Porifera collection of the Italian National Antarctic Museum (MNA), with an updated checklist from Terra Nova Bay (Ross Sea)

Abstract This new dataset presents occurrence data for Porifera collected in the Ross Sea, mainly in the Terra Nova Bay area, and curated at the Italian National Antarctic Museum (MNA, section of Genoa). Specimens were collected in 331 different sampling stations at depths ranging from 17 to 1,100 meters in the framework of 17 different Italian Antarctic expeditions funded by the Italian National Antarctic Research Program (PNRA). A total of 807 specimens, belonging to 144 morphospecies (i.e., 95 taxa identified at species level and 49 classified at least at the genus level) is included in the dataset. Nearly half (45%) of the species reported here correspond to species already known for Terra Nova Bay. Out of the remaining 55% previously unknown records, under a third (~29%) were classified at the species level, while over a quarter (~26%) were ascribed to the genus level only and these would require further study. All vouchers are permanently curated at the MNA and are available for study to the scientific community. A 3D model of an uncommon species from the Ross Sea, i.e. Tethyopsisbrondstedi (Burton, 1929), is also presented and will be made available for outreach purposes.


Purpose
Since the very beginning of the Italian expeditions in the Ross Sea, which started in 1985, sponges have been one of the most studied taxa, due both to the high number of species found in the Terra Nova Bay area (where Italy has its coastal station "Mario Zucchelli", 74°41'42"S, 164°7'23"E), and to the strong community of Italian taxonomists specialized in Antarctic sponges. This sponge collection has been progressively growing each year, with new collections of specimens at each expedition of the Italian National Antarctic Research Program (PNRA) in the Terra Nova Bay area. These specimens have been studied and exchanged among different researchers for comparisons and publications for years, they represented the base of a PhD thesis (Sarà 2002), and then the whole collection was finally acquired by the Italian National Antarctic Museum (MNA, section of Genoa, Italy) in 2010. Since then the sponge collection has been restored, crosschecked for distributional data with the original labels, matched with a collection of permanent glass slides of spicules and updated in terms of taxonomy or new identifications of specimens.
This study aims at publishing and valorising occurrence data of the Porifera collected during several scientific expeditions of the Italian National Antarctic Program (PNRA) in the Ross Sea.
This collection is amongst the largest for Antarctic sponges and despite it being mainly focused on the Terra Nova Bay area, it represents a unicum given the amount of permanent glass slides with spicules available and the large database of images of sponges documented in situ, or freshly collected.
This distributional dataset is the fifth MNA contribution to the Antarctic Biodiversity Portal (www.biodiversity.aq), which is the thematic Antarctic node for both the Ocean Biogeographic Information System (AntOBIS) and the Global Biodiversity Information Facility (ANTABIF), based on materials stored at the MNA. The previous contributions were: Ghiglione et al. (2013), Piazza et al. (2014), Selbmann et al. (2015) and Cecchetto et al. (2017).

Project description
Project title: Antarctic Porifera in the collection of the Italian National Antarctic Museum (MNA) Curator and Promoter: Stefano Schiaparelli

Study area description:
The specimens were collected in the Ross Sea sector of the Southern Ocean in a bathymetric range from 17 to 1,100 meters of depth (Fig. 1).
Design description: Data were assembled by revising all the distributional records of the specimens deposited in the collections of the Italian Antarctic National Museum (MNA, section of Genoa, Italy). The samples were collected in the framework of several expeditions of the Italian National Antarctic Research Program (PNRA) from 1987 to 2014.

Method step description:
See sampling description below and flowchart of Fig. 2.
Study extent description: The distributional data considered here originated from 331 different sampling stations ranging between 17 and 1,100 metres of depth (Figs 1,3,4,5).
Sampling description: Sampling was performed on a total of 331 different sampling stations (Figs 1,3,4,5) through the deployment of a variety of sampling gears, mainly dredges (Charcot dredge, Naturalist dredge, Triangular dredge and Picard dredge) and Van Veen grabs of different volumes. Some samples were also opportunistically collected by long fishing lines, mid water trawls (that touched the bottom due to a failure of the winches), trammel nets, and other fishing nets that provided additional material to standard techniques. Some samples, from the XIV, XV, XVII, XVIII, XXIII, and XXV PNRA expeditions, were hand-collected by SCUBA diving. In one case (i.e., Sycetta antarctica (Brøndsted, 1931), MNA 8847) the specimen was collected beached and sampling coordinates refer to the coastline. In another case, i.e., Lycopodina cf. vaceleti (van Soest and Baker 2011), the specimen record is based on a georeferenced ROV video frame (Fig. 6) and no physical samples are available.
Once the material has been acquired by the MNA after sorting and shipment activities, all the specimens were classified to the lowest possible taxonomical resolution. In the years, different researchers have contributed to the classification of the specimens: Homoscleromorpha were studied by Maurizio Pansini and Antonio Sarà; Demospongiae and Hexactinellida were studied by Rachel Downey, Antonio Sarà, Marco Bertolino, Maurizio Pansini and Barbara Calcinai; Calcarea were studied by Hans Tore Rapp. The present dataset has been formatted in order to fulfil the Darwin Core standard protocol required by the OBIS scheme (http://www.iobis.org/manual/lifewatchqc/) and according to the SCAR-MarBIN Data Toolkit (available at http://www.scarmarbin.be/documents/SM-FATv1.zip). The dataset was uploaded in the ANTOBIS database (the geospatial component of SCAR-MarBIN).
Vouchers are now preserved in 90% ethanol (~53% of the entire collection), frozen (~23%), or dried (~24%). The data flow chart illustrating the sampling, sorting, and storing procedures for specimens, data, and image availability is reported in Fig. 2.
Quality control description: Specimens were identified at the finest possible taxonomic resolution and only those that have been classified at least at the genus level were included in the present dataset. During all the phases of sorting, classification, and storage of samples at the MNA, quality control and data cleaning have been un-     dertaken at various stages in order to produce high quality data and make consistent cross-references between the database and sample labels. The MNA uses an SQL-based database (Specify 6) and a R-Shiny web application to manage its collections and link all the data (photos, glass slides, etc.) to the physical samples.
Due to the large amount of researchers that managed the material before the acquisition of the collection by the MNA, all the specimens and distributional records were rechecked and then imported in the museum database. During this phase it emerged that only ~75% of the Porifera collection fulfilled the expected minimum set of data fields to be included in GBIF. The remaining ~25% of the material present in the MNA collection can be divided in a ~7%, represented by material not yet classified, and another ~18% represented by old materials that cannot be ascribed to a specific sampling station due to missing labels or incomplete information about sampling.
Georeferencing on board each of the different research vessels is based on the interpolation of GPS satellite receivers and a gyrocompass. Station coordinates and sampling events were recorded during sampling activities based on various GPS systems.

Taxonomic coverage
General taxonomic coverage description: This dataset focuses on all classes (Calcarea, Demospongiae, Hexactinellida, and Homoscleromorpha) of the Phylum Porifera (Kingdom Animalia), and includes a total of 807 specimens belonging to 144 mor-phospecies (with 95 taxa classified at species level and 49 at genus level), and representing 12 orders and 30 families (Fig. 7). Nearly half (~45%; 65 species) of the collected taxa correspond to records already known for the Terra Nova Bay area Cerrano et al. 2000;Sarà et al. 1990Sarà et al. , 1992, about one third (~29%; 42 species) correspond to new records classified at specific level, and just over a quarter (~26%; 37 species) to new records classified at the genus level. The new records for Terra Nova Bay are reported with the acronym 'NR' immediately after the species name in the following taxonomic ranks section.
Permanent glass slides with the spicules are available for ~88% of the species of the MNA collection while SEM pictures are available for ~31% of the species (some of which are already available on the Antarctic Field Guide project at: http://afg.biodiversity.aq/pdfs/144164-a-field-guide-to-antarctic-sponges.pdf ).
In  , 2011), is without a MNA collection code, as it was identified from an ROV video frame (Fig. 6), and represents the first record of this carnivorous sponge at Terra Nova Bay and, globally, the second record of this species (Van Soest and Baker 2011). This species was identified thanks to Rob Van Soest and Claire Goodwin and represents an important "visual record" because only a few pictures of carnivorous sponges in situ are available. In the case of the species Tedania (Tedaniopsis) oxeata Topsent, 1916 (MNA 8244), the specimen record was obtained by using the ROV arms (Fig. 8).
The MNA collection also includes several sponge holotypes (Table 1). Some species (i.e. Haliclonissa verrucosa Burton, 1932     Virtual collection of vouchers and 3D models: The species used in the 3D model, Tethyopsis brondstedi (Burton, 1929) (MNA 2839, Fig. 9) presents a spherical body from which 2 fragile oscular tubes protrude from the upper part of the main sponge body, reminiscent of a bull head with horns. The long oscular tubes are believed to serve as both inhalant and exhalant orifices for the sponge (Hajdu et al. 1994). The surface of the sponge's main body is often found covered in small pebbles and other sandy debris, indicating that this species may live partially buried within the seabed sediment. The main sponge body is radial and is composed of oxeas (long spicules which are pointed at both ends) and triaenes (long spicules which are pointed at one end and the other is composed of three equal rays, reminiscent of a wind turbine). These triaenes can be bifurcated at each end of the ray (dichotriaene), the rays can be curved backwards (anatriaene), or are a triaene with only two rays, that are usually bifurcated at each end (orthodiaene). The oscular tubes of the sponge are composed of intricate layers of orthodiaenes. The main body of the sponge also contains tiny asters (star-shaped spicules, with rays radiating equally from a central point), which are composed of strongylasters (blunt-tipped or slightly bulbous ended rays) and oxyasters (with sharply pointed rays).
The model of the sponge was obtained through micro-CT imaging performed at the Department of Geosciences (University of Padova) by CM. A bench-top Skyscan 1172 micro-CT system (Bruker®), equipped with a Hamamatsu 100/250 microfocus X-ray Figure 9. Video of the 3D model of Tethyopsis brondstedi (Burton, 1929) (MNA 2839), a very uncommon sponge only occurring in the Ross Sea. The diameter of the main spherical body is ~10 mm and the length/width of the oscular tubes are respectively ~30 mm and ~5 mm. source (80 kV, 124 μA) and a Hamamatsu C9300 11 megapixel camera (with a pixel size of 8.68 μm) filtered by a 0.5 mm Aluminium foil was used. Projection images were acquired with 1200 ms exposure time, 2x2 binning mode, 0.30° rotation step over 360° rotation, averaged over 10 frames and in vertical random movement mode to minimise noise, providing an image pixel size of 13.2 μm. An oversized sample option was applied with 4 connected scans, leading to a total acquisition time of about 1170 min. Post-acquisition reconstruction was performed using the NRecon (Bruker microCT®) software package, starting from raw projection images, and applying thermal correction, misalignment compensation, ring artefact reduction and beam hardening correction. Segmentation was then performed with CT Analyser (Bruker microCT®) software package, using a 3D adaptive thresholding procedure (mean of minimum and maximum value) within spherical kernels of radius 8 pixels, starting from a pre-determined pre-thresholding value. Resulting images were saved as monochrome (1 bit) bitmaps and imported in the CT-Vox (Bruker microCT®) software package to perform 3D rendering and animations. The model will be available on the MNA web site (www.mna.it) and on Sketchfab (https:// sketchfab.com/MNA). The species chosen for the model corresponded to one of the few specimens collected in the Ross Sea area after the species description (Burton 1929).

Datasets
Dataset description: This dataset contains data about all four classes (Calcarea, Demospongiae, Hexactinellida and Homoscleromorpha) of the Phylum Porifera, based on vouchers from the Ross Sea (with a special focus on Terra Nova Bay) curated at the MNA. In total, the dataset includes 144 different morphospecies, and a total of 807 specimens. Several studies were based on this dataset: Alvizu et al. (in press); Bertolino et al. 2009;Calcinai et al. 2000;Cattaneo-Vietti et al. 2000;Fondi et al. 2014;Mangano et al. 2009;Orlandini et al. 2014;Papaleo et al. 2012Papaleo et al. , 2013Romoli et al. 2011;Sarà 2002;Sarà et al. 1992. The validity and synonyms of each species name were checked in WORMS (World Register of Marine Species; http://www.marinespecies.org; last check made on 2018-03-28). The Darwin Core elements included in the dataset are: ID, Institution code (i.e. the name of the institution where the samples are kept), basis of record, occurrence ID, catalogue number (i.e. MNA catalogue number), individual count, preparation (preservation method and more info about the sample, e.g., ETOH, dry, glass slides, etc.), event ID (i.e. original sampling station code), sampling protocol (sampling gear), event date, year, month, day, verbatim event date, field number (sampling station code as showed in the maps), event remarks (i.e., expedition), maximum depth meters, decimal latitude, decimal longitude, taxon ID, scientific name ID, scientific name, kingdom, phylum, class, order, family, genus, subgenus, specificEpithet, infraspecificEpithet, scientific name authorship, and taxon remarks. Some of the sampling stations are dredge stations, which have two sets of coordinates: the starting and end points. In these cases the coordinates reported in the dataset refer to the starting point of the dredge station.