First record of Longosomatidae (Annelida: Polychaeta) from Iceland with a worldwide review of diagnostic characters of the family

The family Longosomatidae Hartman, 1944 (Annelida: Polychaeta) is newly reported for Icelandic waters. The diversity, taxonomy and ecology of this poorly-known monogeneric family are reviewed based on material collected during the BIOICE project. Two species of the genus Heterospio are recorded and redescribed: Heterospio longissima Ehlers, 1874 sensu Hartman (1965) and Heterospio reducta Laubier, Picard and Ramos, 1972–73. Several body structures of high taxonomic relevance in the genus are reviewed and discussed.


Introduction
Longosomatidae Hartman, 1944(formerly Heterospionidae Hartman, 1963) is a poorly-known family of benthic polychaetes that is rather fragile and easily overlooked in benthic fauna surveys. The body is characterized by an anterior region composed of seven to nine segments with three to eight pairs of elongate branchial filaments usually called "thorax", a "mid-body" region composed of elongate segments and a bulbous "posterior region" with terminal pygidium. The prostomium and peristomium are usually described as reduced and have a pair of nuchal organs and a pair of grooved palps, respectively. Chaetae occur in dense fascicles on the anterior segments and normally form complete rings on elongated segments; they include capillaries, thickened aristate and subuluncini-like capillaries and spines (Borowski 1994;Fauchald and Rouse 1997;Wilson 2000a;Rouse 2001;Hernández-Alcántara and Solís-Weiss 2009).
The family comprises a single genus, Heterospio Ehlers, 1874;Borowski (1994) recognized six species and, recently, Bochert and Zettler (2009) have described a new one. Species have been reported from widely scattered areas, mainly in deep shelf and slope sediments (Fauchald 1977;Rouse 2001). The phylogenetic position of Longosomatidae is still questioned; Rouse and Fauchald (1997) and Blake and Arnofsky (1999) placed it in the clade Spionida, but other authors like Wilson (2000a) consider it more similar to Cirratulidae, in the clade Terebellida. Three species have been described in North Atlantic waters ( Figure 1): Heterospio longissima Ehlers, 1874 (type species of the genus) from west Ireland, Heterospio mediterranea Laubier, Picard andRamos, 1972-73 andHeterospio reducta Laubier, Picard andRamos, 1972-73 both from the Mediterranean Sea. Furthermore, Hartman (1965) and Uebelacker (1984) reported and described H.longissima and H. cf. longissima, respectively, from New England and the Gulf of Mexico, and Uebelacker (1984) reported Heterospio sp. A also from the Gulf of Mexico. All these reports probably represent new taxa.
The BIOICE (Benthic Invertebrates of Icelandic Waters) project was conceived with the aim of improving knowledge on benthic biodiversity in the waters around Iceland, by means of a number of oceanographic cruises. The sampling area covered a depth range from 20 to 3500 m on both sides of the Greenland-Iceland-Faroe Islands Ridge, above which the waters are generally less than 500 m deep. Examination of polychaete material collected during the BIOICE expeditions has led to several papers on polychaete taxonomy and ecology, including descriptions of a number of new taxa (Kirkegaard 2001;Sanfilippo 2001;Sigvaldadóttir 2002;Chambers and Woodham 2003;Parapar 2003;2006;Parapar et al. 2011a;2011b;Moreira and Parapar 2012). During the study of the samples collected by the BIOICE project, two species of longosomatids were also found. In this paper, both species are redescribed using scanning electron microscopy (SEM) and a preliminary review of key diagnostic characters in the genus is provided to improve our scarce knowledge about this family.

Material and methods
This study is based on material collected by several oceanographic cruises within the BIOICE Project. Specimens were fixed in a 4% formaldehyde-seawater solution, and preserved in 70% ethanol. Animals were sorted from samples by the staff at the Sandgerdi Marine Centre, and then examined by the authors. The studied material was deposited in the collections of the Icelandic Institute of Natural History, Reykjavik (IINH) and in the Museo Nacional de Ciencias Naturales de Madrid, Spain (MNCN). Data on the BIOICE samples with Heterospio specimens are presented in Table 1. For SEM, the specimens were dehydrated in a graded ethanol series, prepared by critical-point drying using CO 2 , mounted on aluminium stubs, covered with gold in a BAL-TEC SCD 004 evaporator, and examined and photographed under a JEOL JSM-6400 scanning electron microscope at the Servizos de Apoio á Investigación, University of A Coruña, Spain.
Chaetigers progressively longer from CH9 onwards. CH9 first elongated segment, longer than wide; length (as distance from chaetal bundle to chaetal cincture of CH10) about three times longer than CH8. Branchiae lacking in all specimens but eight pairs of branchial scars present from CH2 to CH9 (Figures 2A, B, 5B). CH1 abranchiate (no scars). The remaining chaetigers strongly elongated and cylindrical in cross-section; length increasing backwards; CH10 about four times longer than CH9, CH11 about 2.5 times longer than CH10. CH1-9 with biramous parapodia; with notopodial and neuropodial chaetal fascicles well separated. From CH10 onwards parapodia as elongated ridges forming a nearly closed flange-like cincture near anterior margin of segment ( Figure 4C). Aristate or acicular spines not observed.

Distribution
Heterospio longissima was described by Ehlers (1874) based on an incomplete specimen from the northeast Atlantic. Hartman (1965) reported the species from the western Atlantic. However, Laubier et al. (1972-73) pointed out that Hartman's material was different from that of Ehlers, but further comparisons to test whether the latter corresponded to a different species were not possible because the holotype was apparently missing (Borowski 1994, p.130). Therefore, Laubier et al. (1972-73) cautiously considered the existence of two forms of this species: the nominal species and H. longissima sensu Hartman (1965). Since then, most subsequent reports of longosomatids were assigned to either form of H. longissima (e.g. Imajima 1974;Intès and Le Loeuff 1977;Kirkegaard 1980;Amoureux 1982;Rosenfeldt 1989; see Figure  8), but the character combinations of these specimens did not always match those of the two forms (cf. Hartman 1974;Uebelacker 1984). In view of the aforementioned worldwide reports attributed to the species, it is likely that several different species are still waiting to be described, particularly all those reported outside the North Atlantic Ocean (e.g., Hartman 1974;Imajima 1974;Intes and Le Loeuff 1977;Rosenfeldt 1989).

Remarks
Both Heterospio longissima sensu Hartman (1965) and the nominal species have in common that the first elongated segment is CH9, but the degree of elongation is different. In H. longissima sensu Hartman CH9 is about three times longer than previous chaetigers, whereas in H. longissima sensu Ehlers CH9 is about as long as all the anterior segments together (see Laubier et al. 1972-73, fig. 3; Borowski 1994,   Figure 5D), H. peruana has only four pairs ( Figure 6C). Although aristate chaetae as illustrated by Bochert and Zettler (2009) were not observed, the shaft of the subuluncini seems to be articulated with the blade ( Figure 4C-H) in a similar manner to that illustrated by Wu and Chen (1966); Borowski (1994) and Wilson (2000a) (see Discussion).

Material examined
Thirty-one specimens were collected in 10 BIOICE samples (

Distribution
Originally described from off Algiers by Laubier et al. (1972-73), the species was later reported by Amoureux (1982) from the continental slope (500-1400 m) off west Ireland.

Remarks
Heterospio reducta is distinguished from most longosomatids by the smaller number of short anterior chaetigers (CH1-CH6), with CH7 being the first elongated chaetiger, while in other species the latter is either CH8, CH9 or CH10. Laubier et al. (1972-73), while describing and illustrating CH7 slightly longer than CH6, consider CH8 as the first posterior chaetiger; this is followed by Bochert and Zettler (2009) in their key of worldwide species of Heterospio. As no chaetal cinctures are present in this species, Laubier et al. (1972-73) used the segment size as evidence of the beginning of the mid-body region, instead of the shape of the chaetal bundles. Hence, in H. reducta the first mid-body chaetiger would actually be CH7 rather than CH8 (see above description and Figure 6B). The only substantial difference between Icelandic specimens of H. reducta and those from the type locality is the presence of very long chaetae in the notopodia of CH8, which are located more dorsally ( Figures  9A, 10A). Heterospio reducta differs from specimens of H. longissima sensu Hartman, 1965 also found in Icelandic waters in the number of short anterior chaetigers (six in H. reducta and eight in H. longissima sensu Hartman) and by having only one type of chaeta; furthermore, in the latter species the chaetae form almost complete circles on elongated chaetigers, which are not observed in H. reducta.
Heterospio reducta differs from H. mediterranea from the Mediterranean Sea and Heterospio sp. A as described by Uebelacker (1984) from the Gulf of Mexico in the number of short anterior chaetigers (seven in these species and six in H. reducta) and in the number of branchiae and chaetal composition, respectively. Hence, H. mediterranea has seven pairs of branchiae while H. reducta has only three; Heterospio sp. A has acicular hooks in the neuropodia of CH1, whereas in H. reducta there are simple capillaries in both the notopodia and neuropodia.
The species most similar to H. reducta is Heterospio angolana Bochert and Zettler, 2009 from the southeast Atlantic Ocean (Angola). This species has the same number of pairs of branchiae (three) and short anterior chaetigers (six)-although Bochert and Zettler (2009) consider CH9 as first elongated chaetiger-but they differ in length of CH8, being more elongated in H. reducta than in H. angolana ( Figure 6B,D) and in the very long chaetae of H. reducta on CH8.

Discussion
Because the number of anterior segments ("thoracic region") and the relative length of the following elongated segments ("mid-body region") are two of the key characters in longosomatid taxonomy, and since segmental furrows are normally not evident in this family, we believe that it is necessary to establish the position of segmental limits to avoid confusion in their designation. According to our observations, parapodia are located at the anterior border of segments and therefore the length of a segment can be considered as the distance between these parapodia and the parapodia of the next chaetiger.
The transition between the thoracic and mid-body regions has not been clearly established in the literature; both the elongation of the segments (e.g. Laubier et al. 1972-73) and the shape of the parapodia (Wu and Chen 1966;Borowski 1994) have been used to indicate the onset of the mid-body region. However, the transition between the two presumed parts of the body is not conspicuous enough in all species. Hence, in most species apart from H. longissima Ehlers, the first elongated segment is just slightly longer than the previous ones, and therefore it is difficult to discern whether it belongs to the thorax or to the mid-body. Similarly, the change in shape and arrangement (biramous vs cincture) of the chaetae on these transition chaetigers is sometimes not clearly defined but shows an intermediate aspect (e.g. H. reducta). Therefore, we recommend not using the terms "thorax" and "midbody" in longosomatids to avoid further confusions when characterizing species.
Aristate chaetae have been reported in many polychaete families, such as Acoetidae Kinberg, 1856 (e.g. Pettibone 1989), Trochochaetidae Örsted, 1843 (e.g. Wilson 2000b), Cirratulidae Ryckholdt, 1851 (e.g. Glasby 2000), Lumbrineridae Schmarda, 1861 (e.g. Aguirrezabalaga and Carrera-Parra 2006), Spionidae Grube, 1850 (e.g. Hourdez et al. 2006), Pilargidae Saint-Joseph, 1899 (e.g. Darbyshire and Mackie 2003) and Flabelligeridae Saint-Joseph, 1894 (e.g. Salazar-Vallejo 2012). This term actually encompasses different types of chaetae (both in shape and origin), which only match a general shape of a "simple chaeta with smooth shaft and a tuft of hairs or a single spine" (Rouse and Pleijel 2001). This type of chaeta was first described in longosomatids by Wu and Chen (1966, p.166, text and table; "they resemble aristate setae") when describing those of H. sinica; this was also followed by Borowski (1994, p.135, pl. 2C) for H. peruana and Wilson (2000a, p.194, pl. 1.105) for one still undescribed Heterospio species. The observation under the SEM of the chaetae of Icelandic H. longissima sensu Hartman revealed that the ones here named "subuluncini" closely resemble those that Wu and Chen (1966) and Wilson (2000a) named "aristate chaetae" or the ones illustrated by Borowski (1994) as "transitional subuluncini capillary-like". We also noticed that the distal ends of these chaetae are deciduous, resulting in chaetae very similar to the "pointed distally" chaetae of Bochert and Zettler (2009). However, the chaeta illustrated by Borowski (1994) as "aristate" seems to be different from any other chaetae described so far in Heterospio. Therefore, we think that confusion regarding the nomenclature of chaetae in Heterospio exists, making necessary a redefinition of terms like "subuluncini", "spine" and "aristate", because it is possible that all of them are simply stages of transition of the same type of chaetae, in the sense already discussed by Borowski (1994).
The presence of palps in the genus Heterospio has previously been questioned. For instance, Hartman (1965, p.163) recognized that even though no palps were observed attached to individuals of H. longissima "several palpi are present in the sample, which are believed to belong to this species; their bases fit well in the scars of attachment at the posterior margin of the prostomium" and she had also illustrated them as such (Hartman 1965, fig. 30f). Laubier et al. (1972-73, p. 245) presented an ambiguous description: "En vue latérale, des organes nucaux sont visibles á la limite postérieure. Les palpes son toujours absents," which is not clarified in either the drawings or in the discussion. Later, Uebelacker (1984, pp. 10.1) provided a different opinion, interpreting the deep grooves behind prostomium as nuchal organs: "Neither palps nor scars of their attachment were observed on Gulf of Mexico specimens, although distinct, usually everted nuchal organs were readily apparent". In contrast, Wu and Chen (1966) actually described and illustrated the palps attached to the animal (Wu and Chen 1966, p. 166, pl IA,B). Borowski (1994) reviewed the family, in confirming the presence of these structures: "Nuchal organs sac-like, in deep grooves between prostomium and peristomium … Peristomium asetigerous, with deep transverse grooves; tentacles in all specimens lost." However, Uebelacker's (1984) view was recently revived by Bochert and Zettler (2009, p.736), who also denied their presence: "Nuchal grooves posteriolateral to prostomium. Tentacles or tentacle scars in all specimens absent." The study of our specimens showed that the only structure observed under the SEM is a deep groove between prostomium and peristomium (never everted in our specimens), which does not look like a tentacle scar as much as a nuchal organ.