus, S. hystrix Claparède 1863, and for S. boeroi Musco, Çinar and Giangrande, 2005. Re-examination of material of S. hystrix revealed that some individuals -contrary to descriptions available in the literature-possess a subdistal spine not only on the blades of the falcigers in anterior but also in posterior chaetigers. Consequently, this characteristic could not be used to unambiguously distinguish the Israeli material from S. hystrix. In order to clarify the relationship between the three very similar species possessing falcigers with a subdistal spine, a morphometric analysis has been performed, a method allowing not only to discriminate statistically significant groupings but also to identify taxonomically important characters (Costa-Paiva and Paiva 2007).

Specimen collection and processing
Specimens were collected on 11 Oct. 2009 in Haifa Bay, (Israel, Eastern Mediterranean Sea) from fine to medium sands in shallow waters (10 m). Sediment samples were taken with a Van-Veen grab (KAHLSICO, model WA265/SS214) 32×35 cm, volume 20 l, penetration 20 cm. The sediment was preserved in buffered formalin 10% for 3-7 days, then sieved through a 250 µm mesh sieve and subsequently stored in 70% ethanol. Specimens were examined under an Olympus SZx12 stereomicroscope and an Olympus BX50 microscope. Illustrations in pencil were made by means of a drawing tube, subsequently scanned, imported into a graphic program (GIMP), re-drawn and saved as a vector graphic. Three specimens selected for obtaining Scanning Electron Microscope (SEM) images were dehydrated, critical point dried (Bal-Tec CPD 030), sputter-coated with gold (Bal-Tec SCD 050) and examined under a JEOL JSM-6390LV at the Department of Biology, University of Crete. Specimens are deposited in the invertebrate collection of the Smithsonian National Museum of Natural History, Washington D.C., USA (USNM) and in the Tel Aviv University Zoological Museum, Israel (TAU).

Morphometric analyses
A total of 30 individuals belonging to three species (S. boeroi: 3 individuals; S. hystrix: 21 individuals; S. levantina sp. n.: 6 individuals) were analysed. Twenty-five variables were measured: I. body length, to account for size-dependencies of other characters; II. number of chaetigers; III. length of blade of dorsalmost falciger of a) anterior, b) midbody, c) posterior chaetigers; IV. length of blade of ventralmost falciger of a) anterior, b) midbody, c) posterior chaetigers; V. ratio of length of blades of dorsalmost to ventralmost falciger in a) anterior, b) midbody, c) posterior chaetigers; VI. ratio of length of blades of falcigers in anterior to posterior chaetigers for a) dorsalmost; b) ventralmost falciger; VII. Ratio of length of dorsalmost falciger to body length in in a) anterior, b) midbody, c) posterior chaetigers; VIII. Ratio of length of ventralmost falciger to body length in in a) anterior, b) midbody, c) posterior chaetigers; IX. maximum length of serration of falcigerous blades in a) anterior, b) midbody, c) posterior chaetigers (smooth, finely serrated, strongly serrated); X. presence of a subdistal spine in dorsalmost falcigerous blades of in a) anterior, b) midbody, c) posterior chaetigers.
Body length was measured excluding antennae, anal cirri and palps. Falciger blade lengths were measured from point of insertion into shaft to distal tip. Falciger blade lengths could not always be measured on the same chaetiger in all animals if blades were broken. Instead, measurements were made in predefined body regions (anterior: first 1-5 chaetigers; posterior: last 5-7 chaetigers; midbody: in between). Three individuals of S. hystrix were excluded from the multivariate statistical analysis due to missing values for some characters.
Multidimensional Scaling (MDS) was subsequently employed to display the similarities of the different individuals. To test for significance of differences between species a PERMANOVA (Permutational Multivariate Analysis of Variance) was performed (Anderson 2001). A Principal Component Analysis (PCA) was used to determine variability of characters and to identify characters for the species differentiation. To determine the importance of the characters discriminating the species, the Principal Component Scores were correlated (Spearman's correlation coefficient) with the measured character values of each individual.
Multivariate statistical analyses were performed with PRIMER V6, correlation of the Principal Component Scores were calculated with the R package (R package version 2.10; http://www.R-project.org).

Sphaerosyllis levantina
Etymology. Derived from the type locality (Levantine Basin), levantina being a neo-Latin adjective meaning "pertaining to the region where the sun raises"; feminine declination in accordance with the genus name (Syllis was a river nymph in the greek mythology and thus female).
Habitat. Fine to medium sands. Taxonomic remarks. S. levantina sp. n. is similar to S. minima Hartmann-Schröder, 1960 in having blades of falcigers with strong serration throughout the body. However, S. minima has a stronger dorso-ventral gradation of the blades of falcigers (dorsal ones twice as long as ventral ones) than S. levantina sp. n. (dorsal ones 1.5 times longer than ventral ones) and the ventral cirrus is longer than the parapodial lobe in S. minima, whereas is is half as long as the parapodial lobe in S. levantina sp. n. S. capensis Day, 1953, S. taylori Perkins, 1981, and S. sandrae Álvarez and San Martín, 2009 are similar to S. levantina sp. n. in the shape and serration of the blades of the falcigers, but S. capensis has all antennae positioned in line (median one posteriorly of lateral ones in S. levantina sp. n.), S. taylori shows no dorso-ventral gradation of the falciger blade length (dorsal blade 1.5 times longer than ventral one in S. levantina sp. n.) and  (Figs 2f, 3, 6). Furthermore, the blades of the dorsalmost falcigers show an anteroposterior gradation in length in S. hystrix (anteriorly 1.5 times longer than posteriorly), whereas they are of similar length throughout the body in S. levantina sp. n. (Figs 3, 6, see also tables in online supplementary material). Finally, S. hystrix has a very narrow pharynx (almost half the width of proventricle), whereas the pharynx of S. levantina sp. n. is wider than ¾ of the width of the proventricle. An identification key to the Mediterranean Sphaerosyllis species is provided at the end of this manuscript.
Ben-Eliahu (1977) discusses two different morphological forms of S. hystrix occurring in her samples from Israel. Based on her description and illustrations, the animal identified as S. hystrix sensu Westheide 1974 could potentially belong to S. levantina sp. n. because of the similar characters of falcigers and papillation. However, the description does not report the characteristic subdistal spine on the blades of the posterior falcigers. In addition, Westheide's (1974) description of S. hystrix from the Galápagos Islands differs from both Ben-Eliahu's specimen and the present material by the absence of parapodial glands (Westheide 1974), a character considered as variable and thus of no taxonomic value by Ben-Eliahu (1977) but recently accepted as a taxonomically stable character (Riser 1991).

Multivariate morphometrical analysis
The results of the Principal Component Analysis show that the first principal component (PC1) account for 77.4% of the variability, the second (PC2) for 16.4% and the remaining 3 PCs for 5.1% (eigenvector values available at http://polychaetes. marbigen.org/content/morphometric-analysis-pca-eigenvectors). The Spearman's correlation of the Principal Component scores with the measured character values of the individuals revealed that the length of the dorsalmost falcigerous blades in all body parts (anterior, midbody, posterior), as well as the ratio of the anterior to posterior ventralmost falcigerous blade are the most important characters discriminating between the three species (ρ-values >0.8 / <-0.8 at p < 0.005) (http://polychaetes. marbigen.org/content/spearmans-correlation-principal-component-scores-vs-measurements).
The PCA plot of the first two components show a discrimination of species into three groups, with individuals of S. levantina sp. n. having the lowest PC1 scores, S. boeroi the highest scores. S. levantina sp. n. and S. hystrix show similar PC2 scores, whereas S. boeroi shows lower scores, and, except for one small-sized individual, forms a distinct group apart from the remaining species. Individuals of S. levantina sp. n. likewise form a close group, however, a couple of individuals of S. hystrix cannot be distinguished from this cluster (Fig. 7). The MDS diagram gives similar results, with individuals of S. boeroi and S. levantina sp. n. forming distinct groups, whereas individuals of S. hystrix are spread as a heterogeneous group, with some of them being plotted close to individuals of either S. boeroi or S. levantina sp. n. (Fig. 8).
The PERMANOVA analysis results in a p-value of 0.001 as calculated by 999 permutations, thus the null-hypothesis (no differences between the groups) cannot be sustained. Subsequent analyses of the differences between species through pairwise tests reveals significant differences between species (S. hystrix / S. boeroi: p = 0.003, 713 permutations; S. hystrix / S. levantina sp. nov: p = 0.001, 995 permutations; S. boeroi / S. levantina sp. n.: p = 0.015, 84 permutations).

Discussion
The genus Sphaerosyllis -like many of the small-sized Exogoninae genera-has a difficult and often confused taxonomy and biogeography. Among the potential causes contributing to the current confusion the following could be cited: a) lack of detail in older (before ca. 1970) species descriptions; b) difficulties of observing certain characters in fixed material (Riser 1991); c) descriptions of new species without examination of comparative material; d) ongoing discussions on the taxonomic value of characters such as the presence or absence of dorsal cirri on the second chaetiger (Fauvel 1923, San Martín 2005, presence and type of parapodial glands (Westheide 1974, Ben-Eliahu 1977, Riser 1991 and variations in chaetal structures (Riser 1991). These factors have lead to the assignment of individuals with very different character sets to the same species name and thus to wide-spread distribution records of some species. S. hystrix (type locality Normandy, France) is included among those species with an alleged cosmopolitan distribution, since it has been recorded from most European coasts including the Mediterranean Sea, the north-western coasts of America (Berkeley andBerkeley 1948, Hartman 1968), the Galápagos Islands (Westheide 1974), China (Men et al. 1993, Ding andWestheide 2008), Australia (Hartmann-Schröder 1984, 1985 and the Western Atlantic Fauchald 1971, Temperini 1981), among others. However, recent studies suggest that the North American records of S. hystrix and S. pirifera Claparède, 1868 are in fact  individuals of S. californiensis Hartman, 1966 and that the two European species are not represented in the American Pacific fauna (Kudenov and Harris 1995). Similarly, some specimens from the Mediterranean Sea previously identified as S. hystrix had been re-examined and found to exhibit significant morphological differences to S. hystrix, leading to the establishment of a new species, S. boeroi . In the light of an ever-increasing number of molecular analyses revealing cryptic species complexes in morphologically indistinguishable polychaete species with an assumed cosmopolitan distribution (e.g. Westheide and Hass-Cordes 2001, Westheide and Schmidt 2003, Barroso et al. 2009, Bleidorn et al. 2006 it is likely that the various specimens recorded under the name S. hystrix may in fact form a complex of similar species, especially since many descriptions differ substantially from each other (see Ben-Eliahu 1977).
The morphometric analysis conducted in this study support the hypothesis of several morphologically very similar species co-existing in the Mediterranean. The individuals of S. levantina sp. n. and S. boeroi form distinct groups in the PCA and MDS plots, however the individuals of S. hystrix show a much wider spread, marginally overlapping with the other two species when only the meristic characters are taken into account. This is explained through a high character variability in the examined individuals, especially concerning the presence of a subdistal spine on the blades of the posterior falcigers and the length of the falciger blades. The presence of a subdistal spine on all dorsal falcigerous blades is invariable in S. boeroi and S. levantina sp. n., wheras individuals of S. hystrix with otherwise very similar chaetal structures might or might not possess such spine. Another feature that seems to be highly variable in S. hystrix is the length of the falciger blades in relation to body size. In fact, individuals of S. levantina sp. n. with short falciger blades are located at the lower end of the size spectrum of all measured blades, S. boeroi with almost spiniger-like blades at the higher end, whereas the blade lengths of the examined individuals of S. hystrix form a smooth transition between the other two species.
However, when tested by strict statistical criteria, the hypothesis of different coexisting species is significantly supported, and based on their meristic characters the species show significant differences. The results of the current study suggest that S. hystrix may well constitute a species complex. Given the difficult taxonomic status of the genus, similar results might be expected for other species as well, and consequently, distributions of several Sphaerosyllis species might be in fact questionable or unknown.

Key to the Mediterranean Sphaerosyllis species:
The three species S. claparedei Ehlers, 1864, S. papillifera Naville, 1933 andS. ovigera Langerhans, 1879 are poorly known. All have been described as having dorsal cirri on the second chaetiger, however, other species, such as S. hystrix, were also originally described or illustrated with dorsal cirri on the second chaetiger whereas they are in fact absent. Since the three aforementioned species are exclusively known from their original description (or partly reproductions of these) and have never been re-described based on new material, they are tentatively included in the key below, but their identity remains questionable.