From caves to seamounts: the hidden diversity of tetractinellid sponges from the Balearic Islands, with the description of eight new species

Study area

The Balearic Promontory (Fig. 1) is a seafloor elevation in the Western Mediterranean, of approximately 400 km length and 105 km wide, containing the Mallorca-Menorca shelf to the east and the Ibiza-Formentera shelf to the west. The continental shelf is narrow and shallow, with a mean depth of 87 m, characterized by the presence of calcareous sediment and by the scarcity of terrigenous input. It harbors rich photophilic habitats of soft and calcareous red algae that develop until depths of 100–150 m, leading to detrital muds of the shelf border. The slopes are very steep and descend until the surrounding abyssal plains of the Valencia Trough and the Algerian Basin (Acosta et al., 2003). As other areas of the Western Mediterranean, most sedimentary bottoms of the continental shelf and the upper and middle slope around the Balearic Islands, between 50 and 800 m depth, have been exploited by the trawling fleet for several decades (Farriols et al., 2017).

Two channels are present in the Balearic Promontory, the Menorca Channel (MeC), between Menorca and Mallorca, and the Mallorca Channel (MaC), between Mallorca and Ibiza. The first channel is narrow and shallow, and it is influenced by the strong northern winds originating in the Gulf of Lion and the hydrodynamic conditions of the Balearic sub-basin, mainly shaped by Mediterranean waters and under the influence of the Balearic Current, flowing along the northern shelf edge of the Balearic Promontory (Massutí et al., 2014). In 2014, the MeC was included in the Natura 2000 Network, in the light of the singularity of its habitats and its high diversity of benthos (Barbera et al., 2012). Conversely, the MaC separating the two shelves (Mallorca-Menorca and Ibiza-Formentera) is wider, deeper and more heterogeneous than the MeC, containing not only continental shelf and slope bottoms but also abyssal plain. The MaC, being located in the Algerian sub-basin, is more influenced by the Atlantic waters (Massutí et al., 2014); the MaC also contains several seamounts, among which stand out Ses Olives (SO), Ausias March (AM), and Emile Baudot (EB).

SO rises from 650–900 to 250 m depth at its shallowest part; it has a flat summit composed of fine sediments. AM has a minimum depth of 86 m and a height 264 m, with a summit in the mesophotic zone, where sediments are coarser, mainly composed of gravel and sand. Finally, EB represents a strongly irregular and uneven elevation, that rises from 900 to 94 m, with numerous mounds, depressions and rocky outcrops. Both SO and AM are of tectonic origin, while EB is of volcanic origin (Acosta et al., 2004). Patches of bio-constructions have been found in the summits of AM and EB, where rhodolith beds and coralligenous outcrops predominate, while rocky bottoms predominate in the flanks of SO, AM and EB, mainly colonized by filtering species, such as sponges and corals. According to these authors, in the less steep flanks and bathyal terraces of the upper and middle slope of these seamounts, muddy soft sediments are found, accumulating facies of the brachiopod Gryphus vitreus (Born, 1778), burrowing megafauna, small sponges and/or dead coral debris. The deepest areas of the middle slope at the base of seamounts are dominated by the finest muddy sediments and the presence of pockmarks fields (Massutí et al., 2022).

Sampling

Specimens were collected during (i) seven MEDITS research surveys on board the R/V Miguel Oliver, carried out annually from 2016 to 2021 on fishing grounds of the Balearic Islands shelf and slopes, between 50 and 800 m depth; (ii) one MEDITS survey carried out in 2020 on board the R/V Miguel Oliver along the northeastern Iberian Peninsula within the same bathymetric range; and (iii) four research surveys carried out on board R/V Ángeles Alvariño and R/V Sarmiento de Gamboa, within the framework of the LIFE IP INTEMARES project at the SO, AM and EB seamounts of the MaC, in August 2018, October 2019 and July–August 2020 (Figs. 1A–1B). The MEDITS program is carried out in most of the northern coast of the Mediterranean and aims to assess the state of the demersal resources and nekton-benthic ecosystems (Spedicato et al., 2019). The objective of the LIFE IP INTEMARES project at the MaC is to improve the scientific knowledge on biodiversity, benthic habitats and human activities, to include SO, AM and EB seamounts in the Natura 2000 network (Massutí et al., 2022). Several sampling devices were used in mesophotic and bathyal bottoms for both MEDITS and INTEMARES surveys: the experimental bottom trawl gear GOC-73 (Bertrand et al., 2002; Spedicato et al., 2019), a Beam Trawl (BT), Rock Dredges (RD) and the Remote Operated Vehicle (ROV) Liropus 2000 which was also used to film underwater. Screenshots of the film were used to study the in situ morphology of the specimens (Fig. 2).

Shallow water caves were explored by scuba diving or free apnea in May and November 2020 and in January, May and August 2021. Most of them can be classified as littoral marine caves created by sea erosion. They have salty water and marine fauna, being shallow (0–10 m depth) and located eastern (“Cova de sa Figuera”, “Cova de ca’n Rafalino”, “Cova de Cala Sa Nau”), northern (“Coves de Na Dana”), and western (“Cova Caló des Monjo”) off Mallorca island (Fig. 1B). Their sizes are quite variable, “Cova de Cala Sa Nau” being the largest and the most important in terms of benthic organisms, with a spacious entrance and a main chamber having 76/36/8 m in maximum length/width/depth (Gràcia, Clamor & Watkinson, 1998). This cave is commonly frequented by scuba divers, especially in summertime, potentially having a negative impact on the sponge community. In contrast, the “Cova de ca’n Rafalino” is the smallest, being a short tunnel only several meters long, with a depth of 1–2 m and between 0.5 and 3 m wide. Inland freshwater infiltration has been observed in “Cova de ca’n Rafalino” and “Coves de na Dana”. Due to the cave architecture, benthic organisms inhabiting the caves are relatively well protected from the action of waves. Details of sampling stations are summarized in Table 1. In situ images of the specimens were taken with an Olympus Tg5 digital camera (Fig. 3).

For oceanographic surveys, once the sampling gear was on board, sponges were separated from the rest of the catch and photographed, then macroscopic characters like morphology, color and texture were annotated prior to sample fixation. Samples for both morphological and molecular analysis were preserved in absolute ethanol (EtOH).

Specimens from this study were all deposited in the zoological collection at the Museum of Evolution, Uppsala University (Uppsala, Sweden) with UPSZMC# for non-type specimens and UPSZTY# for type material (Table S1), under Material Transfer Agreement 2023:074. Two exceptions: the holotype of Geodia phlegraeioides sp. nov. was deposited at the MNCN in Madrid (Spain), and the holotype of Caminus xavierae sp. nov. was already deposited at Naturalis in Leiden (The Netherlands). DNA extractions from the Balearic islands new species were deposited at the Museum of Evolution (holotypes) and at the Balearic Biodiversity Center (https://centrebaleardebiodiversitat.uib.eu/; paratypes), with same deposit numbers as the UPSZTY museum numbers.

The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: [urn:lsid:zoobank.org:pub:A88AE49E-B422-4F9A-A5E0-BB6C6B8FC185]. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central SCIE and CLOCKSS.

Morphological descriptions

To obtain dissociated spicules preparations, a fragment of tissue was digested with bleach, the remaining spicules washed with pure water first, then with 50% EtOH and finally with 96% EtOH. Spicules were observed and measured with an optical microscope. For each sample, unless otherwise indicated, 25 spicules per spicule category were counted. Spicule measurements given in the text are always the range observed from all measured specimens, unless otherwise stated. Handmade thick sections with a scalpel were made to study the skeleton oganization of every species. For precious type material, such as the Schmidt (1868) collection, regular thick sections (100–800 µm) were made by embedding small pieces of the specimens using an Agar Low Viscosity Resin kit (Agar Scientific, Essex, UK). Embedded pieces were sectioned with a diamond wafering blade on a Buehler IsoMet™ Low Speed cutting machine. For SEM images, aliquots of suspended spicules were transferred onto aluminum foil, air dried, sputter coated with gold and observed under a HITACHI S-3400N scanning electron microscope (SEM) at the Serveis Cientifíco-tècnics of the University of the Balearic Islands (UIB). The terminology applied for the morphological description of the spicules follows Boury-Esnault & Rützler (1997) and Hooper & Van Soest (2002).

Molecular analysis

DNA was extracted from a piece of choanosomal tissue (∼2 cm³) using the DNeasy Blood and Tissue Extraction kit (Qiagen, Hilden, Germany). Polymerase chain reaction (PCR) was used to amplify the Folmer fragment (658 bp) of the mitochondrial cytochrome c oxidase subunit I (COI) and the C1-C2 (∼369 bp) or C1-D2 (∼800 bp.) fragments of the nuclear rDNA 28S gene.

For COI, the universal Folmer primers LCO1490/HCO2198 were used (Folmer et al., 1994), except for the Craniella species for which we used primers LCO1490/COX1R1 (Rot et al., 2006); this primer set amplifies a longer fragment ca. 1180 bp (Folmer + Erpenbeck fragments). When LCO1490/HCO2198 failed to amplify COI (especially for some Erylus and Penares species), the primers LCO/TetractminibarR1 were used to amplify the first 130 bp of the Folmer marker, also called the Folmer COI minibarcode (Cárdenas & Moore, 2019). The primers jgHCO (Geller et al., 2013) and ErylusCOIF2 (5′-CTCCYGGATCAATGTTGGG-3′) were then used to amplify the rest of the Folmer fragment (Cárdenas et al., 2018). For 28S, the primer set C1’ASTR/D2 (Vân Le, Lecointre & Perasso, 1993; Cárdenas et al., 2011) was used to get the C1-D2 domains. When the C1’ASTR/D2 primers failed to amplify 28S, we used the primers C1’/Ep3 to get the shorter C1-C2 fragment. PCR was performed in 50 µl volume reaction (34.4 µl ddH20, 5 µl Mangobuffer, 2 µl DNTPs, 3.5 MgCl_2, 1 µl of each primer, 1 µl BSA, 0.1 µl TAQ and 2 µl DNA). PCR thermal profile used for amplification was [94 °C/5 min; 37 cycles (94 °C/15 s, 46 °C/15 s, 72 °C/15 s); 72 °C/7 min]. PCR products were visualized with 1% agarose gel and purified using the QIAquickR PCR Purification Kit (Qiagen, Hilden, Germany) and sequenced at Macrogen Inc. (Seoul, South Korea).

Sequences were imported into BioEdit 7.0.5.2. (Hall, 1999) and checked for quality and accuracy with nucleotide base assignment. Sequences were aligned using Mafft (Katoh et al., 2002). The resulting sequences were deposited in GenBank (http://www.ncbi.nlm.nih.gov/genbank/) with the following accession numbers: ON130519 –ON130569, OR045842 –OR045844 and OR045913 –OR045914 for COI and ON133879 –ON133850 and OR044718 for 28S (Table S1). Eight COI minibarcodes (111–130 bp), too small to be submitted to GenBank, were deposited on the Sponge Barcoding Project instead (https://www.spongebarcoding.org) with sequence numbers 2683 to 2690. The final COI and 28S alignment fasta files were deposited as Data S1.

Phylogenetic analysis were conducted using two different approaches: Bayesian Inference (BI) and maximum likelihood (ML), performed with the CIPRES science gateway platform (http://www.phylo.org; Miller, Pfeiffer & Schwartz, 2010) using MrBayes version 3.6.2 (Ronquist et al., 2012) and RAxML (Stamatakis, 2014). For MrBayes, we conducted four independent Markov chain Monte Carlo runs of four chains each, with 5 million generations, sampling every 1,000th tree and discarding the first 25% as burn-in, while RAxML was performed under the GTRCAT model with 1,000 bootstrap iterations. Convergence was assessed by effective sample size (ESS) calculation and was visualized using TRACER version 1.5. Genetic distance (p-distance) and number of base differences between pairs of DNA sequences were estimated with MEGA version 10.0.5 software (Kumar et al., 2018).

Comparative material and abbreviations

To help with our specimen identifications and descriptions, comparative material was used from the following institutions, for which we provide their abbreviations: BELUM Mc, Ulster Museum Belfast (Northern Ireland, UK); CEAB.POR.BIO, Porifera Collection at the ‘Centro de Estudios Avanzados de Blanes’ (Blanes, Spain); COLETA, ‘Coleção de Referência Biológica Marinha dos Açores’, reference collection of the Department of Oceanography and Fisheries, University of the Azores (Portugal); CPORCANT, Colección PORíferos del CANTábrico, IEO-CSIC (Gijón, Spain); HBOI, Harbor Branch Oceanographic Institute, Florida Atlantic University (Fort Pierce, FL, USA); MNCN, Museo Nacional de Ciencias Naturales (Madrid, Spain); MNHN, Muséum National d’Histoire Naturelle (Paris, France); MSNG, Museo Civico di Storia Naturale “G. Doria” (Genoa, Italy); NHM, Natural History Museum (London, UK); PC, personal collection of P. Cárdenas, Uppsala University (Sweden); RMNH, Rijksmuseum van Natuurlijke Historie, Naturalis Biodiversity Center (Leiden, The Netherlands); SME, Station Marine d’Endoume (Marseille, France); UPSZMC/UPSZTY, zoological collection at the Museum of Evolution (Uppsala, Sweden); ZMBN, zoological collection at the Bergen Museum (Bergen, Norway); ZMUC, Zoological Museum, University of Copenhagen (Denmark).

Type material from different museums were revised or re-examined for comparison with our specimens, especially from the natural history museums collections in London (UK), Paris (France) and Genoa (Italy). Notably, tetractinellids described by Schmidt (1868) from Algeria, currently stored at the MNHN Paris, were all examined. This historical collection gathers samples from the French ‘Exploration Scientifique de l’Algérie’ in 1842 and those collected by French zoologist Henri Lacaze-Duthiers in La Calle (El Kala) in 1860–1862, while he was studying the red coral.

Systematics

Material examined

UPSZMC 190946, field#i416_A, MaC (EB), St. 177 (INTEMARES1019), 151 m, beam trawl, coll. J. A. Díaz; UPSZMC 190944, field#i589_1, MaC (AM), St. 21, (INTEMARES0720), 112 m, beam trawl, coll. J. A. Díaz; UPSZMC 190945, field#i715_2, MaC (EB), St. 45, (INTEMARES0720), 147 m, beam trawl, coll. J. A. Díaz.

Comparative material

Stelletta dichoclada, holotype, MSNG 47152, NIS.83.34a, off Calvi (Corsica) 123–147 m, detrital bottom, July 1969, dredge (Fig. 5).

Stelletta lactea Carter, 1871, UPSZMC 190949, Strangford Lough (Northern Ireland), 0 m, October 2021, collected by hand at low tide, coll. C. Morrow, id. C. Morrow.

Outer morphology (Figs. 4A–4C and 5A)

Small subspherical, up to one cm in diameter, completely encrusted by calcareous sediment (Figs. 4A–4C), cortex and choanosome grayish in life and in EtOH. Sponges are slightly compressible, hispid to the naked eye. Cortex patent (Fig. 4C), about 1 mm in width.

Spicules (Figs. 4D–4LTable 2)

Plagiotriaenes (as the ones observed in the holotype: Fig. 5C) small, fusiform, slightly curved rhabdome, with a slight swelling just below the cladome. Clads are pointed upwards. Rhabdome: 167–650 × 6–19 µm, clads: 26–96 × 6–16 µm. Plagiotriaenes are very scarce, and may represent immature stages of the dichotriaenes because it is common to find small and incipient bifurcated clads in the plagiotriaenes; also, the dichotriaenes are always larger.

Dichotriaenes (Fig. 4D), rhabdome robust, straight or slightly curved, fusiform. The short-sized dichotriaenes have protoclads longer or the same length as deuteroclads while in large-sized dichotriaenes, protoclads are shorter than deuteroclads. Occasionally, 1–2 clads may not be bifurcated. Rhabdome: 305-1,515/12-52 µm, protoclad 33-102/12-43 µm and deuteroclad 14-181/6-33 µm.

Anatriaenes (Fig. 4E), very scarce, most are broken, with slightly curved rhabds. Rhabdome: 1,659-2,799/8-14 µm, cladi 14-41/6-10 µm.

Oxea (Fig. 4F), thin, slightly curved, and fusiform, 829-3,100/7-32 µm.

Oxyasters (Figs. 4G–4K), spherical, with short centrum and large actines, spiny along the whole actine with a clear pointy end. Only one size category, 7–31 µm in diameter.

Raphides in trichodragma (Fig. 4L), trichodragma length/width: 18-38/4-18 µm.

Ecology notes

Always found on sedimentary bottoms with sand and gravels, at mesophotic depths between 112 m and 151 m.

Genetics

We have sequenced the Folmer fragment in two pieces for specimens i589_1 and i715_2 (ON130566 and ON130567) and 28S (C1-D2) for all three specimens (ON133865, ON133864 and ON133866). The Folmer fragments of i589_1 (AM Seamount) and i715_2 (EB Seamount) have 1 bp difference. The 28S (C1-D2) of i589_1 (AM) and i416_A/#i715_2 (EB) have 1 bp difference.

Remarks

Specimens were found on the EB and AM seamounts at 112–151 m. This is the second record of S. dichoclada in the literature since its original description by Pulitzer-Finali (1983) from a specimen collected off Corsica at similar depths (123–147 m). The holotype is a small hemispherical sponge (Fig. 5A), 0.7 cm in diameter × 0.7 cm in height; openings not visible. Cortex is conspicuous, about 1 mm in width, beige in color, crusty to the touch, resilient, and incorporates sediment. Choanosome dirty orange, softer than the cortex (fleshier). Spicules of the holotype have been re-measured (Table 2) and examined with SEM (Figs. 5B–5H). Our material differed from the holotype by the presence of a few anatriaenes in specimens i589_1 and i715_2, but not in the holotype, nor i416_A. However, anatriaenes were very uncommon, which may explain their absence in the holotype and i416_A slides. Importantly, trichodragmas were not mentioned in the original description but they are definitely present in the holotype and our specimens (Figs. 4 and 5). To ensure that those are not foreign material, we have made digestions from two different parts of the holotype body: both contained trichodragmas. The holotype contains several foreign spicules, notably microtriods to microcalthrops with annulated rugose surface, identical to those described in the same work for Annulastrella verrucolosa (Pulitzer-Finali, 1983), collected at the same station. Also, there are foreign spirasters similar to those from the order Clionaida Grant, 1826.

We have not tried to amplify the DNA from the holotype because it had been conserved in formalin. Surprisingly, in the Balearic specimens we detected two haplotypes for both markers, each time with 1 bp difference. One haplotype corresponded to i589_1, collected at the AM, while the other haplotype was shared with i416_A and i715_2, from the EB. This may suggest that each seamount harbors isolated populations, or perhaps that S. dichoclada represents a species complex with two cryptic species. This should be assessed in further studies, by using more variable markers and sequencing more individuals.

Stelletta dichoclada appears to belong to a clade with North Atlantic species (S. lactea, Stelletta normani Sollas, 1880 and Stelletta rhaphidiophora Hentschel, 1929) (Fig. 6 and Figs. S1–S2). In this clade all species share dichotriaenes, trichodragmas and one category of oxyasters. For both markers, the closest sister-species is the shallow-water to intertidal North Atlantic S. lactea. We sequenced a specimen of S. lactea from the intertidal area off Northern Ireland (UPSZMC 190949, ON130565 (COI), OR044718 (28S)). Stelletta dichochlada and S. lactea UPSZMC 190949 have respectively 17–18 pb difference in COI and 19–20 bp difference in 28S so they are significantly different genetically, despite their morphological strong similarities. A SEM plate for S. lactea has also been made to compare the microscleres (Fig. S3). In fact, both S. lactea and S. dichoclada share similar spicular types, with similar morphologies: there is no clear spicule or external morphological difference between these species. Our S. lactea comparative specimen did have shorter/thicker oxeas, dichotriaenes with shorter rhabds, smaller oxyasters and longer trichodragmas (Table 2) but this would need to be confirmed with the measurements of several more S. lactea specimens.

However, there are also several Mediterranean records of S. lactea: Gulf of Lion (Boury-Esnault, 1971; Pouliquen, 1972), the Tyrrhenian Sea (Sarà, 1958; Sarà & Siribelli, 1960) and the Ionian Sea (Pulitzer-Finali, 1983), all from shallow waters, in agreement with most North Atlantic specimens but unlike S. dichoclada, which seems to be a mesophotic species. There are no sequences of Mediterranean S. lactea specimens, so its presence in the Mediterranean Sea cannot be confirmed, and relatedness with S. dichoclada cannot be currently assessed.

Material examined

UPSZMC 190943, field#i757, MaC (EB), St. 53, (INTEMARES0720), 97–102 m, rock dredge, coll. J. A. Díaz.

Comparative material

Stelletta mediterranea, holotype, MNHN DT2305 (two spicule slides), Cap l’Abeille, Banyuls, France, 30–40 m.

Outer morphology

Small, about 1.3 cm in diameter (Fig. 7A). Hemispherical body polarized in upper (rounded) and basal (flattened) parts; hispid surface. Externally pink when alive, grayish after preservation. Choanosome color not recorded on deck, grayish after preservation. Free of agglutinated sand. Cortex 1 mm thick. Hard consistency, barely compressible. Openings inconspicuous.

Spicules

Plagiotriaenes (Fig. 7B), very scarce, rhabdome slightly curved, fusiform, measuring 240–1070 (N = 3)/13-27 (N = 6) µm. Cladome, with clads measuring 73-132/13-24 (N = 6) µm. A single plagiotriaene with two bifurcated clads was observed (Fig. 7C), of the same length as the regular plagiotriaenes.

Anatriaene (Fig. 7D), rare, with teratogenic clads in form of aborted hooks just beneath the cladome.

Oxea I (Figs. 7E–7F), large, robust, fusiform, most are bent at the middle, 1,528-2,189-2,699/24-47-74 µm

Oxea II (Fig. 7G), small, thin, fusiform, bent at the middle, some slightly flexuous, 694-1,151-1,418/5-11-17 (N = 23) µm.

Oxyasters (Figs. 7H–7LL), abundant, only one size category (9-14-24 µm) but small ones are strongylaster-like, while larger ones are more like oxyasters; 5–18 actines, less actines in larger oxyasters. Spines are distributed all along the actine in small oxyasters, and absent near the centrum in large ones.

Raphides in trichodragma (Fig. 7M), length/width measuring 13-16-20/3-6-10 (N = 14).

Ecology notes

Found at the shallowest part of the EB summit (104 m). The area was rich in sponges and in coralligenous red algae. Epibiont on a large Irciniidae.

Genetics

Only the second part of the Folmer COI fragment (ON130568) was obtained; 28S (C1-D2) was also sequenced (ON133867).

Taxonomic remarks

This specimen is assigned to S. mediterranea, a poorly known species described by Topsent (1893) in Banyuls (France), and later recorded in the Alboran and Aegean seas (Pansini, 1987; Vamvakas, 1971). The spicules from the type slides were re-measured for the present study (Table 3). Similarities between our material and S. mediterranea are: (i) presence of two categories of oxeas, (ii) presence of plagiotriaenes (although Topsent called those spicules orthotriaenes, they are clearly pointing forward, (see Topsent, 1984, Plate XIV, Fig. 3)) and especially (iii) characteristic anatriaenes with teratogenic clads (Fig. 7D). Also, spicular sizes of both megascleres and microscleres fit with those of the holotype and Pansini (1987) (Table 3): the plagiotriaenes in the type are slightly more robust and its oxeas II are slightly thinner. Our specimen and the one from Pansini (1987) share a similar pink color when alive. Trichodragmas were not reported by Pansini (1987), but they could have been missed since they are not abundant in our specimen. The 28S tree (Fig. 6B) clearly suggests that this species groups with Stelletta grubii Schmidt, 1862 and Stelletta carolinensis (Wells, Wells & Gray, 1960) (well-supported) while its position with COI (Fig. 6A) is more ambiguous (not supported) and could be explained by the fact that we only have a small sequence.

Etymology

Due to its resemblance to a “morter”, a type of ancient pottery kitchen bowl commonly used in Mallorcan cuisine.

Material examined

Holotype: UPSZTY 190957, field#i714_1, St. 45 (INTEMARES0720), MaC (EB), beam trawl, 150 m, coll. J. A. Díaz (Fig. 8).

Paratypes: UPSZTY 190950-51, field#i352_1 and field#i352_2, St. 136, MaC (EB), beam trawl, 146 m, coll. J. A. Díaz; UPSZTY 190952, field#i401_2, St. 167, MaC (EB), beam trawl, 151 m, coll. J. A. Díaz; UPSZTY 190953-54, field#i406-A and field#i406-B, St. 167, MaC (EB), beam trawl, 151 m, coll. J. A. Díaz.

Other specimens: UPSZMC 190955-190956, field#i582 and field#i594 (Fig. 9), St. 21 (INTEMARES0720), MaC (AM), beam trawl, 109 m, coll. J. A. Díaz.

Comparative material

Stelletta defensa Pulitzer-Finali, 1983, holotype, MSNG 47153, NIS.83.36, Calvi, Corsica, Ligurian Sea, July 1969, dredge, 121–149 m, detrital bottom (Fig. S4); paratype, MSNG 47154, NIS.85.3, July 1969, dredge, 121–149 m, detrital bottom.

Stelletta dorsigera Schmidt, 1864, MNHN DCL4070, Roches Toreilles, France, 25 m, Oct. 1994, id. J. Vacelet and N. Boury-Esnault, COI: HM592750; 28S: AY348892.

Stelletta grubii Schmidt, 1862, BELUM Mc2668, Rathlin Ireland, Northern Ireland, summer 2005, id. B. Picton, 28S: HM592786 (Cárdenas et al., 2011).

Stelletta hispida (Bucchich, 1886), ZMBN 25636, Gulf of Cadiz, 1215 m, id. (Arnensen, 1920 (1932)).

Stelletta tuberosa (Topsent, 1892), MNHN DCL4066, Bay of Biscay, 4400 m, BIOGAS V expedition (Centob), id. P. Cárdenas.

Stelletta simplicissima (Schmidt, 1868), holotype, MNHN Schmidt collection#62, Algiers.

Stelletta stellata Topsent, 1893, UPSZMC 190958, South of Porto Cesareo lagoon, Apulia, SE Italy, 0.5 m, 27 July 2017. coll. P. Cárdenas and F. Cardone, id. P. Cárdenas and F. Cardone.

Outer morphology

Massive, circular to ellipsoid sponges, 3–6.5 cm in diameter, 2.5–6.5 cm in height with an atrium on its upper side (Figs. 8A; 9A). The atrium also has an ellipsoid shape, the opening 1.5–3 cm in diameter, and subsequent hole 1.5–3 cm deep (Fig. 9A). In specimen i582, the atrium does not generate a hole, but a concave depression at the surface. Color alive grayish (Fig. 8A). In EtOH, surface color dark gray and choanosome cream (Fig. 9A). Hard consistency, slightly compressible. Hispidation visible to the naked eye, present all over the surface, including the atrium. The atrium contains many small uniporal oscules, each with its own sphincter. Minute cribriporal pores are distributed on the sides of the specimens. Cortex ∼0.5 mm thick. Abundant sediments or pebbles are incorporated into the surface, but not in the choanosome, which is fleshy.

Spicules

Orthotriaenes (Fig. 8B), stout rhabdome, slightly curved, fusiform, with a sharp tip, measuring 482-1865/12-65 µm. Cladome also stout and with a sharp tip, 43-287/11-50 µm. The smallest triaene showed a marked swelling beneath the cladome, and its clads were more triangular.

Dichotriaenes (Fig. 8C), rare, only in specimens i714_1 and i401_2. Same size and morphology as orthotriaenes, rhabdome measuring 1,159-1,498/39-52 µm, while cladome measuring 29-71/33-44 µm (protoclad) and 86-133/24-36 µm (deuteroclad).

Anatriaenes (Fig. 8D), uncommon, rhabdome straight and stout, 1,626-3,055/5-20 µm. Cladome with tips of the cladi sharp, curved inwards, 18-87/3-18 µm. Some with underdeveloped cladome, resembling oxeas.

Protriaenes, very rare, rhabdome thin and slightly curved, measuring 1,156-1,267/5-9 µm. Cladi measuring 38-82/3-7 µm. Not found in i402_1 and i582.

Oxea I (Fig. 8E), robust, slightly curved, fusiform, 791-2,762/8-58 µm.

Oxea II, slender, slightly curved or flexuous, 753-1,627/5-11 µm. Common in specimen i352_1, very rare in i401_2, i582 and i714_1, and not observed in i594. Some have at their tip structures that remind of an aborted cladome so they may be anatriaenes with underdeveloped clads: this was pointed out by Topsent (1893) when describing similar oxeas II in S. mediterranea. which has similar oxeas II.

Oxyasters I (Fig. 8F), choanosomal, having 6–11 long actines, faintly spined (Fig. 8G), small centrum, 11–47 µm in diameter.

Oxyasters II (Figs. 8H–8J and 9B), ectosomal, having 9–15 short actines with more robust spines than oxyasters I and a centrum that is about 1/3 of the total diameter. The centrum is devoid of spines; overall measuring 11–24 µm.

Ecology notes

The species was found in the AM and the EB, on detrital bottoms with gross sand and gravels, from 111–152 m depth.

Genetics

COI (ON130562, ON130563, ON130564) and 28S (C1-C2) (ON133861, ON133862, ON133863) markers were obtained from i352_1 (paratype), i594 and i714_1 (holotype).

Taxonomic remarks

There are 10 species of Stelletta without raphides in the Northeast Atlantic/Mediterranean region: S. dorsigera, S. grubii, S. hispida, S. addita (Topsent, 1938), S. simplicissima, Stelletta pumex (Nardo, 1847), S. defensa, S. stellata, S. tuberosa and Stelletta ventricosa (Topsent, 1904). S. dorsigera has a conspicuous dark cortex with characteristic conules, unlike the cortex of S. mortarium sp. nov. Also, S. dorsigera is subspherical and not bowl shaped like our specimens. S. dorsigera has smaller oxyasters than S. mortarium sp. nov., measuring 8–12 µm (Uriz, 1981) versus 11–47 µm. S. grubii lacks anatriaenes and protriaenes and its orthotriaenes have much shorter and downwards curved clads. Besides, COI/28S of S. dorsigera and S. grubii are far apart from our COI/28S sequences in our phylogenetic analyses (Fig. 10).

S. hispida has large plagiotriaenes instead of ortho/dichotriaenes, anatriaenes and protriaenes. Besides, it has styles instead of oxeas and a small (2, 5 cm) spherical body shape (Bucchich, 1886). The size of the ortho/dichotriaenes in S. mortarium sp. nov. are four times longer and nearly two times thicker than in S. addita: dichotriaenes of S. addita have a rhabdome 225-350/25 µm (vs. 1,159-1,498/39-48 µm in S. mortarium sp. nov.). Also, the triaenes of S. addita are mostly dichotriaenes, while S. mortarium sp. nov., has mostly orthotriaenes. Moreover, no anatriaenes nor protriaenes were described for S. addita. Descriptions of the type of S. pumex are very poor (Nardo, 1847; Schmidt, 1864) but if we follow the short redescription made by Sollas (1888) it appears S. pumex has only one type of aster which can be quite variable (vs. two types of aster in S. mortarium sp. nov.) and only plagiotriaenes (versus essentially orthotriaenes in our species, along with some dicho-, ana- and protriaenes).

We found that in the MNHN Schmidt (1868) collection the holotypes of S. mucronatus (found in jar#63 labeled ‘Myriastra simplicissima’ and ‘Myriastra addita’) and Stelletta simplicissima (found in jar#62 labeled ‘Stelletta mucronata’) had been exchanged. Furthermore, according to the labels of jar#63 (MNHN DT 758), the holotypes of S. simplicissima and S. addita should have been stored together, since originally, they were both identified as S. simplicissima by Schmidt (1868). However, the holotype of S. addita was missing from either jar (#62 or #63). This may have happened when Topsent (1938) revised the Schmidt collection and described S. addita, the types were not placed back properly and the holotype of S. addita was misplaced and is presumably lost. The holotype of S. simplicissima is a brown subglobular specimen 2 × 3.5 cm. It has only stout plagiotriaenes with short clads (70–153 µm, our measurements) and very robust oxeas (2,300-2,700/80 µm, our measurements), while the largest oxeas of S. mortarium sp. nov. are 8–58 µm thick. Also, newly made thick sections of the holotype showed that Schmidt (1868) and Sollas (1888) overlooked short trichodragmas ∼12–15/5 µm long, which are not very abundant but clearly present close to the cortex. As for S. addita, it has essentially dichotriaenes (and some rare orthotriaenes) and two sizes of strongylasters (Topsent, 1938) while S. mortarium sp. nov. has essentially orthotriaenes and two sizes of oxyasters. S. tuberosa is a deep-sea species from the North Atlantic found deeper (454–4,400 m) than our specimens, they have a subspherical shape, much bigger megascleres and oxyasters (Cárdenas & Rapp, 2015). Their COI/28S is also quite different from those of our new species (Fig. 10).

The secondary loss of sterrasters in some Geodia (Geodiidae family) results in the same spicule repertoire as Stelletta species (Ancorinidae family), with triaenes, oxeas and asters (Cárdenas et al., 2011). Stelletta is therefore currently polyphyletic, with several of its representatives (e.g., S. tuberosa) grouping in a temporarily named ‘Geostelletta’ clade while others are true ancorinids (Stelletta sensu stricto) (Cárdenas et al., 2011). Both COI and 28S suggest that S. mortarium sp. nov. groups with Geodia, but not in the ‘Geostelletta’ clade, thereby suggesting there are several Stelletta-like Geodia clades amongst the Geodia. Actually, the position of S. mortarium sp. nov. is somewhat uncertain and poorly supported even within Geodia (Fig. 10). We refrain from allocating this species in the genus Geodia, until more species of Stelletta are sequenced so that new genera or subgenera can be formally created and defined based on shared morphological characters. We further note that so far none of the Stelletta-like Geodia possess raphides/trichodragmas, a spicule absent in the Geodiidae in general, so the presence of this spicule could be a good character to discriminate more efficiently some of the Stelletta sensu stricto species.

Material examined

UPSZMC 190959, field#i827_1, St. 25 (INTEMARES0820), MaC (EB), ROV, 100 m; UPSZMC 190960, field#POR1196, St. 212 (MEDITSGSA521), east of Mallorca (Cala Ratjada), 63 m, GOC-73, coll. J. A. Díaz; UPSZMC 190961, field#POR715, St. 184 (MEDITSGSA519), east of Mallorca (Portocolom), 52 m, GOC-73, coll. J. A. Díaz.

Comparative material

Stryphnus mucronatus, holotype, MNHN DT758, Schmidt collection#63, Algeria, ‘Exploration Scientifique de l’Algérie’, 1842; PC440, field#GOR 06.80, Gettysburg Peak, Gorringe Bank, scuba diving, 36–42 m, LusoExpedição 2006, coll. J. R. Xavier, specimen mentioned in Xavier & Van Soest (2007).

Outer morphology

Massive sponges, 5–12 cm in diameter (Fig. 11A). All three specimens grow on several calcareous red algae, which are included in the sponge body. Specimen i827_1 serves as a substrate for Haliclona (Flagellia) sp., P. monilifera and a calcareous sponge. Oscula cribiporal, with several orifices surrounded by a sphincter. On the video recording we observe an oscule, 1 cm in diameter (Fig. 2D). However, on deck we observe 2–4 contracted orifices, measuring 2–4 mm in diameter. Pores inconspicuous. Same color on deck and in EtOH: dark black cortex and a slightly paler choanosome (Fig. 11B). EtOH strongly colored dark by the specimens. Hard but slightly flexible consistency. Surface visually smooth but rough to the touch. Cortex patent, 3 mm in thickness (Fig. 11B, arrow).

Spicules

Dichotriaenes (Figs. 11C–11D), scarce, only found in i827_1. Rhabdome: 333-420-498/15-22-29 (N = 7), the protoclad: 40-49-62/13-20-25 µm and the deuteroclad: 21-49-73/8-13-17 µm (N = 11).

Plagiotriaenes (Fig. 11F), only two found, one in specimen POR715 and one in i827_1. This spicule showed tuberculous processes below the cladome. Plagiotriaenes probably represent immature stages of the dichotriaenes. Rhabdomes measure 209-302/9-12 µm (N = 2), while clads measure 54-79/9-10 µm (N = 2)

Oxeas (Fig. 11E), Fusiform, large, bent at the middle, rarely modified to styles: 400-2471/6-59 µm.

Oxyasters (Fig. 11I), small centrum and long actines, spined all over its shaft: 14–38 µm with 5–11 actines.

Amphisanidasters (Figs. 11G–11H), actines radiating from both ends of the shaft, spined: 7–15 µm long. In POR715 most are underdeveloped and have extra actines on its shaft.

Ecology notes

The species is not common in the Balearic Islands. It was found at the summit of the EB and in the fishing grounds east of Mallorca. Fishing ground stations were shallower (52 and 63 m) than the EB station (100 m). In both cases, sponges were growing on red algae bottoms.

Genetics

The COI Folmer region from specimen i827_1 (Figs. 2D and 11A) from the EB seamount was sequenced in two parts (ON130556). The miniCOI was sequenced for POR715 and the second part of COI for POR1196, both from fishing grounds east of Mallorca (SBP#2688 and ON130557). The COI of individual i827_1 and the COI fragments of POR715 and POR1196 match, and are identical to the sequence from specimen Po.25733 from the Eastern Mediterranean (Israel) (Idan et al., 2018).

Taxonomic remarks

Easily recognizable species, macroscopically characterized by its massive shape, black color, and thick cortex. The spicular set is quite homogeneous between the individuals of the Balearic Islands, as well as with individuals from other localities of the Mediterranean (Table 5). Triaenes are very rare, to the point that we did not find them in one specimen. As far as we know, this rarity of triaenes has not been reported before in this species (Topsent, 1894; Topsent, 1925; Vacelet, 1961; Vacelet, 1969; Pansini, 1987). The scarcity of triaenes is shared with Stryphnus raratriaenus Cárdenas et al., 2009, from the Caribbean. As already stated (Cárdenas et al., 2009), scarcity (or lack) of triaenes support the phylogenetic closeness to the genus Asteropus Sollas, 1888, which is essentially a Stryphnus without triaenes. Our findings (morphology and phylogenetic tree) suggest once again that both genera are probably synonymous, a fact that requires further revision of Atlantic species and sequencing of more species of Asteropus in particular.

Material examined

UPSZMC 190963, field#i208_b, St. 68 (INTEMARES0718), MaC (EB), rock dredge, 135 m, coll. F. Ordines & H. Marco; UPSZMC 190964, field#POR778_1, (MEDITSGSA06N), St. 3 (2020), GOC-73, 76 m (Benicassim), coll. J. A. Díaz; UPSZMC 190966, field#POR798, (MEDITSGSA06N), St. 14 (2020), GOC-73, 96 m (St Carles de la Rápita), coll. J. A. Díaz.

Outer morphology

Large specimens, up to 25 cm in diameter, 1.5–3 cm in width (Fig. 12A). Flattened, concave or irregular and slightly lobulated. With large holes and depressions that increase its exposed surface. Color in life of a dark tint on the upper side, with some whitish areas on its lower side. Color remains after EtOH preservation, yet the EtOH gets colored in black. Openings up to one mm in diameter, essentially located on the upper side of the body, surrounded by circular areas without pigment. Hard consistency. Hispidation localized. Poorly-delimited cortex, less than 0.5 mm thick.

Spicules

Dichotriaenes (Fig. 12B). Rhabdome short, straight, and fusiform, 310-732/17-58 µm. Cladome with protoclads projected outwards in a 120° angle with respect to the rhabdome 39-158/14-48 µm. Deuteroclads in a 90° angle, 31-170/8-35 µm. Rarely some of the clads may not be bifurcated.

Plagiotriaenes (Fig. 12C). Rabdome straight, most with a swelling just below the cladome. Scarce to absent in some specimens. Rhabdome 146-757/9-51 µm, cladi: 46-344/8-50 µm.

Oxeas (Fig. 12D). Stout, fusiform, slightly curved, 1,058-2,850/11-62 µm.

Oxyasters (Figs. 12E–12F). with numerous and sharp actines, more or less spined, 11–23 µm.

Amphisanidasters (Figs. 12G–12H). Strongly spined, 8–15 µm long.

Ecology notes

In the MaC, only found at the summit of the EB, an area with gross sand and both dead and live rhodoliths. Other large sponges were also collected in the same dredge, including Jaspis sp., P. compressa or P. monilifera. Also found at two stations of the fishing grounds in front of the Ebro delta, at 76 and 95 m depth. Individual i208_b from EB was free of epibionts but specimens from the fishing grounds were almost entirely overgrown by other sponges (e.g., Desmacella annexa Schmidt, 1870, Haliclona cf. fulva (Topsent, 1893)). The specific association with D. annexa has been reported before in the Northeast Atlantic (Cárdenas & Rapp, 2015).

Genetics

Only the miniCOI was obtained for i208_b from the EB seamount (SBP#2689).

Remarks

This species occurs in the Northeast Atlantic and the Western Mediterranean, from the intertidal to mesophotic depths (0–200 m, this study). For most of the records, only dichotriaenes were found, and no plagiotriaenes reported. Subsequently, a variety of S. ponderosus with plagiotriaenes was reported under the name S. ponderosus var. rudis, because Stryphnus rudis Sollas, 1888, a junior synonym of S. fortis, was known to have both dichotriaenes and plagiotriaenes. Latter, Cárdenas & Rapp (2015) synonymized S. ponderosus var rudis as S. ponderosus, also establishing the morphological and ecological distinctions between S. ponderosus (shallow temperate North Atlantic and Mediterranean species) and S. fortis (deep-sea boreo-arctic species). The presence, absence or abundance ratio between the plagiotriaenes and dichotriaenes seems to have a specific, although highly variable, value: S. ponderosus has a prevalence of dichotriaenes while S. fortis has a prevalence of plagiotriaenes.

We have found both plagiotriaenes and dichotriaenes in all the specimens. However, plagiotriaenes were abundant in the shallowest specimen (POR778_1, N = 17, 76 m), as opposed to very rare in the two deepest specimens (POR798, N = 6, 95 m; i208_b, N = 1, 135 m). It may suggest that plagiotriaenes are early dichotriaenes stages, and that specimens living in deeper waters having more silica at their disposal, their spicules are more easily fully developed and mature. Cárdenas & Rapp (2015) suggested that population differences could explain the different ratios between plagiotriaenes and dichotriaenes found in the different specimens of S. fortis. Future laboratory experiments with different silica levels could help understand the relationship between plagiotriaenes and dichotrianees in these species.

We only managed to sequence a Folmer minibarcode (130 bp) from specimen i208_b. The sequence is identical to a sequence of S. ponderosus from Northern Ireland (HM592685), a fact that confirms the Mediterrano-Atlantic distribution of the species. It has a 2 bp difference with S. fortis from Norway (HM592697) and a 4 bp difference with S. mucronatus.

In the Mediterranean, S. ponderosus has been widely reported. It is known from the Gulf of Lion (Vacelet, 1969), the Alboran Sea (Maldonado, 1992), the Catalan Coast (Uriz, 1981), the Adriatic Sea (Babiç, 1922) and the Aegean Sea (Voultsiadou, 2005). In the Balearic Islands, it is only known by an undocumented report at the Cabrera archipelago littoral (Uriz, Rosell & Martín, 1992). The present record is the second in the Balearic Islands, and the first on a Mediterranean Seamount.

Material examined

UPSZMC 190806-07, field#i693-i682, St. 43 (INTEMARES0720), 116–118 m, rock dredge, MaC (EB), coll. J. A. Díaz.

Comparative material

Calthropella (C.) pathologica, lectotype, MNHN DT753, Schmidt collection#66, Algeria, ‘Exploration Scientifique de l’Algérie’, 1842; paralectotype, MNHN DT754, Schmidt collection#87, Algeria, ‘Exploration Scientifique de l’Algérie’, 1842; PC324, La Ciotat, 3PP cave, 28S: AF062596; MNHN DCL4076, field#ASC6/327-6, Apulian Platform, off Cape Santa Maria di Leuca, Southern Italy, 39.56, 18.43, 560–580 m, ROV dive 327-6, Ifremer MEDECO leg1 (ifremer), 17 Oct. 2007, coll. J. Reveillaud, erroneously assigned to C. (C.) geodioides (Carter, 1876) in Cárdenas et al. (2011), COI: HM592705, 28S: HM592826.

Calthropella (C.) geodioides, holotype, NHM 82.7.28.16, Cape St. Vincent, Portugal, 534 m; ZMAPOR 21667, EMEPC/G3-D4-Ma11a, SE of Terceira Island, Azores, 38.4265°N, 26.8206°W, 1201 m, 18 May 2007, COI: HM592734, 28S: HM592825 (Cárdenas et al., 2011), SEM images presented in Van Soest, Beglinger & de Voogd (2010, Fig. 24).

Outer morphology

Massively irregular. Specimen i682 measures 7 × 9 × 4 cm and specimen i693 measures 5 × 4 × 3.5 cm, both overgrowing coralligenous red algae. Color in life beige with areas of dirty green, more localized in one of the body sides. On a transversal section, the green color persists in the first mm and then fades away turning into a beige choanosome. This first mm corresponds to a well-delimited cortex, which can be distinguished with the naked eye. Stony hard consistency, surface optically smooth, slightly rugose to the touch, overgrown by encrusting sponges (e.g., Jaspis sp.); pores inconspicuous.

Spicules

Calthrops (Figs. 13B–13E), with 2–5 cladi and large variability, including aborted actines, teratogenic modifications and stylote ends. On a wide but continuous size range, overall measuring 54-879/6-94 µm. Potentially divisible in two categories (54-278/6-37 µm and 318-879/31-94 µm) but with intermediate sizes. Mesocalthrop modifications present in both large and small sizes but more common in small ones.

Oxeas (not shown), thin, invariably broken, scarce, up to 2,839/17 µm.

Oxyasters (Figs. 13F–13G), scarce, with many rays and sometimes a few small spines (only detectable by SEM), 9–18 µm.

Strongylasters to spherasters (Figs. 13H–13K) with more or less spiny actines, variations with large centrum/short actines to small centrum/longer actines. Sizes of both kinds overlap so they probably belong to the same category, 6–23 µm.

Ecology notes

Both specimens were found at the same station; rhodolith bed at the summit of the EB, and both included rhodoliths to their bodies, which probably served as substrate. The station was rich in massive sponges like P. monilifera, Jaspis sp. or Spongosorites spp.

Genetics

We obtained the Folmer COI of i693 (ON130548) but only the miniCOI from i682 (SBP#2687). 28S (C1-C2) was obtained from i693 (ON133856).

Remarks

We assign with hesitation the Balearic material to C. (C.) pathologica. As stated by Van Soest, Beglinger & de Voogd (2010), C. (C.) pathologica and C. (C.) geodioides are morphologically similar, and may be differentiated by (i) the shape of the asters (longer actines in C. (C.) pathologica), (ii) occasional calthrops with dichotomous clads in C. (C.) geodioides (absent in C. (C.) pathologica) and (iii) mesocalthrops and dimesocalthrops in C. (C.) pathologica (absent in C. (C.) geodioides).

New spicules preparations were made from the holotype of C. (C.) geodioides and thick sections were made of the lecto- and paralectotype of C. (C.) pathologica; SEM images of the spicules of the lectotype of C. (C.) pathologica were previously presented by Van Soest, Beglinger & de Voogd (2010, Fig. 26). Based on re-examination of the types, MNHN DCL4076 from Santa Maria di Leuca was re-identified as C. (C.) pathologica and not C. (C.) geodioides as originally identified by Cárdenas et al. (2011). C. (C.) pathologica possesses long thin oxeas that form occasional large bundles, visible on the thick sections of the lectotype. We also found fragments of long oxeas in C. (C.) geodioides.

We confirm that asters usually have longer actines in C. (C.) pathologica than in C. (C.) geodioides. Dichocalthrops have not been found in any of the C. (C.) pathologica examined, including the holotype, which is in accordance with the literature. On the other hand, dichocalthrops of all sizes are more or less abundant in C. (C.) geodioides, including the type, where they are particularly numerous. Given the extreme variability in calthrop morphology (including cladi size, number and aberrant forms), this character may depend on ecological factors and is potentially misleading so it needs to be further tested in the future. After examining our comparative material, the presence of one or two (mesocalthrop) actines appears to be a solid character: in C. (C.) pathologica, small calthrops show ‘meso’ modifications, both in the form of 4 (fully developed) + 1 underdeveloped actine (Fig. 13C) and in the form of five (fully developed) actines (Fig. 13E) while in C. (C.) geodioides they always have 4 (fully developed) + 1 (underdeveloped) actine. Regarding the large calthrops, in C. (C.) pathologica those can have 4+1 and five actines while in C. (C.) geodioides they always show three, four or 4+1 actine. Analysis of COI/28S sequences revealed 2 bp. differences between the Azores specimen ZMAPOR 21667 (HM592734), and Mediterranean ones (including those from the EB seamount (i682 and i693), and specimen MNHN DCL4076 (HM592705) from Santa Maria di Leuca, Italy). To conclude, both species appear valid for now, based on three morphological differences and COI/28S.

Note: we noticed a typo in the C. (C.) geodioides material described by Van Soest, Beglinger & de Voogd (2010, p. 59 and Fig. 22): it should not be ‘ZMAPOR 21666’ (EMEPC/G3-D03A-Ma012) which is a C. (C.) durissima, IDed and sequenced by Cárdenas et al. (2011). Van Soest, Beglinger & de Voogd (2010) meant ‘ZMAPOR 21667’ (EMEPC/G3-D4-Ma11a). However, the collecting information is correct.

Material examined

UPSZMC 190808, field#LIT05, “Cova de sa Figuera” (cave), east of Mallorca, free apnea, 0–0.5 m, coll. J. A. Díaz.

Outer morphology

Hemispherical, 1.5 cm in maximum diameter (Fig. 15A). Ectosome light brown on live specimen, and after fixation in EtOH. Choanosome color not recorded on live specimen, whitish after fixation. Cortex has a stony hard consistency and is slightly rough to the touch, choanosome is hard but compressible. The cortex can be separated from the rest of the body, ∼0.5 mm thick. Surface covered with circular pores, which have a ring-shape in life (open), and a dot-shape after fixation (contracted). Three circular oscula, 1–2 mm in diameter in live specimen, slightly smaller due to contraction after fixation.

Spicules

Dichotriaenes, robust, with short conical rhabdome, 398–399 (N = 2)/65-72 (N = 3), long protoclads and short deuteroclads, 60-95/67-150 (N = 5) and 248-360/49-68 (N = 5), respectively.

Oxeas, fusiform, scarce, 1313-1860/15-34 (N = 3).

Sterrasters (Figs. 15B–15D), spherical to oval; no clear rosettes but intricate brain-like surface covered with small warts (Fig. 15BC)

Spiny spherasters to spherules (Figs. 15E–15G), 5-9-12 µm in diameter.

Oxyasters (Figs. 15H–15I), 4–15 actines, 13-19-27 µm in diameter. Actines are acanthose with robust and triangular spines that are also microspined. Large oxyasters tend to have less actines than smaller ones.

Ecology and distribution

A single specimen found in the dark area of a littoral cave, firmly attached to a vertical wall, no more than 30 cm deep. Our specimen was living in an area that emerges with high waves, a fact suggesting that the species can survive short periods of air exposure. No epibionts.

Genetics

Folmer COI (ON130547) and 28S (C1-C2) (ON133877) were obtained. COI was identical to previously sequenced C. intuta from Lebanon, and Portugal. The 28S was also identical to specimen ZMAPOR 21653 (Portugal), but differs in 1 bp with specimen SME PL617PC-7 from Cosquer Cave, Marseille, France.

Taxonomic remarks

The external morphology, spicular set/sizes matches well with a revision of the species, including a redescription of type material (Cárdenas et al., 2018). Genetically, the only discrepancy with the published sequences is the 1 bp difference in the 28S sequence when compared to a specimen from Marseille (PL617PC-7). Unfortunately, no COI is available for that same specimen. This is the first record of the species in the Balearic Islands, and the shallowest for the species.

Material examined

UPSZMC 190809, field#i142_C, St. 51 (INTEMARES0718), MaC (EB), beam trawl, 135 m, coll. J. A. Díaz; UPSZMC 190810, field#i254_4, St. 50 (INTEMARES1019), MaC (AM), beam trawl, 102 m, coll. J. A. Díaz; UPSZMC 190811, field#i391_2, St. 158 (INTEMARES1019), MaC (EB), beam trawl, 146 m, coll. J. A. Díaz; UPSZMC 190812, field#i526, St. 18 (INTEMARES0720), MaC (AM), beam trawl, 114 m, coll. J. A. Díaz.

Comparative material

Caminus vulcani, MNHN DT2288, slide, Banyuls, France, 30–40 m, specimen studied by Topsent (1894); SME, wet specimen, Cassidaigne canyon, off Marseille, France, 100–150 m, trawl, 16 June 1961, specimen studied by Vacelet (1969).

Caminus xavierae sp. nov., ZMAPOR 20422, holotype, Cueva Agua Dulce, Tenerife, Canary Islands, 5–10 m, scuba diving, field#CAN.07.05, 15 Jan 2007, coll. J. R. Xavier.

Outer morphology

Subspherical sponges (Fig. 16A) 3–8 cm in diameter, with an apical rounded oscula (2–4 mm in diameter), with a raised rim. Larger specimens (i254_4, i526) tend to acquire an ellipsoid, constricted shape. Same color in life and after preservation in EtOH: dark grayish ectosome and brownish choanosome. Hard (1 mm thick) but breakable cortex, pulpy choanosome. Smooth surface, with a mosaic visible to the naked eye, consisting of whitish polygonal patterns, which reflect the distribution of the pores; these patterns are even more obvious on dried cortex.

Spicules

Orthotriaenes (Fig. 16B), few, with clads curved forward, sometimes sinuous, short straight rhabdomes, only slightly longer than the cladi. Malformations like stylote termination or aberrant actines present in both rhabdome and cladome. Rhabdome: 606-668/15-28 µm, clads: 253-537/11-25 µm.

Strongyles (Fig. 16C), straight or curved, with round tips, 510-962/10-22 µm. Immature stages with oxeota ends are also present, 424-787/8-22 µm.

Sterrasters (Figs. 16D–16G), spherical, with a very pronounced hilum (Figs. 16E and 16F), 72–133 µm in diameter.

Spherules (Fig. 16H), rugose, 3–6 µm in diameter.

Oxyasters (Figs. 16I–16J), with small spines, 34–99 µm in diameter and having 2–8 actines, oxyasters with 2 actines tend to be larger.

Ecology notes

Only found at the summit of the EB and the AM, inhabiting the mesophotic zone (depth range 102–146 m).

Genetics

COI (ON130546) and 28S (C1-C2) (ON133892) were obtained from specimen i526. This is the first 28S fragment published for the genus Caminus. COI sequence was 8 bp different with C. xavierae sp. nov. from Tenerife (EU442205), and only 1 bp different with Caminus carmabi Van Soest, Meesters & Becking, 2014 from the Caribbean (MT815828).

Taxonomic remarks

See remarks below for C. xavierae sp. nov.

As Caminus vulcani: (Cárdenas et al., 2011; Cárdenas, 2020) (Figs. 5F and 5G).

Etymology

Named after sponge biologist Joana R. Xavier for collecting this species in the Canary Islands, and for her continuous efforts and leadership to support deep-sea sponge research.

Material examined

Holotype: ZMAPOR 20422 (wet specimen), UPSZTY 190813 (thick section and spicule slide), field#CAN.07.05, Cueva Agua Dulce, Tenerife, Canary Islands, 5–10 m, scuba diving, 15 Jan. 2007, coll. J. R. Xavier.

Paratype: UPSZTY 190814, spicule slide preparation, same locality as holotype, field#CAN.07.06, 15 Jan. 2007, coll. J. R. Xavier.

Comparative material

Caminus vulcani (this study).

Caminus carmabi, HBOI 11-V-00-1-007, specimen code 200005111007, Kaap Sint Marie, South coast, Curacao, 12.180550, -69.083980, 282 m, Johnson Link II-3209, 11 May 2000, id: P. Cárdenas, COI: MT815828.

Outer morphology

Holotype is 2.5 × 1 cm, massive encrusting with a unique oscule opening with raised margins (2–5 mm high) (Fig. 17A). Pores are distributed on the body of the sponge with the typical Caminus star-shaped pattern. Color in EtOH is light yellow to light brown; the raised oscule is lighter. Choanosome is beige.

Skeleton

Typical Caminus skeleton (Fig. 17B) with a distinct cortex (0.5 mm thick) essentially made of sterrasters, and covered with a thin layer of spherules. Below, a subcortical fibrous layer and a few triaenes supporting the cortex with their clades. In the choanosome, bundles of strongyles with no particular orientation, and numerous sterrasters, spherules and oxyasters.

Spicules (holotype and paratype)

Orthotriaenes (Fig. 17C), not abundant. sometimes ectopic clades on the rhabdome, rhabdome: 326-458/5-20 µm, clads: 77-326/14-23 µm.

Strongyles (Fig. 17D), straight or curved, 232-639/4-18 µm.

Sterrasters (Figs. 17E–17G), spherical to elongated, common unequal length of the actines giving a cauliflower-like aspect, 50–92 µm in diameter.

Spherules (Fig. 17H), 2–5 µm in diameter.

Oxyasters, sometimes looking like strongylasters (with thicker actines), 2–5 actines, many irregular, 13–37 µm in diameter.

Ecology notes

Until now only found in underwater caves (Agua Dulce and San Juan) in Tenerife, Canary Islands.

Genetics

COI from the holotype (EU442205) and the paratype (COI unpublished) are identical and differ in 8 bp with the COI of C. vulcani and 7 bp with the COI of C. carmabi.

Taxonomic remarks on Caminus vulcani and Caminus xavierae sp. nov.

Caminus vulcani is a well-known species, easily recognizable due to its macroscopical habit and its spicular set. In the Mediterranean, the species is recorded from shallow waters including caves (Topsent, 1894, 30–40 m; Pulitzer-Finali, 1983, 15 m; Grenier et al., 2018) to mesophotic depths (Vosmaer, 1894, 150–200 m; Vacelet, 1969, 100–150 m; Maldonado, 1992, 70–120 m). In general terms our material matches with the revision by Uriz (2002) only differing in that MaC specimens have strongyles in a wider size range (510-962/10-22 µm vs 850 vs 880/15-17 µm) and larger oxyasters (36–95 vs 35–42 µm). Besides, type material of C. vulcani comes from the Adriatic Sea (Schmidt, 1862), so we consider that our material is conspecific with the type because of geographical proximity and morphological similarities. It is reported here for the first-time in the Balearic Islands.

The only published Atlantic records of C. vulcani came from shallow caves in Tenerife, Canary Islands (Cruz, 2002; Cárdenas et al., 2010). Interestingly, COI from those Canary Island specimens differs by 8 bp with COI of our specimens from the MaC seamounts, which clearly indicates that they are two separate species. A new species is proposed for Canary Island specimens, Caminus xavierae sp. nov., characterized by three main spicule differences with C. vulcani: (i) shorter strongyles (232-639/4-18 µm vs. 510-962/8-22 µm in C. vulcani), (ii) smaller sterrasters on average (average sizes of 73–80 µm vs. average sizes of 92–104 µm) and (iii) shorter oxyasters (13–37 µm vs. 36–99 µm) (Table 7). The “cauliflower” morphology of the sterrasters (vs. regular subspherical shape in the other Caminus species) may be another specific character but SEM examination of more specimens is required to confirm this.

Caminus xavierae is the fourth species of Caminus in the Atlantic after the Caribbean C. carmabi, the Caribbean/Brazilian Caminus sphaeroconia Sollas, 1886 and the Mediterranean C. vulcani. C. sphaeroconia differs from all with the absence of oxyasters while the sterrasters (average of 190/144 µm) and oxyasters (average of 65 µm) of C. carmabi are much larger than in C. xavierae.

Interestingly, the COI sequence of C. vulcani (i526) and C. carmabi showed they are genetically closer to each other (1 bp. difference) than to C. xavierae sp. nov. (7–8 bp). This confirms the morphological similarities observed between them, highlighted in the original description of C. carmabi: “Our material is most similar to Mediterranean Caminus vulcani Schmidt, 1862, but in that species sterrasters are smaller (105-115/85-88) and calthrops have also shorter and thinner cladi” (Van Soest, Meesters & Becking, 2014)”. C. vulcani and C. carmabi are therefore mesophotic sister species. Such small genetic differences in COI on either side of the Atlantic is in accordance with what is observed in other tetractinellid sister-species (Cárdenas P., unpublished data). In the 28S tree, C. vulcani groups with Geodia and not with the Erylinae (Fig. 10). This unexpected result implies a long branch and no bootstrap support so it may be explained by the short length of our fragment and low polymorphism of the sequence obtained.

Material examined

UPSZMC 190838, field #LIT10, Calo d’en Rafalino (Cala Morlanda), Mallorca, semi-submerged cave, 0–0.5 m, free apnea, coll. J. A. Díaz.

Comparative material

Erylus deficiens, holotype, MNHN DT1111 (slide), Porto Santo Bay, Madeira, 33°2′N, 16°19′45″W, 100 m, St. 801, 2 July 1897, trawl; ZMAPOR 21693, Gettysburg Peak, Gorringe Seamount, 32 m, 3 June 2006, coll: J. Xavier, field# GOR 06.01, COI: HM592687, 28S: HM592823, specimen identified as Erylus sp. in Xavier & Van Soest (2007) and Cárdenas et al. (2011); ZMAPOR 20419, Reserva do Garajau, Madeira, 7 m, 17 Feb. 2005, coll: J. Xavier, field#MAD.05.02.31, COI: EU442204, 28S: EU552088.

Outer morphology

Small crust (Fig. 18A), 4–5 cm in maximum diameter, 0.2–0.4 cm in width. Whitish in life, dark brown after EtOH fixation. Skin smooth and hard to the touch, choanosome crumbly. Minute pores, <0.1 mm in diameter (visible to the naked eye), gathered (Fig. 18B). Oscules not observed.

Spicules

Dichotriaenes (Figs. 18C–18D), uncommon malformations like aborted actines and stylote terminations. Juvenile stages with protoclads much longer than deuteroclads, which are very small, sometimes barely visible. Rhabdome:145–182 (N = 4)/8-17-26, the protoclad: 47-67-86/8-14-25 µm and the deuteroclad: 11-79-135/4-11-18 µm.

Orthotriaenes, rare, rhabdome 76 (N = 1)/8-11 µm (N = 4), cladi 61-133/6-10 µm (N = 4).

Oxeas (Fig. 18E), robust, straight to slightly curved, and fusiform, 280-503-647/4-11-17 µm.

Styles to strongyles (Figs. 18F–18G) look like oxea modifications, tend to be shorter and thicker, especially the strongyles. Styles measure 193-520/14-15 µm (N = 2), while strongyles measure 248-412-504/11-17-27 µm (N = 14).

Aspidasters (Figs. 18H–18J), very scarce, always with an irregular shape due to unequal actine lengths, 31-42-56 µm (N = 13).

Oxyasters (Figs. 18K–18L), with 4–9 spined actines, 12-21-36 µm.

Microrhabds (Fig. 18LL), very spiny, with less spines at the center, some being centrotylote, 22-36-55/2-3-6 µm.

Ecology and distribution

Single specimen found in a littoral, semi-submerged cave in the intertidal zone, periodically exposed to the air. The cave receives freshwater inputs that may increase silicon levels.

Genetics

Folmer COI (ON130535) and 28S C1-C2 (ON133853) were obtained.

Taxonomic remarks

See general discussion on E. discophorus, E. mamillaris and E. deficiens below, after the description of E. mamillaris.

Material examined

UPSZMC 190846, field#POR785, St. 6 (MEDITS06N20), fishing ground off Columbretes Islands, GOC-73, 144 m, coll. J. A. Díaz.

UPSZMC 190847-49, field#LIT71, field#LIT72 and field#LIT74, Coves de na Dana (Alcudia), Mallorca, semi-submerged cave, scuba diving, 0–1 m, coll. J. A. Díaz and A. Frank.

Comparative material

Scutastra cantabrica Ferrer-Hernández, 1912, paratype, NHM 30.1.21.5, wet specimen, Santander, Spain; Erylus discophorus, ZMUC, off São Pedra Bay, São Vincente, Madeira, 40 m, St. 40, originally identified as S. cantabrica by Burton (1956), here re-identified.

Outer morphology

Fishing ground specimen POR785 is massive (Fig. 20A), lobulate, measuring 10 cm in height and 7 cm in diameter. Color beige with dark shades in life, and after fixation in EtOH. Dark shades more present on the upper part and around the oscula. Choanosome beige. Surface smooth. Consistency hard but slightly flexible. Circular oscula, 2–3 mm in diameter, placed at the top of the lobules. Inhalant pores not observed.

Cave specimens are encrusting (Fig. 20B), 0.3–0.5 cm width, spreading 7–8 cm on the vertical walls. Color in life whitish to beige with some brownish areas, probably caused by diatoms. Same color for the ectosome and the choanosome. Color slightly paler after fixation in EtOH. Inner channels visible only in areas where the body was thinner. Surface smooth, but wrinkled after collection due to contraction. Hard consistency. When alive, many small and circular oscula visible, 1–2 mm in diameter, aligned on the top of small ridges (Fig. 20B). Pore groupings visible to the naked eye, in depressed parts of the specimen.

Spicules

Dichotriaenes, robust, with short and fusiform rhab. Rhabdome: 123-557/8-62 µm, protoclad: 46-111/7-52 µm and deuteroclad: 10-262/3-44 µm.

Orthotriaenes, very scarce, only found in POR785 (N = 3) and LIT72 (N = 1), rhabdome 135-464/9-18 µm, clad 96-164/7-17 µm.

Oxeas, slightly curved and fusiform, 388-978/5-29 µm, sometimes modified to styles (384-650/10-25 µm) and strongyles (183-796/11-22 µm).

Aspidasters (Figs. 20C–20F), circular to slightly elongated, 33–95 µm (max. diameter).

Oxyasters (Figs. 20G–20I), with 4–12 spined actines, 6–30 µm.

Microrhabds I (Fig. 20J), densely recovered with robust spines, centrotylote, 14-68/1-5 µm.

Microrhabds II (Figs. 20K–20LL) uncommon, smooth to microspined, curved and centrotylote, 41-89/3-6 µm.

Ecology and distribution

Species found in a fishing ground close to Columbrets (POR785) and in a shallow water cave with freshwater inflow (LIT71, LIT72 and LIT74). Cave specimens were very abundant, and found just below the water surface. The only previous mention of E. discophorus in the Balearic Islands was from shallow caves off Cabrera Archipelago (Uriz, Rosell & Martín, 1992).

Genetics

Folmer COI was obtained from all specimens (POR785, ON130531; LIT71, ON130532; LIT72, ON130533; LIT74, ON130534) whereas 28S (C1-C2) was obtained only from POR785 (ON133854).

Taxonomic remarks

See general discussion after the description of E. cf. mamillaris.

Material examined

UPSZMC 190850, field#i142_B, St. 51 (INTEMARES0718), MaC (EB), 128 m, beam trawl, coll. F. Ordines; UPSZMC 190851-52, field#i179_A-179_B, St. 60 (INTEMARES0718), MaC (EB), 138 m, beam trawl, coll. F. Ordines; UPSZMC 190853-56, field#i314, field#i329_A, field#i329_B, field#i329_C, St. 124 (INTEMARES1019), MaC (EB), 152 m, beam trawl, coll. J. A. Díaz.

Comparative material

Erylus cf. mamillaris, ZMAPOR 20421, Ponta Furada, Faial Island, Azores, 2–8 m, 6 Sept. 2005, field#FUR05.09.14, coll: J. R. Xavier, COI: EU442207, 28S: EU552090.

Outer morphology

Massive, ovoid and lobated sponges (Fig. 21A), which often agglomerate foreign sediments, pebbles, worm tubes. The largest specimen (i314) is 9.5 × 5 cm. Single apical oscule on each lobe. Dark brown on its upper side, progressively fading to light brown or whitish at its basal area; choanosome lighter, cream colored. Surface visually smooth but rough to the touch, texture is quite firm, only very slightly compressible. Cortex less than 1 mm thick, clearly distinguishable. Choanosome fleshy, light brown and showing a well-developed aquiferous system.

Spicules

Dichotriaenes (Fig. 21B), scarce, may be significatively thick. Rhabdome: 335 (N = 1)/27-40-57 (N = 5) µm, protoclad: 115-133-151/27-34-52 (N = 5) µm and the deuteroclad: 119-139-161/19-27-39 (N = 5) µm.

Oxeas (Fig. 21C) robust and fusiform, slightly curved 429-1,037/5-28 µm. One single stylote modification was observed in specimen i179_B (not measured).

Aspidasters (Figs. 21D–21H), slightly elongated, 64-121/33-91 µm (length/width).

Oxyasters (Figs. 21I–21J), 4–13 spined actines, 14–56 µm in diameter.

Microrhabds (Fig. 21K), spined and centrotylote, 15-31/1-4 µm.

Ecology and distribution

Species found at several stations on the EB summit, always on sedimentary bottoms at mesophotic depths, just below the photic zone.

Genetics

Folmer COI (ON130529) and 28S C1-C2 (ON133884) obtained from specimen i329_B.

Taxonomic remarks on Erylus discophorus, E. mamillaris and E. deficiens

E. mamillaris, E. discophorus and E. deficiens form a poorly understood complex of Mediterranean and Northeast Atlantic species with similar spicule sets: dichotriaenes, spiny microrhabds, spiny oxyasters and aspidasters (Sollas, 1888; Topsent, 1927; Topsent, 1928; Cárdenas et al., 2011; Cárdenas, 2020). The conspecificity of specimens assigned to one or the other species have been extensively debated in the literature (Sollas, 1888; Von Lendenfeld, 1894; Marenzeller, 1889; Topsent, 1901; Topsent, 1928; Pulitzer-Finali, 1972). A character proposed to differentiate these three species are aspidasters size, morphology and abundance: larger and more elongated in E. mamillaris vs. small and more rounded in E. discophorus and extremely scarce to absent in E. deficiens (Topsent, 1928). We note that this has some consequences on the rosette arrangement of aspidasters in the three species, giving more of a radial pattern in E. discophorus, a more regular pattern in E. mamillaris and a more irregular/disorganized pattern in E. cf. deficiens. Also, the size and abundance of microrhabds seem to differ between species, being smaller and less common in E. mamillaris and larger and more common in E. discophorus (Sollas, 1888; Von Lendenfeld, 1894) and E. deficiens (Topsent, 1928; Vacelet, 1976; cf. holotype redescription in Table 8). Von Lendenfeld (1894) also mentioned microrhabd morphology as a significant character, being more heterogeneous in E. discophorus than in E. mamillaris (with pointed and rounded variations). In the same work, the presence of differently spined to smooth microrhabds in E. discophorus is also mentioned (see Table 8 for details). The Balearic specimens were identified primarily according to aspidaster characters, then the microrhabds were carefully compared (Table 8). Microrhabds are indeed shorter in E. mamillaris (15–31 vs. 14–68 µm). A second rare category of microrhabds (microrhabds II) was found in E. discophorus: larger, smooth to minutely spined. These microrhabds II tend to have pointed tips, close to those drawn by Von Lendenfeld (1894, Taf III Fig. 41A). This rare spicule type is not singled out in previous descriptions (Table 8) maybe because they are too rare and not easily spotted; for instance, we could not find them in the paratype of Scutastra cantabrica. So we are hesitant to consider them as a specific character of E. discophorus. However, microrhabds II were absent in all E. mamillaris and E. deficiens specimens examined, and must be taken into account for a future revision of the types and this species complex.

Regarding spicule size variation in E. discophorus, the deeper specimen POR785 had wider oxeas, triaenes and microrhabds and larger oxyasters than the cave specimens (LIT71, LIT72, LIT74). This may be explained by ecophysiological differences, deeper specimens are usually subjected to higher silica concentrations and therefore larger spicules, a phenomenon well described in the Geodiidae (Cárdenas & Rapp, 2013). There may be other parameters at play: fishing grounds near Columbrets are a more eutrophic habitat than shallow caves of Mallorca, mainly because of the influence of the river inflows from the Iberian Peninsula.

Macroscopically, E. deficiens tends to be described as having a massive lobated external morphology with one single large oscule at the summit of each lobe, and a dark smooth cortex (Topsent, 1928). The species was first considered a variety of E. discophorus by Topsent, but later erected as a valid species. It was described in Madeira (at 100 m depth), and later collected in the Gorringe Bank (Xavier & Van Soest, 2007) and the Ligurian Sea (Topsent, 1927; Vacelet, 1976). The most remarkable character of E. deficiens is the rarity to complete absence of aspidasters. Moreover, Topsent (1928) mentions the higher abundance of microrhabds in E. deficiens than in E. discophorus, to compensate the deficit of aspidasters (the more aspidasters the less microrhabds and vice versa). Topsent (1928) also proposed growth habit as a character to separate both species: E. deficiens being much larger with lobate processes and E. discophorus encrusting and smaller. A spicule slide (MNHN DT 1111) of the holotype of E. deficiens from Madeira was compared with specimens morphologically identified as E. deficiens from Gorringe Bank (ZMAPOR 21693) (Xavier & Van Soest, 2007), Madeira (ZMAPOR 20419) and our cave specimen (LIT10). All specimens were massive except for the encrusting Mallorcan specimen; spicule sizes were all quite similar (Table 8). Microrhabds were similar as in E. discophorus but comparatively slightly shorter and thinner.

Type material of S. cantabrica from the northern coast of Spain (Santander, Cantabria) is reexamined here for the first time: a new spicule slide was made and spicules measured (Table 8). Shape and size of the spicules conform to those of E. discophorus, except for irregular megascleres and the discoid aspidasters which are common but slightly smaller (32-41-50 µm in diameter) and underdeveloped (actines not fused). Such underdeveloped aspidasters and irregular megascleres are actually not uncommon in several E. discophorus specimens from Banyuls, France (Boury-Esnault & Lopes, 1985) or Portugal (Cárdenas & Rapp, 2013, Fig. 12A) and could be linked to low silica concentrations (Cárdenas & Rapp, 2013). We therefore confirm the suggestion of Boury-Esnault & Lopes (1985) in considering S. cantabrica a junior synonym of E. discophorus. Another specimen from Madeira identified as S. cantabrica by Burton (1956) was also re-examined: it has abundant aspidasters (discoidal to elongated, some underdeveloped, 30–55 µm), oxyasters (8–22 µm), mostly straight non-centrotylote spiny microrhabds and regular megascleres: it is therefore re-identified as a typical E. discophorus.

Cárdenas et al. (2011) had shown that these three Erylus species were genetically very close. The present study enriches the sequence sampling to further test the validity of these species. The COI tree (Fig. 19A) is more informative than the 28S tree because we only obtained the more conserved 28S (C1-C2) region which showed no bp differences between the species (Fig. 19B), so in this case we will only discuss the COI tree. In the COI tree, E. discophorus specimens POR785, LIT71, LIT72 and LIT74 strongly group together in a cluster that also includes Erylus sp. (EU442206) from Portugal (6 m depth), and E. discophorus (HM592692) from Slovenia (5 m). The cluster includes several haplotypes differentiated by 0–2 bp differences. This number of bp differences in COI may indicate the presence of two cryptic species, but also be caused by intraspecific variability. It is far beyond the scope of the present work to elucidate this, and future works with a more extensive sampling shall be conducted. Identical COI sequences were obtained for the encrusting LIT71 and the massive deeper POR785 suggesting that shape is not a diagnostic character. More variable genetic markers would be needed to see if they could be ecotypes. A second paraphyletic group is represented by Atlantic specimens E. cf. mamillaris (Azores, 7 m) and two E. deficiens (Gorringe Bank, 32 m and Madeira, 7 m) with 1 bp. difference. E. cf. mamillaris has 2 bp differences with both E. deficiens sequences. A third group is only represented by the cave Mallorca specimen LIT10 identified as E. cf. deficiens, which is 6–8 bp different from the other sequences. Since the type locality of E. deficiens is Madeira, we can be more or less confident that our specimens sequenced from Gorringe and Madeira are closer to the type locality. This suggests that LIT10 most probably represents a new species from the Mediterranean Sea. However, the difficulty to find clear diagnostic morphological characters and the presence of only one specimen pushes us to delay a new species description. This result also suggests that the character of rare aspidasters may have appeared in different lineages of Erylus independently. Finally, the COI sequence of specimen E. cf. mamillaris i329_B diverges alone, clearly apart from another E cf. mamillaris from the shallow Azores. This suggests that slightly elongated aspidasters and small microrhabds may not be good specific characters either. All the deep specimens collected from the EB seamount (i179_1, i179_2, i314, i329_A and i329_C) shared the same spicular characters so they must all belong to the same species, which may be conspecific with E. mamillaris from the Adriatic Sea. However, comparison with type material would be necessary to address this matter in future works.

Altogether, these phylogenetic results cast doubt on the current spicule characters (aspidaster morphology mainly) used to discriminate these species. To our knowledge, this is the first time such a sponge species complex is revealed in the Mediterranean Sea with many COI haplotypes from populations from different depths and habitats. Additional genetic markers and specimens from these different populations will be necessary to resolve the Erylus discophorus/mamillaris/deficiens complex.

Material examined

UPSZMC 190845, field#i707, St. 45 (INTEMARES0720), MaC (EB), 150 m, beam trawl, coll. J. A. Díaz; UPSZMC 190841, field#i389_1, MaC (EB), St. 158 (INTEMARES1019), beam trawl, coll. J. A. Díaz; UPSZMC 190840, field#i356_A, MaC (EB), St. 136 (INTEMARES1019), beam trawl, coll. J. A. Díaz; UPSZMC 190843, field#i402_B, MaC (EB), St. 167 (INTEMARES1019), 151 m, beam trawl, coll. J. A. Díaz.

Comparative material

Erylus corsicus, holotype (slide), MSNG 47157 (NIS.85.14), off Calvi (Corsica), 121–149 m, 14 July 1969.

Erylus papulifer Pulitzer-Finali, 1983, holotype, MSNG 47155 (NIS.19.4a), wet specimen, off Calvi Corsica, 135 m, 18 July 1975; paratype, MSNG 47156 (NIS.19.4c), same locality as holotype.

Erylus expletus Topsent, 1927, holotype, MNHN DT837 and DT1326, two slides, off Säo Jorge, Azores, 1 Aug. 1895, St. 616, 1022 m; ZMAPOR 18142, SE of Rockall Bank, West coast of Ireland, 55°30′13.93″N, 15°47′6.18″W, 679 m, 2 Sept. 2004, coll: R. van Soest, field# M2004/33-05, id. P. Cárdenas, COI: EU442208.

Outer morphology

Massive globular sponges, 1 to 3.5 cm in diameter (Fig. 22A). Hard, slightly compressible cortex, fleshy choanosome. After fixation in EtOH, the choanosome contracts and tends to get separated from the cortex, which remains resilient. Beige color in life, some individuals have a brownish tinge in localized areas, product of diatom colonization. The brownish stains disappear after EtOH fixation, and then both cortex and choanosome are beige. Small abundant openings, often with a lighter colored ring, scattered all over the surface, 0.1–1 mm in diameter.

Spicules

Orthotriaenes, very scarce, cladi curved outwards and rhabdome straight. Rhabdome: 340-654/12-31 µm. Cladi: 172-387/11-31 µm.

Oxeas (Fig. 22B), slightly curved and fusiform, 534-1,371/6-25 µm.

Aspidasters (Figs. 22C–22F), discoid to elongated but most are oval, being up to two times longer than wider. Underdeveloped or aberrant forms are present, more or less common depending on the specimen. Rosettes of type 3 (Fig. 22D) sensu Cárdenas (2020); well defined hilum (Fig. 22F). On average, length: 133–250 µm, width: 106–202 µm.

Microrhabds (Figs. 22G–22H), on a wide size range, smooth, slightly curved and faintly centrotylote, 35-171/2-10 µm.

Toxas (Figs. 22I–22J), scarce, most 2-actined, rarely 3-actined. With a central swelling, overall measuring 204-425/6-10 µm.

Ecology notes

Species found at the summit of the EB, between 146–151 m, on sandy sedimentary bottoms. The species is always covered by a brownish coating produced by diatom aggregations (observation with microscope). The diatoms seem to be favored by the smooth surface that offers the aspidasters on its external side, acting as substrate.

Genetics

COI was obtained from specimens i707 (ON130536) and i402_B (ON130537) while the 28S (C1-C2) fragment has only been obtained from specimen i707 (ON133851).

Taxonomic remarks

Thought to be divergent reduced oxyasters (Topsent, 1928), toxas are a rare spicule type in Erylus species. And yet, three very similar Erylus with toxas are recorded from the Mediterranean Sea: E. corsicus, E. papulifer and E. expletus (this last one is also reported from the Northeast Atlantic). The type material of these three species was compared here for the first time to test their validity and to compare with our material (Table 9). E. expletus is the first one to be described, from the deep waters of the Azores, 1022 meter depth (Topsent, 1927; Topsent, 1928) and subsequently from Rockall Bank (Van Soest et al., 2007), shallow waters of the Canary Islands (Cruz, 2002) and underwater caves around Marseille, France (Pouliquen, 1972). E. papulifer is based on six specimens collected from mesophotic depths in Corsica (Pulitzer-Finali, 1983). The type material has been re-examined here (Figs. 23A–23K): they are subglobular sponges, the holotype measures 2.5 cm in diameter (Fig. 23A) while paratypes measure about one cm in diameter. They have a hard but breakable consistency, slightly compressible. Color after formalin fixation is dirty beige with orangish pink (peach) areas. Small abundant openings are scattered all over the surface, 0.1–1 mm in diameter. In terms of external morphology, E. expletus (Topsent, 1928, Pl. I, Fig. 20), E. papulifer (Fig. 23A) and our material (Fig. 22A) are very similar. Actual differences between E. papulifer and E. expletus were not clearly stated by Pulitzer-Finali (1983) except for “the shape and size of aspidasters”. E. corsicus is also described by Pulitzer-Finali (1983), based on a single tiny specimen, which was completely digested to make a single spicule preparation, which was examined here. The E. corsicus specimen was collected in the same dredge as several E. papulifer specimens. Pulitzer-Finali (1983) based E. corsicus on having larger and more irregular aspidasters than E. papulifer, and a wider length range of microrhabds (48–160 µm vs. 50–80 µm in E. papulifer).

After careful observation and measurements of spicules (Table 9), we first note that E. papulifer is the only species with dichotriaenes while E. expletus and E. corsicus have very few orthotriaenes. E. expletus seems to be discriminated by (i) a majority of aspidasters with a characteristic lemon-shape (Topsent, 1928, pl. V, Fig. 10), while E. corsicus/papulifer have oval to lemon-shape or irregular aspidasters (Figs. 22C–22E and 23D), (ii) two separate sizes of microrhabds vs. only one in E. corsicus/papulifer (Figs. 22G and 23K) and (iii) absence of 3-actin toxas (in the holotype), which are present in E. corsicus/papulifer, but quite rare (Fig. 23F), which makes this third character difficult to use. Following the two first characters mentioned, our material is not E. expletus. No 3-actin toxas were found in our material but since they can be quite rare, we may have missed them. Then our material shares some characters with E. corsicus and some with E. papulifer. As in E. corsicus, our material has rare orthotriaenes, irregular aspidasters (not in all our specimens), and larger aspidasters, toxas and microrhabds. As in E. papulifer, some of our specimens have mostly regular aspidasters and have 1-actin toxas. This suggests that the irregular aspidasters may be due to the environment and therefore would not be a reliable specific character. SEM observations of toxas in the holotype of E. papulifer revealed spines at the tips of the actines (Figs. 23I–23J), which were not present in our material (Figs. 22I–22J) or in E. expletus from Rockall Bank (data not shown). This new character needs further confirmation in order to consider it as diagnostic. All things considered, our specimens were identified as E. corsicus, the second report of E. corsicus after its original description. However, the validity of E. corsicus originally based on a single tiny specimen from the same locality as E. papulifer remains dubious and would need to be further tested with additional material, and more importantly genetic sequences from other Mediterranean populations. In the Mediterranean Sea, E. expletus has been reported from shallow water caves off Marseille, France (Pouliquen, 1972). Since the Marseille specimens have smaller aspidasters, toxas and oxeas than E. expletus and only one size of microrhabds, we propose that they are instead E. papulifer, thus restricting E. expletus to the North Atlantic. More specimens and sequences of E. expletus, E. papulifer and E. corsicus are now necessary to test the robustness of these spicular characters. The COI tree currently confirms that our E. corsicus and E. expletus are different but very close species with only a 3 bp difference (Fig. 19A). The phylogenetic position of E. corsicus/E. expletus within the Erylinae is still ambiguous and not supported (Fig. 19).

Material examined

UPSZMC 190921-190922, field#i142_A-i146_4, St. 51 (INTEMARES0718), MaC (EB), 128 m, beam trawl, coll. J. A. Díaz; UPSZMC 190924, field#i524_b, St. 17 (INTEMARES0720), MaC (AM), 113 m, beam trawl, coll. J. A. Díaz; UPSZMC 190926-190927 and UPSZMC 190920, field#i528-529 and field#i530, St. 18 (INTEMARES0720), MaC (AM), 112 m, coll. J. A. Díaz, beam trawl; UPSZMC 190928, field#POR469, St. 219 (MEDITSES052017), Son Bou (South-west of Menorca), 65 m, GOC-73, coll. J. A. Díaz; UPSZMC 190929, field#POR932_1, St. 74 (MEDITSES052020), Favàritx (Northeast off Menorca), 75 m, GOC-73, coll. J. A. Díaz; UPSZMC 190930, field#POR975, St. 78 (MEDITS052020), Ciutadella (West off Menorca), 56 m, GOC-73, coll. J. A. Díaz; UPSZMC 190931, field#POR1141, St. 185 (MEDITS052021), Sa Costera (North of Mallorca), 61 m, GOC-73, coll. J. A. Díaz; UPSZMC 190932, field#POR1253, St. 02 (MEDITS0521_PITIUSSES), South of Formentera, 54 m, GOC-73, coll. J. A. Díaz.

Comparative material

Penares euastrum, holotype, MNHN Schmidt collection#76, wet specimen, La Calle, Algeria, coll: H. de Lacaze-Duthiers.

Outer morphology

Massive, irregular lobose sponge, reaching about 12 cm in maximum diameter (Fig. 24A). Surface dark brown except at the basal area, where it is whitish. Pores are located on white circular areas throughout the body, giving a characteristic appearance. Circular oscula placed apically. Cortex <1 mm.

Spicules

Dichotriaenes (Fig. 24B), scarce, whose cladome has longer protoclads than deuteroclads. Rhabdome: 344-387/19-36 , protoclad: 148-206/20-45 , deuteroclad: 46-220/11-36 µm. In the observation of several P. euastrum specimens we have observed modified dichotriaenes with one, two and three (orthotriaenes) of its clads non-bifurcated. Also, we observed one rhabdome modified to a rounded end (Fig. 24B).

Oxeas, with variable morphology; some are straight while others are bend or abruptly curved. Tips may be pointed, stepped or rounded, resembling strongyles, measuring 465-1390/4-38 µm.

Aspidasters (Figs. 24C–24E), shape variable between individuals, some showing a predominance of discoid to oval (POR932_1) or only oval (i529, i530, POR469, POR1253), sometimes irregular (Fig. 24E). Stumps in the hilum area found only in POR932_1, having several stumps per aspidaster and showing a bulbous shape, type 3 rosettes sensu Cárdenas (2020), measuring 97-182/74-142 µm (maximum diameter/min diameter).

Oxyasters I (Fig. 24F), with 2–10 actines, smooth or with few spines, barely visible with an optical microscope (Fig. 24G). There is a relationship between size and number of actines, being in general larger those with less actines. Measuring 25–74 µm in diameter.

Oxyasters II (Figs. 24H–24I), with >10 actines, smooth or with few spines, not visible to the optical microscope, measuring 8–28 µm in diameter.

Microrhabds (Fig. 24J), smooth, centrotylote, with blunt ends 27-74/2-6 µm.

Ecology and distribution

Common species usually found at mesophotic depths: on fishing grounds of the platform, on Peyssonnelia and rhodolith bottoms and at the summit of the AM and the EB seamounts. May have several sponge epibiont species like Timea sp. and some poecilosclerids.

Genetics

The Folmer short miniCOI was obtained from i530 and POR932_1 (SBP#2683 and SPB#2684), while 28S (C1-C2) (ON133852) was obtained from i530.

Taxonomic remarks

See remarks below for Penares cavernensis sp. nov.

Etymology

Penares cavernensis, from the Latin cavernae (cave), referring to the habitat where the species was discovered.

Material examined

Holotype: UPSZTY 190918, field#LIT55, Cova Cala Sa Nau, east of Mallorca, 3–4 m, scuba diving, coll. J. A. Díaz & J. Cabot.

Paratypes: UPSZTY 190917, field#LIT45, Cova Cala Sa Nau, east of Mallorca, 3–4 m, scuba diving, coll. J. A. Díaz; UPSZTY 190919, field#LIT65, Caló des Monjo, west of Mallorca, 6 m, scuba diving, coll. J. A. Díaz & A. Frank.

Outer morphology

Massive-encrusting, lobated sponges (Figs. 3F and 25A), about 2–5 cm in maximum diameter, and 1–3 cm in height. Surface color can be dark brown to beige, whitish rim around the oscula. Choanosome whitish. Same coloration in life and after fixation in ethanol. Several circular oscula located apically, 1–3 mm in diameter, smaller whitish pores are visible with the naked eye. No hispidation present, surface smooth to the touch. Very thin cortex present, about 0.4 mm. Hard but slightly compressible consistency. The skin wrinkles after EtOH preservation. In the holotype, a large cloaca can be observed below the main apical oscula.

Spicules

Dichotriaenes (Fig. 25B), scarce, morphology and abundance varies with the individuals. They are very scarce in specimens found in Cala sa Nau (LIT45 and LIT55), with robust rhabdome and cladome. In the specimen from Caló des Monjo (LIT65), dichotriaenes are more abundant but with thinner rhabdome and cladome, also showing teratogenic modifications and aborted actines. Overall measuring: rhabdome 407-515/7-27 µm, protoclad: 86-213/9-22 µm, deuteroclad: 49-139/7-15 µm.

Orthotriaenes (not shown), scarce and mostly broken, with long clads slightly curved inwards, rhabdome: 289-548/9-18 µm, cladi: 109-291/7-17 µm.

Oxeas (Fig. 25C), straight or curved, with slightly stepped tips, 426-1,259/5-22 µm.

Aspidasters (Figs. 25D–25F), discoid to oval (the majority are oval), very thin, several having one or more characteristic stump(s) in the hilum area (Fig. 25E), which are isolated actines; type 3 rosettes sensu Cárdenas (2020), measuring 67-135/54-98 µm (maximum diameter/min diameter). Translucent appearance with an optical microscope due to their extreme thinness.

Oxyasters I (Figs. 25G–25J), with very thin actines, smooth or minutely spined, 2–5 actines, 32–68 µm.

Oxyasters II (Figs. 25K–25LL), many spined actines, (Fig. 25LL), 6–14 µm.

Microrhabds (Fig. 25M), smooth, curved and centrotylote, with blunt ends, 28-58/1-3 µm.

Ecology and distribution

Found on vertical walls of shallow caves, Mallorca Island, at 3–4 m depth.

Genetics

The Folmer miniCOI has been obtained from the holotype (SPB#2685).

Taxonomic remarks on Penares euastrum and Penares cavernensis sp. nov.

Penares euastrum is a former Erylus species (i.e., Geodiidae with aspidasters) that was moved to the genus Penares by Cárdenas et al. (2010) based on its close phylogenetic position with P. helleri, the type species of the genus Penares. The Schmidt holotype of P. euastrum from Algeria was here examined, and all the spicules remeasured for the first time (Figs. 26A–26I, Table 10). SEM pictures notably showed that the oxyasters I and II can be spiny to slightly spiny (Figs. 26D–26H) and not always smooth as most of the previous descriptions suggested. Oxyasters of our P. euastrum material (Figs. 24F–24I) were similar, smooth to spiny. P. euastrum is a common Mediterranean species, regularly reported from caves to mesophotic depths. We had originally identified our cave specimens from Mallorca as P. euastrum, however, while the two miniCOI from our mesophotic specimens (i530 and POR932_1) were a perfect match to the COI of a mesophotic P. euastrum from Italy (MT815827), the miniCOI sequence from our cave specimen LIT55 showed a 2 bp difference with the Italian and the Balearic islands mesophotic specimens. This is a significant difference considering how conserved COI is in demosponges (Schuster et al., 2017), the geographical closeness of the two habitats and the short size (130 bp) of the miniCOI. In fact, having 2 or more bp differences in a full (about 640 bp) COI sequence is generally considered indicative of different species. External morphology differences are few and subtle but consistent: mesophotic specimens are larger and with a darker coloration while cave specimens are encrusting to massive-encrusting and of a paler color. Although spicules are very similar with overlapping size ranges (Table 10), overall cave specimen spicules are thinner and smaller. Aspidasters of P. euastrum are slightly larger and wider (97-182/74-142 µm vs 67-135/54-96) with rosettes more densely disposed (Figs. 24D, 25E and 26C). Besides, aspidasters of P. cavernensis sp. nov. are more translucid under the light microscopy than those of P. euastrum, probably because of having less densely arranged rosettes and being thinner (although we did not manage to get thickness measurements of the aspidasters). Moreover, aspidasters from cave specimens are occasionally discoid (much rarer in mesophotic specimens) and commonly have ectopic actines in the hilum area, which is quite distinct (Figs. 25D–25F). Ectopic actines have been found in all three specimens of P. cavernensis sp. nov. examined but only in one of the nine Balearic Islands P. euastrum studied (specimen POR932_1), and are absent in the holotype (see Table 10, Figs. 24 and 26). The holotype of P. euastrum from Algeria is a large mesophotic specimen (3.5 × 3 cm) (Fig. 26); its spicule measurements are closer to our mesophotic specimens, which are therefore formally identified as P. euastrum. Topsent (1928) had also distinguished a variety of P. euastrum called Erylus aspidodiscus, which, based on its morphological similarity with P. euastrum should actually be moved to the genus Penares as well, as Penares aspidodiscus comb. nov. This species was unfortunately based on a single specimen, from mesophotic depths close to Monaco (123 m) with spicule sizes closer to those of the holotype of P. euastrum. It was characterized by having only discoid aspidasters, commonly with ectopic actines, just like in P. cavernensis sp. nov. Since one of our P. euastrum also showed ectopic actins (POR932_1), it seems that this character may be shared by P. euastrum and P. cavernensis sp. nov., although being much more common in the second. Also, P. cavernensis sp. nov. does not have exclusively discoidal aspidasters like in P. aspidodiscus, in fact the majority of the aspidasters are oval in P. cavernensis sp. nov. P. aspidodiscus thus shares some characters from P. euastrum and some from P. cavernensis sp. nov. and is of uncertain taxonomic status at this point.

Regarding 28S, we only manage to get a C1-C2 fragment of mesophotic P. euastrum (i530). Interestingly, it showed 3 bp difference with another mesophotic P. euastrum from the ‘Banc de l’Esquine’ (off La Ciotat, France, AF062600), which may indicate that there is even more diversity than suspected in the P. euastrum complex.

Material examined

UPSZMC 190913, field#i143_G, St. 51, INTEMARES0718, MaC (EB), 128 m, beam trawl, coll. F. Ordines; UPSZMC 190914-15, field#i315_A-315_B, St. 124, (INTEMARES1019), MaC (EB), 152 m, beam trawl, coll. J. A. Díaz.

Comparative material

Penares candidatus, holotype, MNHN Schmidt collection#80, wet specimen, Algeria, ‘Exploration Scientifique de l’Algérie’, 1842.

Outer morphology

Massive, lobated sponges (Fig. 27A), up to 5 cm in maximal diameter. Greenish to dirty gold in life, light to dark brown with a faint yellowish tinge after preservation in ethanol. Choanosome beige after ethanol fixation, cavernous. Cortex about 0.2 mm thick. Hard consistency, only slightly compressible. No hispidation, smooth surface and heavily wrinkled. Several circular oscula, 2–3 mm in diameter, located at the apex of the lobules and spread over the body. A characteristic dark rim surrounds the oscula (Fig. 27B). Pores inconspicuous.

Spicules

Dichotriaenes (Figs. 27D–27E), with a short, straight, and sharp rhabdome, measuring 197-310/24-39 µm (N = 5). Cladome with cladus of approximately the same length as rhabdome. Protoclads measuring 52-97/9-49 µm, deuteroclads measuring 37-258/6-40 µm.

Oxeas (Fig. 27C), robust, fusiform, slightly bent, with sharp tips, 402-1,156/5-26 µm.

Microrhabds (Figs. 27F–24H), smooth, curved and with sharp ends (Fig. 27H), measuring 35-216/2-12 µm. Some are very slightly centrotylote, more clear in small spicules.

Ecology notes

Species found at two stations, both at the EB summit: one at 132 m and the other one at 152 m. Both stations corresponded to sponge grounds, with a great sponge diversity and abundance, including large tetractinellids like Discodermia polymorpha (Pisera & Vacelet, 2011) or Erylus spp., small axinellids and haplosclerids like Petrosia (Petrosia) ficiformis (Poiret, 1789) and Petrosia (Strongylophora) vansoesti (Boury-Esnault, Pansini & Uriz, 1994).

Genetics

COI (ON130543) and 28S (C1-C2) (ON133857) sequences of specimen i315_B were obtained. The 28S sequence represents the first published sequence of this marker for the species.

Taxonomic remarks

This is the first report of P. candidatus in the Balearic Islands, and the deepest for the species, extending its bathymetric range down to 152 m. The species is also reported for the first time on a seamount. P. candidatus is easily recognizable by its external appearance and the absence of oxyasters, a unique feature amongst North Atlantic/Mediterranean Penares species. However, there is some heterogeneity in the literature regarding spicule sizes (Table 11). Microrhabds have smaller size ranges in Bibiloni (1981) (50–100/4–5 µm), and Cruz (2002) (40–60 µm) and are never centrotylote in contrast to the holotype (50–250 µm), specimens from Banyuls, France (30–250 µm) (Topsent, 1894) and our material (35-216/2-12 µm). A similar tendency is observed with the oxeas, which are smaller in Bibiloni and Cruz (450-500/10 µm and 320–700 µm, respectively) than in our specimens (825-1,200/23-25 µm and 402-1,156/5-26 µm, respectively) and the rest of the literature (Table 11). These discrepancies may be explained by the depth, and thus silica availability, where individuals were collected, as Bibiloni specimens came from shallower waters (Blanes, 6 m), while possibly the holotype and our specimens were collected at greater depths (>30 m, and up to 152 m). Another explanation is that P. candidatus represents a species complex between some shallow and deeper populations, as for Penares helleri and Penares euastrum (see discussion above and below), but this is currently not supported by COI sequences, which are 100% identical for specimens from 5 m (Berlengas Islands, Portugal, HM592719) and our mesophotic specimens.

Material examined

UPSZMC 190933, field#i142_D, St. 51 (INTEMARES0718), MaC (EB), 128 m, beam trawl, coll. F. Ordines; UPSZMC 190934, field#i152, St. 52 (INTEMARES0718), MaC (EB), 107–110 m, rock dredge, coll. F. Ordines; UPSZMC 190935, field#i233, St. 48 (INTEMARES1019), MaC (AM), 123 m, beam trawl, coll. J. A. Díaz; UPSZMC 190936, field#i739, St. 52 (INTEMARES0720), MaC (EB), 300-330 m, beam trawl, coll. J. A. Díaz; UPSZMC 190937, field#POR946, St. 76 (MEDITS2020), s’Algar (east of Menorca), 131 m, GOC-73, coll. J. A. Díaz.

Comparative material

Penares sp.1, PC325, 3PP cave, La Ciotat, France, scuba-diving, 28S: AF062598, originally identified as P. helleri in Chombard, Boury-Esnault & Tillier (1998).

Penares sp.2, ZMAPOR 21658, field#FLW.06.48, Gruta Enchareus (cave), Flores, Azores, 12 m, scuba-diving, 11 July 2006, coll: J. R. Xavier, 28S: HM592828, originally identified as P. helleri in Cárdenas et al. (2011).

Outer morphology

Massive ovoid and lobated sponges (Fig. 28A), colors alive are dark brown on its upper side, progressively fading to light brown or whitish in its basal area. Some specimens are entirely whittish (e.g., i739). Colors after ethanol fixation are brown (basal area) to dark brown (apical area). Surface visually smooth but rough to the touch, slightly compressible. Cortex 0.5–0.7 mm thick clearly distinguishable. Choanosome fleshy, light brown after ethanol fixation and showing a well developed aquiferous system. 1–3 small oscules (mm size) irregularly distributed, uniporal pores, much smaller (<1 mm) grouped in some areas. In some specimens the ethanol acquires a dirty gold coloration during the fixation.

Spicules

Dichotriaenes (Figs. 28B–28C), abundant, rhabdome: 236-293/24-62 µm, longer deuteroclad than protoclad: protoclad 51-103/18-59 µm, deuteroclad 99-265/17-50 µm.

Oxeas, robust, fusiform and slightly bent, 684-1,641/10-42 µm.

Oxyasters (Figs. 28D–28H), microspined (Fig. 28E), with 4–10 actines, 11–56 µm.

Microrhabds (Figs. 28I–28K), on a wide but continuous size range, smooth, can have mucronated ends, which are only visible under SEM microscope (Fig. 28K), centrotylote, measuring 19-171/2-8 µm.

Ecology and distribution

Species found in sedimentary bottoms from mesophotic to upper bathyal depths (100–300 m). It is common at the fishing grounds of the upper slope of the Mallorca and Menorca shelf and at the summit and slopes of the AM and the EB.

Genetics

The COI (ON130538, ON130539) and 28S (C1-C2) fragments (ON133859, ON133858) have been obtained from specimens i739 (EB) and POR946 (s’Agar). These are the first sequences from mesophotic P. helleri.

Taxonomic remarks

See below taxonomic remarks for Penares isabellae sp. nov.

Etymology

Named after Isabel Fullana Riera (a.k.a. Bel Fullana), a Mallorcan painter.

Material examined

Holotype: UPSZTY 190938, field#LIT48, Cova Cala Sa Nau, east of Mallorca, 3–4 m, scuba diving, coll. J. A. Díaz.

Paratypes: UPSZTY 190939, field#LIT40_1, Cova Cala Sa Nau, east of Mallorca, 3–4 m, scuba diving, coll. J. A. Díaz; UPSZTY 190940, field#LIT66, Caló des Monjo, west of Mallorca, 6 m, scuba diving, coll. J. A. Díaz and A. Frank.

Outer morphology

Massive-encrusting sponges (Figs. 4E and 29A), 8–9 cm in length (sometimes more), 0.5 cm in height. Whitish gray in life and after preservation in ethanol. Surface smooth, slightly rough to the touch. Flexible and compressible. Cortex less than 0.5 mm, clearly distinguishable. Choanosome whitish. Circular oscula 3–5 mm in diameter (Fig. 29B). Minute pores distinguishable with the naked eye when alive, located on depressed areas present all over the surface (Fig. 29C), due to body contraction after fixation, these areas are more difficult to see.

Spicules

Dichotriaenes (Fig. 29D), very scarce (only found in the holotype), with short rhabdome, cladome may show aberrant cladi. Rhabdome: 169 (N = 1)/15-22 (N = 2) µm, protoclad 38-59/16-19 (N = 2) µm, deuteroclad 105-114/12-17 (N = 2) µm.

Oxeas (Fig. 29E), slightly bent, tips may be pointed or blunt (i.e., modifications to style and strongyle), 353-913/6-17 µm.

Oxyasters (Figs. 29F–29J), with up to 17–18 microspined actines (Fig. 29FG). Overall measuring 9–30 µm.

Microrhabds (Figs. 29K–29M), very abundant, on a wide but continuous size range, measuring 31-168/2-6 µm, smooth, slightly bent (rarely abruptly bent), with stepped tips (Fig. 29L) that may be mucronated (Fig. 29M).

Ecology and distribution

This species was discovered in shallow caves, the two exact same localities/habitat as P. cavernensis sp. nov. (Fig. 3I) (see above).

Genetics

The Folmer COI (ON130540, ON130541, ON130542) were obtained from holotype LIT48 and paratypes LIT40_1 and LIT66, while a 28S (C1-C2) fragment was obtained from the paratype LIT40_1 (ON133860).

Taxonomic remarks on Penares helleri and Penares isabellae sp. nov.

Penares helleri is a common Mediterranean species found at both infralittoral caves and the mesophotic zone with one report from the Atlantic side, but close to Gibraltar at 521 m depth, its current deepest record (Boury-Esnault, Pansini & Uriz, 1994). Cave specimens are reported to be massive or encrusting while mesophotic ones tend to be reported as massive. Otherwise cave and open-sea specimens share the same spicular set; smooth microrhabds, oxyasters, oxeas and dichotriaenes. However, reported spicular sizes are very heterogeneous (Table 12), with some specimens having longer and/or thicker microrhabds/oxeas. Also, dichotriaenes seems to be very common in some specimens, but very scarce in others. The mentioned differences are usually explained in the literature (Sarà, 1961) by differences in habitat conditions, for example, currents may affect the growing morphology, leading to encrusting versus massive specimens and depth may influence nutrient availability, affecting the silica content and thus spicule development. Our results reveal that P. helleri is a species complex with at least four cryptic but genetically different species, a fact that challenges the use of ecophysiological traits to explain morphological differences.

A first species (POR946, i739, i152 and i233) corresponds to the true P. helleri, as macroscopic morphology (massive), spicules (robust and abundant dichotriaenes, thick microxeas and large oxyasters with 4–10 actines) and habitat (mesophotic to upper bathyal) matches the holotype.

A second species was revealed by our molecular markers. The full Folmer COI clearly suggested that our cave specimens (LIT40_1, LIT48 and LIT66) was a different species with a significant 7 bp. difference with the mesophotic P. helleri; 28S (C1-C2) shows a 1 bp. difference, knowing that the C1-D1-C2 fragment is quite conserved (we are missing the much more variable D2 part, which is ideal to discriminate species). This whitish encrusting species had thinner microxeas, smaller oxyasters with >10 actines and very scarce dichotriaenes. This species actually resembles P. helleri forma subtilis Sarà & Sribelli, 1960 found on rocky bottoms in the Gulf of Naples (20–25 m) and in caves. P. subtilis is described as whitish encrusting, small, with weak microxeas and with very scarce dichotriaenes, just as in our cave specimens. In fact, in her thesis Bibiloni (1990) assigned cave P. helleri from Mallorca to the subtilis variety. However, the description of P. subtilis is incomplete (the size of oxyasters is notably missing) and it suggests there are two distinct sizes of microrhabds (32–80 µm and 103–213 µm) vs. only one in our material (31–165 µm). Furthermore, the type material of P. helleri var. subtilis (specimens #773:2 and #783:2) could not be revised here, it is unfortunately missing and presumably lost (pers. comm., ‘Stazione Zoologica Anton Dohrn’, Naples, curator Dr. Andrea Travaglini; ‘Museo Civico di Storia Naturale “G. Doria”’, Genoa, curator Dr. Maria Tavano; ‘Museo zoologico, Naples University’, curator Dr. Roberta Improta; ‘Museo di zoologia’, Bari, Dr. Giovanni Scillitani). Instead of risking to mis-identify this common Mallorcan species, we decided to create P. isabellae sp. nov., before someone can sequence several P. helleri from the Gaiola area in the Gulf of Naples for comparison.

The fourth species in this complex was originally identified as P. helleri in Chombard, Boury-Esnault & Tillier (1998) but is here again revealed different thanks to 28S. The full 28S (C1-D2) fragment of a specimen from the 3PP cave (La Ciotat, France, AF062598) has surprising 12 bp. and 13 bp differences with respectively our mesophotic P. helleri and P. isabellae sp. nov. This specimen (PC325) was re-examined here, it is a cave specimen with many dichotriaenes, large oxyasters (13–47 µm) as in P. helleri and non-centrotylote shorter microxeas (25–105/2–4 µm), as in P. isabellae sp. nov. Interestingly, it has common double-bent microxeas, which were never observed in our Balearic specimens. More “P. helleri” from the 3PP cave need to be sequenced in order to confirm and understand this potential new species. Meanwhile, this brings additional data to possible cryptic cave faunas. Marine caves are often considered isolated habitats, with cave fauna being poorly connected and having low gene flow, a fact that promotes speciation and high levels of endemism (Juan et al., 2010). Patterns of genetic connectivity between cave sponges are understudied, but some works pointed to a high isolation pattern (Muricy et al., 1996).

Finally, a fifth species is represented by the specimen ZMAPOR 21658 from another shallow cave, this time on Flores Island, in the Azores. This Flores P. helleri is genetically much closer to the 3PP cave specimen than to the Balearic specimens. Indeed, this Flores specimen and the 3PP cave specimen group in a well supported clade, along with a sequence of Penares sclerobesa (Topsent, 1904) , while the Balearic specimen seem to group closer to the Erylus mamillaris/discophorus/deficiens complex (Fig. 19). The Flores specimen also has double-bent microrhabds, like those observed in the 3PP cave specimen, which thus may be a good character for future discrimination.

To conclude, the discovery of a polyphyletic P. helleri brings new taxonomic issues because if indeed the type of P. helleri from the Adriatic Sea is conspecific with our mesophotic specimens, then the phylogenetic relationship P. helleri/P. euastrum (based on the position of the 3PP “P. helleri” sequence) is no more. This means that the reallocation of P. euastrum (and P. cavernensis sp. nov.) to the genus Penares would not be justified. Our current phylogenetic COI and 28S trees now reveal a complex mix of Erylus and Penares species, with mostly poorly-supported nodes, thus begging for better and additional markers.

Etymology

Named matrix in analogy with something that harbors other elements, because the species always incorporates all kinds of substrata on its body, and also is a substrate to many other sponge epibionts; also after the 1999 film by the Wachowski sisters.

Material examined

Holotype: UPSZTY 190881, field#i577_1, St. 21 (INTEMARES0720), MaC (AM), 109 m, beam trawl, coll. J. A. Díaz.

Paratypes: UPSZTY 190876, field#i146_1A, St. 51 (INTEMARES0718), MaC (EB), 128 m beam trawl, coll. F. Ordines; UPSZTY 190879, field#i244_B1, St. 50 (INTEMARES1019), MaC (AM), 102 m, beam trawl, coll. J. A. Díaz; UPSZTY 190880, field#i545, St. 19 (INTEMARES0720), MaC (AM), 111–94 m, rock dredge, coll. J. A. Díaz; UPSZTY 190882, field#i577_2, St. 21 (INTEMARES0720), MaC (AM), 109 m, beam trawl, coll. J. A. Díaz.

Comparative material

Geodia canaliculata Schmidt, 1868, holotype, MNHN-DT750, MNHN Schmidt collection#61, wet specimen and slide, Algeria, ‘Exploration Scientifique de l’Algérie’, 1842.

Outer morphology

Ramose sponge, growing repent and incorporating all kinds of gravels from the substrate. In life, light brown (Fig. 30A), dark brown after ethanol fixation. Choanosome dirty beige after ethanol fixation. The holotype is 7 × 2.5 × .5 cm. Surface smooth to hispid. Hard but breakable consistency. Small uniporal openings barely visible, often at the tips of the lobes, probably oscules, but maybe also pores. Cortex 0.3–0.5 mm thick. The species is usually covered with other sponges, like Hexadella sp. (Fig. 30A), haplosclerids, poecilosclerids and an orange encrusting Timea sp.

Spicules

Oxeas (Fig. 30B), slightly curved and fusiform, 566-1,949/7-28 µm.

Plagiotriaenes (Fig. 30C), rhabdome stout and long, straight to slightly bent, with a sharp end. Smaller ones with a marked swelling below the cladome. Clads are usually disposed in a 60–70° angle with the rabdome, they are short and triangular, straight or slightly curved upwards, some aberrant or underdeveloped. Rhabdome: 728-1,623/9-36 µm, cladi: 29-179/7-29 µm.

Sterrasters (Figs. 30D–30F), spherical, smooth rosettes (Fig. 30F), measuring 26–71 µm.

Spheroxyasters (Figs. 30G–30I), smooth, with triangular actines. Young spherasters have few spines (Fig. 30G) while fully developed spheroxyasters have many spines concentrated at the tips of the actines (Figs. 30H–30I), resembling the rosettes of the sterrasters. Measuring 11–33 µm in diameter.

Oxyasters I (Figs. 30J–30L), 2–7 actines. The less actines they have, the larger they are. Actines are essentially smooth with a few occasional small spines. Measuring 36–92 µm in diameter.

Oxyasters II (Figs. 30LL–30N), uncommon, with 4–11 slightly microspined actines (Fig. 30N), 16–37 µm in diameter.

Ecology and distribution

Always found associated with rhodolith beds at the summit of the AM and the EB seamounts, where it can be very abundant and significantly contribute to the overall biomass in several kg per 100m2 (Díaz et al., 2024). Due to its relatively large size and high abundance, the species may play a role as habitat builder, providing shelter to smaller associated fauna. It is always overgrown by other sponges, mostly Hexadella sp., Timea sp. and several haplosclerids.

Genetics

COI (ON130523, ON130524, ON130525) and 28S (C1-C2) (ON133885, ON133886, ON133887) were obtained from i146_1A, i244_B1, i577_1 (holotype) and i577_2.

Taxonomic remarks

Geodia matrix sp. nov. appears to have uniporal oscules/pores, which is only found in a few temperate Atlanto-Mediterranean species, previously grouped in the genus Isops: Geodia geodina (Schmidt, 1868), G. canaliculata, Geodia globus Schmidt, 1870 and Geodia pachydermata Sollas, 1886. Two more species should be considered for comparison, since their opening morphologies are unknown: Geodia echinastrella Topsent, 1904 and Geodia spherastrella Topsent, 1904.

Geodia geodina has smaller spheroxyasters (10–18 µm), which do not develop spines at their tips (cf. below). G. globus, is a poorly described species only reported once from Portugal but a re-description of the type suggests it has strongylasters (9 µm), and no spherasters (Burton, 1946). G. pachydermata has a very typical external warty appearance, and much larger sterrasters (>200 µm). G. echinastrella from the Azores, resembles G. matrix sp. nov. in having similar size spherical sterrasters (47–50 µm) and similar sized smooth oxyasters (22–26 µm), however, it has smaller spherasters (15–18 µm) and is lacking the large oxyaster category. G. spherastrella also from the Azores has larger and ellipsoidal sterrasters (90–110 µm), smaller spherasters (14–16 µm) and only one size of spiny oxyasters (25 µm).

Of all known Atlanto-Mediterranean species, G. matrix sp. nov. is closest to G. canaliculata, a poorly-known species reported twice from the coast of Algeria more than 100 years ago (Schmidt, 1868; Topsent, 1901). The holotype (Fig. 31) was revised by Topsent (1938) but re-examined here and thick sections were made (Figs. 31D–31G): it is a massive sponge with uniporal openings (Topsent also could not distinguish oscules from pores), which explains its placement in the former genus Isops by Topsent (1901). It can be added to the detailed description given by Topsent (1938) that the holotype is overgrown by an encrusting poecilosclerid with hymedesmoid skeleton (reddish color, Figs. 31D–31E) and the cortex is 0.25–0.3 mm thick (measured on thick sections). On our sections, oxyasters are essentially found just below the cortex (moderate abundance) while spheroxyasters (at different stages of development) are found throughout the choanosome (low abundance, Fig. 31F), with rare presence in the cortex. More surprisingly, slightly flexuous microtoxas were found, often slightly centotylote (160–220 µm long; Fig. 31G); they look like flattened flexuous microtoxas such as the ones from the encrusting sponge D. annexa, a common epibiont in the Atlanto-Mediterranean Sea region. They clearly appeared on the thick sections we made, fairly abundant throughout the choanosome, and with no particular orientation. Such spicules have never been observed before in Geodia species so we consider them to be probably foreign, but we also note that there are no other spicule contamination in the choanosome than these flexuous microtoxas (i.e., no other typical D. annexa spicules). As a result, despite some shared characters with G. canaliculata (irregular lobose/ramose shape with numerous foreign bodies and epibionts, uniporal oscules/pores, sterraster size, characteristic spherasters), G. matrix sp. nov. has two significant differences: (i) well-developed regular plagiotriaenes (vs. aborted cladomes), and (ii) large and small oxyasters (vs. only small oxyasters).

Attention should be paid to the fact that G. matrix sp. nov. tends to be overgrown by Timea sp. This species has spheroxyasters of a similar shape and size (6-19-30) as those of G. matrix sp. nov. and so they mixed in our first spicule preparations. The subtle difference is that spheroxyasters of G. matrix sp. nov. have a larger centrum and heavily spined actines, especially at their tips. Timea sp. has a characteristic orange color in life but it turns brownish after ethanol fixation, so its presence is hard to notice once specimens are fixed. After carefully digesting separately clean cortex and choanosome, we observed that Geodia spheroxyasters are much more abundant in the cortex, suggesting that spheroxyasters could be ectocortical or located just below the cortex. However, we could not confirm this as in the thick sections the cortex was always covered by the Timea sp. We further note that spheroxyasters are overall much more abundant in G. matrix sp. nov. than in the holotype of G. canaliculata; this characteristic should be confirmed with new specimens of G. canaliculata. According to their similar spicules, G. canaliculata and G. matrix sp. nov. are undoubtedly phylogenetically close species. It is also clear that the typical spheroxyasters in both species are homologous. There has been some confusion regarding these spheroxyasters in G. canaliculata (Topsent, 1901; Topsent, 1938), which SEM observations of G. matrix sp. nov. have helped us to understand. Schmidt (1868) and Topsent (1901) mention the presence of smaller sterrasters, with less actines than the regular ones, later thought to be a different category of spicule (Topsent, 1938). These are actually fully-developed spheroxyasters with spiny tips, as observed in G. matrix sp. nov. (Fig. 30I) or in other species but with a smaller size (G. pachydermata, G. spherastrella). The confusion arises because these spheroxyasters are almost as large as the sterrasters and sometimes mixed with them in the cortex.

COI and 28S tree suggest that G. matrix sp. nov. is related to species Geodia parva Hansen, 1885, Geodia phlegraei (Sollas, 1880), G. geodina, Rhabdastrella intermedia Wiedenmayer, 1989 and Rhabdastrella sp.1 South Africa. (Fig. 10). Those species do not belong to the three main Geodia clades (Cydonium^P, Depressiogeodia^P and Geodia^P)(Cárdenas et al., 2011), and may represent a fourth group in Geodia based on shared smooth oxyasters and presence of spheroxyasters, despite no bootstrap support for the clade at the moment.

Synonym

Not Geodia anceps in Sitjà et al. (2019) from the Gulf of Cadiz: renamed in the present paper as Geodia phlegraeioides sp. nov.

Material examined

UPSZMC 190862-190863, field#i140_A and field#i140_B, St. 51 (INTEMARES0718), MaC (EB), 135 m, beam trawl, coll. F. Ordines; UPSZMC 190868, field#i391_3, St. 158 (INTEMARES1019), MaC (EB), 146 m, beam trawl, coll. J. A. Díaz; UPSZMC 190870-190871, field#i575 and field#i576, St. 21 (INTEMARES0720), MaC (AM), 109 m, beam trawl, coll. J. A. Díaz; UPSZMC 190872, field#i708, St. 45 (INTEMARES0720), MaC (EB), 150 m, beam trawl, coll. J. A. Díaz.

Comparative material

Geodia geodina, lectotype (designated here), MNHN Schmidt collection#91 (large specimen), paralectotype (designated here), MNHN Schmidt collection#93 (small specimen), both specimens in the same jar registered under MNHN DT752, wet specimens, Algeria, ‘Exploration Scientifique de l’Algérie’, 1842; MNHN DCL728, spicule slide, East Gettysburg, Gorringe Bank, St. 149, trawl, 95 m (Lévi & Vacelet, 1958); MNHN (unregistered), field#JC46, Jean Charcot Madeira 1966, SW of Deserta Islands, 32°21′30″N, 16°30′18″W, 100–130 m, wet specimen, 18 July 1966, id: P. Cárdenas.

Geodia anceps, syntype, NHM 1955.3.24.1 (=RMNH POR0655), wet specimen, label from the ‘Rijksmuseum-Leiden’ saying ‘coll. no. 655, fragment of type sp.’, Vosmaer personal number N557, between Capri and Naples, 150–200 m, 12 Feb. 1891.

Outer morphology

Massive, globular (Fig. 32A), to ramose or lobated. Larger specimens up to 12 cm in maximum diameter. Grayish ocher in life, dark brown after ethanol fixation. Always paler on the lower side of the body (protected from the light). Surface smooth to hispid, smooth to the touch. Hard, only slightly compressible consistency. Cortex less than 0.5 mm thick, clearly distinguishable. Choanosome fleshy, whittish. Uniporal oscules are grouped on the top surface of specimens or at the top of lobes, always contracted on deck and after ethanol fixation. Uniporal pores gathered in depressed areas, visible to the naked eye in some specimens.

Spicules

Orthotriaenes (Fig. 32B), rhabdome straight and fusiform; the smaller ones have a triangular swelling at the joint with the cladome. Clads are disposed in a 100° angle with the rhabdome, some being slightly tortuous or having its tips curved inwards. A single cladome modification in the form of dichotriaene was observed. Rhabdome measuring 419-1947/11-54 µm, cladi measuring 71-470/11-50 µm.

Oxeas (Fig. 32C), slightly curved and fusiform, measuring 748-2,614/6-38 µm.

Anatriaenes, uncommon, with straight, fusiform rhabdome, 1,180-2,687/5-19 µm. Clads with short cladi, evenly curved, measuring 3-97/5-16 µm.

Sterrasters (Fig. 32D), spherical, with smooth rosettes having 4–12 conical rays (Fig. 32E), measuring 41–68 µm.

Oxyasters I (Figs. 32F–32I), large, smooth actines, usually with a few spines at its tips (Fig. 32G), 36–88 µm (2–10 actines).

Oxyasters II (Figs. 32J–32K), smaller than oxyasters I, smooth actines. Measuring 14–38 µm (∼6–19 actines).

Spherasters (Fig. 32L), with triangular actines that can be smooth or microspined at its tips, measuring 8–22 µm.

Ecology and distribution

Circalittoral species found at the summit of the AM and the EB, although reaching greater depths at the EB. It can be very abundant in some stations, suggesting that it is a habitat-forming species due to its large size and sometimes intricate body shape. It is used as a substrate by epibionts, especially other sponges, like Hexadella sp., Timea sp., as well as several haplosclerids like Haliclona poecillastroides (Vacelet, 1969).

Genetics

COI (ON130519, ON130520, ON130521, ON130522) has been obtained for i575, i708, i576 and i140_A, while 28S (C1-C2) (ON133879, ON133880) were obtained from i575 and i708.

Taxonomic remarks

See taxonomic remarks on G. geodina and G. phlegraeioides sp. nov. below.

Etymology

Named ‘phlegraeioides’ to highlight its phylogenetic and morphological closeness with Geodia phlegraei (Sollas, 1880) from boreal waters.

Type material

Holotype, MNCN/1.01/1026 (wet specimen), UPSZTY 190886 (thick sections and spicule slide), Almazán mud volcano, Gulf of Cadiz, 36°3′17.39″N, 7°19′43.20″W–36°3′36.6″N, 7°19′13.2″W (INDEMARES-CHICA), beam trawl, 894–896 m, 4 March 2011, coll: C. Farias, originally identified as G. anceps in Sitjà et al. (2019).

Paratype, UPSZTY 190887, field#DR15-972, Le Danois Bank, Cantabrian Sea, station DR-15, 650 m, 44°6′20.64″N, 5°9′16.2″W (SponGES0617), rock dredge, 23 June 2017, coll: P. Rios.

Other non-type material examined

Geodia phlegraeioides sp. nov., CPORCANT, DR10-490 and -500, Le Danois Bank, Cantabrian Sea, station DR-10, 44°6′4.8″N, 4°38′18″W, 541 m, 17 June 2017, rock dredge, Expedition SponGES0617, coll: P. Rios; CPORCANT DR15-869c, -862c, -882, same station as paratype; COLETA#5803 ( =PC566, spicule slide), Banc Princesse Alice, Azores, 37°49′19.2″N, 20°27′43.2″W, 432 m, 28 Feb. 2011, subglobular specimen, bycatch from long line demersal fishery TB/137/MBO/2011; COLETA#6243 (=PC567, spicule slide), Banco Voador, Azores, 37°28′58.8″N, 30°50′34.8″W, 418 m, 24 June 2010, fragment of a specimen, Coral Fish D33-V10, Palangre de fundo.

Comparative material

Geodia phlegraei, holotype, NHM 1910.1.1.840, Korsfjord, SW Bergen, Norway, 1878, 60°9′60″N, 5°10′0″E, 330 m, coll: Rev. A. M. Norman.

Geodia parva, holotype, ZMBN 100, spicule slide, unknown station, Norwegian North Sea Exp. 1876–78.

Outer morphology and skeleton

Massive, subspherical. External color whitish to light brown, alive and in ethanol; choanosome slightly more tanned. Surface is smooth to hispid. Uniporal oscules (up to ∼1 mm in diameter) and minute uniporal pores (∼0.2 mm in diameter). Cortex is ∼0.5 mm thick. Typical geodiid skeleton with ectocortical spheroxyasters and choanosomal oxyasters. The holotype is the largest specimen we have seen so far, 6.5 × 5 × 2.5 cm; for a detailed description of the holotype including its spicules, see Sitjà et al. (2019).

Spicules (Table 14)

Ortho- and dichotriaenes, robust, straight rhabdome 375-2770/11-70 µm, cladi slightly forward oriented, clads of orthotriaenes: 90–580 µm, protoclads: 28-378, deuteroclads: 53–580 µm.

Oxeas, curved and fusiform, 136-3,406/16-45 µm.

Sterrasters, spherical, with smooth rosettes, 41–99 µm.

Oxyasters, smooth, 11–93 µm.

Spheroxyasters, smooth with a few microspines at tips essentially, 9–32 µm. Specimens from the Azores tend to have a larger centrum.

Ecological notes

Some specimens were found growing on other sponges: DR15-869c was found on C. (C.) geodioides, the holotype was growing on a Pachastrella sp. The holotype was budding (Sitjà et al., 2019).

Genetics

COI were obtained from the holotype (OR045844), COLETA#5803 (OR045842), COLETA#6243 (OR045843) and DR15-869c (OR045845).

Taxonomic remarks on G. geodina and G. phlegraeioides sp. nov.

Our material from the EB and AM seamounts matches with G. geodina, a species described from Algeria (Schmidt, 1868) and subsequently reported in the Gulf of Naples (Pulitzer-Finali, 1972) and the Gorringe Bank (Lévi & Vacelet, 1958). We have compared our material with the type material from Schmidt (lectotype (#91) and paralectotype (#93), here designated), for which new thick sections and SEM was done (Fig. 33), and the spicules re-measured (Table 14). The only differences were in the smaller sizes of sterrasters (35–45 µm vs 41–68 µm) and of the orthotriaenes in the type material (Table 14). In this process, a close morphological similarity with G. anceps, a better-known Mediterranean Geodia, was also noticed. A syntype of G. anceps (NHM 1955.3.24.1) was examined, with new spicule and SEM preparations (Fig. 34). The type materials of G. geodina and G. anceps shared the same external morphology, spicule set and morphologies, with similar size ranges (Table 14). Again, the only noticeable difference was the smaller size of the sterrasters (35–45 µm vs. 50–70 µm) and of the orthotriaenes in the types of G. geodina, which may be a result of different depths or habitats, as it is known that sterraster and triaene size can be influenced by these parameters (Cárdenas & Rapp, 2013). All previous records of G. anceps, from the Gulf of Naples (Vosmaer, 1894; Pulitzer-Finali, 1972) or the Alboran Sea (Maldonado, 1992), as well as our material, come from mesophotic depths (70–200 m). Unfortunately, we have no locality/depth data for the types of G. geodina so it is impossible to test this hypothesis at the moment. Also, anatriaenes with small cladomes were occasionally found in the syntype of G. anceps (this study), in previous reports (Vosmaer, 1894; Pulitzer-Finali, 1972; Maldonado, 1992), in a NHM Schmidt type slide of G. geodina (Burton, 1946) as well as in our material from the Balearic Islands; however, no anatriaenes were found in our preparations of the G. geodina types. Not finding anatriaenes is not surprising since they can be rare and are often localized to certain parts of the sponge in some Geodia, so they can be easily overlooked. To conclude, we propose that G. anceps becomes a junior synonym of G. geodina.

One single G. anceps report is from the Atlantic: from the Gulf of Cadiz, 895 m depth (Sitjà et al., 2019). It was stated at the time that sterrasters were larger than in Mediterranean specimens (Sitjà et al., 2019). Besides that, we also noted other unusual features in this specimen, such as a mix of orthotriaenes and dichotriaenes, and very rare oxyasters I. Examination of additional Northeast Atlantic specimens originally identified as G. anceps from Le Danois Bank (Cantabrian Sea) at 650 m and two specimens collected South of the Azores (418–432 m) revealed these exact same characters: (i) larger sterrasters with an average size of 68–94 µm (vs. 40–61 µm in G. geodina), (ii) common dichotriaenes mixed with orthotriaenes, (iii) usually only one category of oxyasters with a continuum of sizes from 11 to 93 (vs. two separate sizes in G. geodina) and (iv) slightly larger spheroxyasters with an average size of 14–21 µm (vs. 13–15 µm in G. geodina). Although Sitjà et al. (2019) report two oxyaster categories, only one category was found during our re-examination of the specimen. Likewise, only one large oxyaster was found in specimen DR15-862c from Le Danois Bank, suggesting that a second larger category of oxyasters may occasionally be produced in Geodia phlegraeioides sp. nov. but they are very rare or integrated in a continuum with the first category. On the other hand, large oxyasters are always very common in G. geodina, and in a clear separate category. These morphological differences were supported genetically: COI Folmer sequences of these Atlantic specimens had a highly significant 18 bp difference with G. anceps. We propose the name Geodia phlegraeioides sp. nov. for this North Atlantic species resembling G. anceps and found deeper, down to the upper bathyal zone (418–895 m). The specimen described by Sitjà et al. (2019) is designated as the holotype, while one specimen from Le Danois Bank is designated as a paratype.

One deep report of G. anceps at 738–809 m, from the deep-sea coral reef off Cape Santa Maria di Leuca, Italy (Longo, Mastrototaro & Corriero, 2005) is possibly G. phlegraeioides sp. nov.: it has larger sterrasters (88–106 µm) and the common dichotriaenes. However, two categories of oxyasters are reported with similar size ranges, which is atypical and needs to be revised. If this was indeed G. phlegraeioides sp. nov., it would suggest that this North Atlantic species can also be found in the upper bathyal Mediterranean Sea, below the G. geodina zone in the mesophotic area. Likewise, G. geodina would also be found in the mesophotic zone in the Atlantic, above the deeper G. phlegraeioides sp. nov. In fact, a G. geodina spicule slide (MNHN-DCL 728) from the Gorringe Bank in the Northeast Atlantic, 95 m depth (Lévi & Vacelet, 1958) was re-measured and it fits clearly better the description of G. geodina than that of G. phlegraeioides sp. nov. (Table 14). This was independently confirmed by the examination of a MNHN large Geodia specimen (∼12 cm long) from Deserta Islands (close to Madeira) from 100–130 m, that we also identified as G. geodina.

With respect to our phylogenetic trees (Fig. 10), both G. geodina and G. phlegraeioides sp. nov. group with G. matrix sp. nov., former species of Isops (G. phlegraei, G. parva), which all share as previously said smooth oxyasters, spheroxyasters, and external morphology. In the COI tree, G. phlegraeioides sp. nov. was the sister species to the clade G. phlegraei +G. parva; there were respectively 5 and 6 bp differences between them. G. phlegraei is a boreal species found from 40 to 3000 m depth, while G. parva is its arctic counterpart (Cárdenas & Rapp, 2013), so G. phlegraeioides sp. nov. is the temperate version of G. phlegraei, thus explaining our choice for the species name. G. phlegraeioides sp. nov. seems to have a more irregular shape than G. phlegraei (which is subglobular when young and then bowl shaped when larger), with fewer oscules, less visible pores, it can have lots of dichotriaenes (G. phlegraei only has orthotriaenes), its sterrasters are round (vs. usually oval in G. phlegraei). G. geodina is also quite close to G. phlegraei but again, the shape is more irregular with smaller, less abundant oscules and invisible pores. G. geodina has smaller spherical sterrasters (vs. oval in G. phlegraei) with two sizes of oxyasters (vs. one size in G. phlegraei), and the large ones have fewer actines than in G. phlegraei.

Etymology

Named after Dr. Maria Antònia Bibiloni, who initiated the studies of sponges from the Balearic Islands, from 1982 to 1993.

Material examined

Holotype: UPSZTY 190857, field#i715_1, St. 45 (INTEMARES0720), MaC (EB), beam trawl, 150 m.

Paratypes: UPSZTY 190860-190861, field#i674-field#i675, St. 42 (INTEMARES0720), MaC (EB), 143–139 m, beam trawl, coll. J. A. Díaz; UPSZTY 190858-190859, field#i780-field#i781, St. 54 (INTEMARES0720), MaC (EB), 124–210 m, rock dredge, coll. J. A. Díaz.

Outer morphology

Found in two different morphologies: globose when living in sedimentary bottoms, with a lot of gravel incorporated on its body (i715_1; Fig. 35A), encrusting, slightly hemispherical, and free of foreign materia when growing on rocks (i674, i675, i780, i781; Fig. 35B). Relatively small, ∼2 cm in diameter. Whitish beige on deck and after ethanol fixation. Choanosome fleshy, pale beige after ethanol fixation. Surface hispid. Hard consistency, incompressible. Small cribriporal openings (about 0.2 mm) present all over the body, visible to the naked eye on deck. After preservation, they are also visible but not so patent. Cortex more than 0.5 mm thick. Typical geodiid skeleton with ectocortical spherasters, endocortical sterrasters and choanosomal oxyasters.

Spicules

Oxeas (Fig. 35C), thin and long, slightly curved, 1,058-2,765/12-32 µm. A second category of smaller oxeas were found in small numbers (N = 5) and only in i675 (Table 15). Those spicules are probably contamination, as its shape was similar to the isoactinal oxeas (oxeas II) of Craniella cf. cranium.

Orthotriaenes (Fig. 35D), with long, fusiform rhabd and slightly curved clads. The clads may end tipping downward. Sometimes a clad may be bifurcated. Juvenile stages show a bulbous swelling at the uppermost part of the rhabd. Rhabdome length: 594-2,224/14-59 µm, cladi 78-306/12-49 µm.

Protriaenes (not shown), only found in specimen i675. With long, straight or slightly curved rhabdome and pointed clads. Rhabdome: 1,472-2,389-3,446/9-11-14 µm (N = 10), cladome: 69-102-157/8-11-13 µm (N = 16).

Anatriaenes (not shown), scarce, rhabdome: 2,567 (N = 1)/2-10 µm, cladome: 16-86/2-14 µm.

Sterrasters (Fig. 35E), rounded, with warty rosettes (Fig. 35F), 39–69 µm in diameter.

Strongylasters-spherasters (Fig. 35G), with spines on the tips of the actines, 9–18 µm.

Oxyasters (Figs. 35H–35I), normally with 6–12 (sometimes up to 16), spined actines, 21–48 µm.

Ecological notes

Species found in the mesophotic to aphotic zone, in both sedimentary and rocky bottoms. In the sedimentary bottoms the species is collected as a rounded mass with many agglutinated sediments while when found on hard substrata it is encrusting, hidden in crevices of rocks. With that in mind, the spherical morphology is likely an artifact, as a consequence of the body contraction when collected, given that the small sediments that act as substrate in these bottoms are not heavy enough to avoid the body contraction. This, however, must be corroborated through direct observation of living specimens.

Genetics

Folmer COI (ON130526, ON130527 and ON130528) and 28S (C1-C2) (ON133882, ON133883, ON133881) sequences were obtained from i780, i674 and i715_1 (holotype). Two haplotypes were found for COI and 28S, which differed in 1 bp for both markers.

Taxonomic remarks

See taxonomic remarks on G. bibilonae sp. nov. and G. microsphaera sp. nov. below.

Etymology

The name ‘microsphaera’ refers to the small size of its sterrasters.

Material examined

Holotype: UPSZTY 190883, field#i589_2, St. 21 (INTEMARES0720), MaC (AM), 109 m, beam trawl, coll. J. A. Díaz.

Paratypes: UPSZTY 190884-190885, field#i589_3-i589_8, same station and collector as the holotype.

Comparative material

Geodia cydonium, ZMBN 85220, field#HC1, Hidden Cleft Cave, Brixham, Devon, England, collected above water at low tide, 1 Sept. 2008, colls. F. Crouch and C. Proctor, id: P. Cárdenas, COI: HM592715, 28S: HM592814.

Outer morphology

Small (0.8 cm in diameter), globose sponge almost entirely covered with gravels (Figs. 34A–34B). Color on deck not recorded, of whitish surface and choanosome after fixation in ethanol. Surface fairly hispid, hard consistency. Cortex patent, 0.5 cm thick. Openings not visible. Typical geodiid skeleton with ectocortical oxyspherasters, endocortical sterrasters and choanosomal oxyasters. Choanosome fleshy.

Spicules

Orthotriaenes (Figs. 36C–36E), fusiform rhabd and curved cladi. In juvenile forms, swellings present at the uppermost part of the rhabd (Fig. 34D). Rhabdome length: 311-1,558/5-29 µm, cladi: 34-332/4-27 µm.

Protriaenes (not shown), rare in i589_2 and i589_3 but common in i589_8. Long and thin, rhabdome straight, slightly curved or bent, cladome straight. Rhabdome: 1,394-3,504/7-13 µm, cladi: 44-118/5-13 µm.

Oxeas (not shown), fusiform, usually straight, some slightly bent or curved, 827-1696/9-23 µm.

Sterrasters (Fig. 36F), small and spherical, with warty rosettes (Fig. 36G), 31–51 µm.

Oxyasters (Figs. 36H–36I), with 4–10 long and spined actines, overall measuring 16–49 µm

Strongylasters (Figs. 36J–36L), with the spines concentrated at its tips, overall measuring 5–23 µm.

Ecological notes

All specimens were found at a single station: the summit of the AM, in the mesophotic zone, composed of detrital bottoms with gross sand and gravels. Due to its small size and the amount of sediment that it agglutinates, the species may have been neglected from many other stations with similar features. In the field, it was almost indistinguishable from G. bibilonae sp. nov. and S. dichoclada.

Genetics

COI (ON130529) and 28S (C1-C2) (ON133884) sequences were obtained from the holotype (i589_2).

Taxonomic remarks on Geodia bibilonae sp. nov. and Geodia microsphaera sp. nov.

Both new species have very similar external morphology, spicule sets and yet clearly different COI/28S sequences; they were also found on different seamounts. Geodia bibilonae sp. nov. probably has cribriporal oscules and pores since only cribriporal openings could be found and we then assumed that some of those were inhalant and others exhalant. It can be therefore compared with North Atlantic/Mediterranean species with indistinctive cribriporal oscules/pores, some of which were formerly grouped in the genus Cydonium Fleming, 1828, now a synonym of Geodia. These species are the shallow/mesophotic species Geodia cydonium, Geodia conchilega Schmidt, 1862, Geodia tuberosa Schmidt, 1862 and the deep-sea North Atlantic Geodia macandrewii and Geodia nodastrella Carter, 1876. To this we need to add the poorly described species Geodia pergamentacea Schmidt, 1870 from Portugal. We can easily discard G. macandrewii and G. nodastrella, which are massive sponges found in the North Atlantic at deeper depths (Cárdenas & Rapp, 2013; Cárdenas & Rapp, 2015) and have for the former much larger sterrasters and for the later very characteristic ectocortical spherasters; both have also been sequenced and have different COI and 28S (Fig. 10). The shallow G. conchilega has a characteristic thick cortex and fairly large oval sterrasters, quite different from those in our new species; its COI/28S sequences are also different from those of our new species. As for the two poorly known Schmidt species, G. pergamentacea and G. tuberosa, for which no illustrations are published, we can rely on redescriptions (Sollas, 1888; Burton, 1946) from fragments and/or slides of type material. Although their succinct and incomplete descriptions make it challenging to identify these species, they are clearly different from our new species with much larger sterrasters (spherical 90 µm for G. tuberosa and oval 60–80 µm for G. pergamentacea). Burton (1946) further suggests that G. pergamentacea is a synonym of G. conchilega but this remains to be confirmed.