Finding of the macrophagous deep-sea ascidian Dicopia antirrhinum Monniot, 1972 (Chordata: Tunicata) in the Tyrrhenian Sea and updating of its distribution

Abstract This study reports the in situ observations of a deep-sea ascidian, Dicopia antirrhinum (family Octacnemidae), on the deep seabed off the Aeolian Islands (Tyrrhenian Sea), between 569 and 813 m depth. These observations represent the first record of the species in Italian waters and the second observation in vivo to date. Peculiar macroscopic features are described here and the main differences from other species of the family Octacnemidae are highlighted, in order to allow the future identification of D. antirrhinum through visual technologies such as remotely operated vehicles. A total of 29 specimens were observed, with a maximum density of 0.012 specimens/m2. The global geographic and bathymetric distribution of D. antirrhinum known so far is also updated and discussed.


Introduction
The current knowledge about Mediterranean deepsea fauna is still scarce compared with the fauna in shallower waters, especially concerning the class Ascidiacea. Unlike their shallow-water counterparts, deep-sea ascidians are rather uncommonly found in the Mediterranean Sea (Mecho et al. 2014). Visual technologies, such as remotely operated vehicles (ROVs), provide the opportunity to find and observe deep-sea habitats and species. This is particularly true for fragile species that are destroyed by trawling or that live in irregular areas where trawls cannot be deployed, to update the knowledge about their distribution (e.g. Mastrototaro et al. 2016Mastrototaro et al. , 2017Chimienti et al. 2018c). Moreover, visual techniques can allow the study of population structure and dynamics in non-invasive ways (e.g. Ambroso et al. 2014;Chimienti et al. 2018b), as well as the observation of interesting or unknown behaviours of benthic invertebrates (e.g. Chimienti et al. 2018a).
However, the overall lack of sampling, which cannot always be carried out using an ROV, represents a bias for the taxonomic identification of some marine invertebrates, such as the ascidians (Sanamyan et al. 2018), as well as for their molecular analysis.
Within the family Octacnemidae, the peculiar features of the body such as the shape of the siphons with the hypertrophy of the oral lobes, as well as the presence of a peduncle (or stalk), can be used as macroscopic diagnostic characters to allow morphological identification even without the collection of samples. The family Octacnemidae includes 10 genera and 30 species of deep-sea ascidians. Among them, the genera Dicopia, Megalodicopia and Situla are characterized by an oral siphon expanded into muscular lobes, forming a lip-like aperture used to catch small motile prey such as crustaceans (Monniot & Monniot 1975a). This unique morphological feature, coupled with descriptions provided in the literature (e.g. Monniot & Monniot 1974, 1977aMecho et al. 2014), can help with the specific identification of some Octacnemidae species, such as Dicopia antirrhinum Monniot, 1972. Within this family, only D. antirrhinum and Situla lanosa Monniot and Monniot, 1973 have been found so far in the Mediterranean Sea, the former supported by scientific literature (e.g. Mecho et al. 2014 and reference therein), the latter reported only in Monniot and Monniot (1990) without any reference. The present study reports the observation of various specimens of D. antirrhinum in the Aeolian archipelago (Tyrrhenian Sea), with a detailed description of the external morphological appearance of the species. Moreover, an update about the currently known distribution of D. antirrhinum is given and discussed. The behaviour of the specimens observed in the Aeolian archipelago is described and a preliminary estimate of the density of this species is reported.

Material and methods
ROV surveys were carried out on board the R/V Oceana Ranger from May to June 2018. Observations were made using a SAAB Seaeye Falcon DR ROV equipped with an HDV (high-definition video) camera of 480 TVL (television lines), with minimum scene illumination of 2.0 lux (F1.4), a ½″ CCD (chargecoupled device) pick-up device, image sensor, spherical ½ of 3.8 mm and wide-angle lenses. The ROV also included a depth sensor, sonar, a compass for underwater navigation, and two laser beams for the measurement of the area surveyed and of the observed specimens. ROV transects were carried out with an average speed of 0.2-0.3 knots and a constant frame of 1.80 m of width. The position was continuously recorded using a LinkQuest Tracklink USBL Transponder with up to 0.25°accuracy.
Five different sites around the Aeolian Islands were investigated, namely Eolo Seamount, Filicudi Bank, Capo Shoal, and areas off Lipari and off Stromboli ( Figure 1). A total of 51 ROV transects were carried out in the study area, covering a linear distance of 52,316 m and a total area of 94,169 m 2 , from 30 to 990 m depth. A total of 76 h 11 min video of the seabed was recorded. In seven of the ROV transects, D. antirrhinum was observed (Table I; Figure 1), and its density was estimated considering the number of specimens and the area surveyed within each ROV transect.
Detailed footage was taken of each specimen observed to attain a complete and precise external view of their morphology. The size of each specimen (i.e. length without the stalk, considering the specimens with the oral lips open) was estimated from the video frames using metric references based on the laser beams.

Results
A total of 29 specimens of D. antirrhinum were observed on the seabed off the Aeolian Islands: two at Eolo Seamount, two at Filicudi Bank, 18 off Capo Shoal, two off Lipari and five off Stromboli (Table I; Figure 1). Sampling by means of the ROV manipulator arm was not possible due to the soft consistency of the specimens.

Description
The body was supported by a thick peduncle inserted in the substrate (Figure 2a, e-f). The oral aperture consisted of two enlarged and lengthened lips that formed a sort of horseshoe. The two lips were almost the same length, with the lower one slightly longer than the upper one ( Figure 2). The distal ends of both lips appeared smooth (Figure 2a-d). The tunic was also quite smooth, except for some small digitations visible on its surface (Figure 2), likely to be papillae, that give it a velvet appearance. The narrow atrial siphon was placed dorsally (Figure 2f). The size of some observed specimens was estimated at between 3.3 and 3.7 cm height.
Among the 29 specimens observed, 18 had their lips open suggesting a feeding behaviour, and six of them were observed closing their lips and performing a rapid contraction of their bodies to considerably reduce their size in case of stress (e.g. the presence of the ROV) (Figure 2g-j; Supplemental material). This contraction took from 2 to 15 seconds, with a mean of 6.17 ± 5.38 seconds. In particular, the first step of the contraction was often slow, until contact was made between the two lips; then it was followed by a rapid contraction of the body, which seemed to collapse on the peduncle. In contrast, 10 other specimens already had their lips closed once approached by the ROV, while one of them opened its lips when illuminated by the lights of the ROV.

Remarks
The detailed observation of specific macroscopic features is essential to avoid the misidentification of D. antirrhinum with respect to similar species. In particular, within the family Octacnemidae, described by Herdman (1888), only three genera are characterized by the presence of an oral siphon developed into two lips: Megalodicopia Oka, 1918, Situla Vinogradova, 1969and Dicopia Sluiter, 1905. The first genus includes species characterized by a wide body attached to the substratum through a very long, thick and muscular peduncle (Oka 1918;Monniot 1989;Sanamyan 1998;Sanamyan & Sanamyan 2002), unlike the genus Dicopia.
The two genera Situla and Dicopia are very similar, and can be distinguished by the shape of the branchial sac, which is conical in Dicopia and like a flat branchial lamina in Situla (Brunetti & Mastrototaro 2017). Unfortunately, for this type of diagnosis it is necessary to sample the animals and to properly anesthetize them in order to observe the branchial sac. However, specimens belonging to Dicopia can also be distinguished from Situla on the basis of macroscopic features, such as the proportions of the upper and lower lips, or the shape of the peduncle. In particular, six species have been described to date within the genus Situla: S. pelliculosa Vinogradova, 1969, S. lanosa, S. rebainsi Vinogradova, 1975, S. macdonaldi Monniot and Monniot, 1977b, S. galeata Monniot, 1991 andS. cuculli Monniot andMonniot, 1991. Situla pelliculosa, the type species of the genus, does not have the ability to close its mouth, having a weak radial musculature (Vinogradova 1969;Sanamyan & Sanamyan 2002), resulting in an easily recognizable feature for differentiating S. pelliculosa from species that actively close their oral aperture, such as D. antirrhinum (Figure 2g-j). Situla lanosa appears very similar to D. antirrhinum (Brunetti & Mastrototaro 2017), but the two species can be distinguished due to the different proportions of the lips. In particular, the upper lip of D. antirrhinum is almost comparable in size to the lower one, while in S. lanosa the upper lip is larger than the lower one ( Figure 3). Moreover, the tunic, the distal end of the lips and the peduncle are characterized by numerous filaments that give S. lanosa a hairy appearance (Figure 3c), hence the name lanosa (woolly) (Monniot & Monniot 1973). Also, in S. galeata and S. cuculli the lower lip appears smaller than the upper one (Monniot & Monniot 1991), while S. rebainsi is characterized by a body which resembles a flat, oval disc, attached to the substratum by a very long peduncle (Vinogradova 1975;Sanamyan & Sanamyan 2002). Lastly, S. macdonaldi is represented by only one specimen collected off the Kerguelen Islands (Indian Ocean) which, unlike D. antirrhinum, had a muscular peduncle similar to those of the genus Megalodicopia (Monniot & Monniot 1977b).
Within the genus Dicopia, three species are currently known: D. antirrhinum, D. fimbriata Sluiter, 1905 andD. japonica Oka, 1913. Dicopia fimbriata was described for the first time from Indonesia (Sluiter 1905), and four specimens were then sampled off New Caledonia (Monniot & Monniot 1991), one in the north Tasman Sea (Sanamyan & Sanamyan 1999) and, recently, one in Papua New Guinea (Monniot & López-Legentil 2017). This species is characterized by large flat lips and a small oral aperture similar to a fissure, with a smooth vitreous tunic featuring a smooth and not rolled margin of the oral lobes, as well as several papillae in the inner part of the tentacular velum (Monniot & Monniot 1991;Monniot & López-Legentil 2017). As recently highlighted by Monniot and López-Legentil (2017), D. fimbriata can be distinguished from D. antirrinum because the latter has a tunic covered with papillae, giving it a velvet aspect, and a different organization of the musculature. Finally, D. japonica was described by Oka (1913), and three specimens were sampled by Millar (1988) off the coast of northern Peru. According to Monniot (1972), this species can be distinguished from D. antirrhinum mainly on the basis of its highly reduced peduncle, and the oral aperture consisting of two flatter and rectangularshaped oral lobes.

Ecology
Specimens are usually found on the deep muddy seabed, directly attached to small rocks or pebbles covered by mud (Monniot 1972), or fixed to the hardened mud, due to the compacting ability of their anchoring rhizoids (Mecho et al. 2014). All the specimens found in the Aeolian Islands were on muddy bottoms.

Distribution
The macrophagous ascidian D. antirrhinum is considered a deep-sea species, characterized by a wide bathymetric range, from 500 to 4300 m depth (Monniot & Monniot 1975b, 1977a. This species shows a North Atlantic distribution, having been recorded mainly along the Atlantic European continental slope (Monniot & Monniot 1990). It was first collected in the Gulf of Biscay between 600 and 1130 m depth, with only one specimen sampled (Monniot 1972).
Later, 22 specimens (17 juveniles and five adults) were collected in the same area at 2100-2200 m depth (Monniot & Monniot 1974), and 95 more specimens from 1845 to 4300 m depth (Monniot & Monniot 1977a). Moreover, 50 specimens of D. antirrhinum were sampled in 1976 along the Rockall Trough (West of Ireland), at about 2500 m depth (Monniot & Monniot 1985), and two specimens in 1988 southwest of the Gibraltar Strait, between 1200 and 2000 m depth (Monniot & Monniot 1988). In 1975, seven juveniles of Dicopia sp. were collected in the Mediterranean Sea, off Malta (Sicily Channel), at 500 m depth (Monniot & Monniot 1975b). More recently, in 2011, five adults were observed in situ for the first time and one of them was sampled by Mecho et al. (2014) in La Fonera Canyon (Balearic Sea), at 1072-1106 m depth, using a ROV (Figure 4).
The 29 specimens observed in this study were widespread along the five ROV transects, at depths ranging from 569 to 813 m. The maximum observed density was 0.012 specimens/m 2 , at Capo Shoal ( Figure 1).

Discussion
Considering the macroscopic features of the species, and based on the comparison with previous observations (Mecho et al. 2014), it was possible to assign all the specimens observed off the Aeolian Islands to D. antirrhinum. Although the morphological analysis of anesthetized samples is often necessary for the correct identification of ascidians, macroscopic diagnostic characters such as body shape and the peculiar lip-like siphons permit the identification of D. antirrhinum from high-quality in vivo images, with a high level of reliability.
The size of the specimens observed (ca. 3.5 cm in height) is comparable to that measured by Mecho et al. (2014) in the Balearic Sea (about 6 cm in height), considering that those authors measured the whole specimen collected, including the part of the rhizoids below the sediment, which is not visible with ROV imaging.
The lack of observations of this species to date could be due to the difficulty in recognizing it from most deep-sea samples, which are mainly collected using sledges and dredges that damage the specimens, making them unclassifiable. For this reason, it is possible that the presence and distribution of D. antirrhinum have been underestimated or not detected with common sampling tools. Due to this difficulty in finding samples for molecular analysis, molecular characterization to describe the species is also missing. Our findings add one more tile to the still incomplete mosaic of distribution of D. antirrhinum in the Mediterranean Sea, extending its known presence to the Tyrrhenian Sea, where the species inhabits muddy bathyal bottoms between ca. 500 and 900 m depth. These new data confirm that the species is less rare than was considered in the past. Future prospects include the possibility to sample some specimens for barcoding purposes, as a complementary taxonomy tool for the morphological and visual analysis. Deepsea visual surveys are providing continuous new findings and information about these still poorly known environments. In particular, such surveys are primarily improving knowledge about soft-bodied species (e.g. gelatinous), which are difficult to sample and identify with traditional tools as trawling, grabs and dredges.