Cheilostomatida (Bryozoa) from the Ionian Apulian coast (Italy) with the description of new species

Abstract The mesophotic zone is a relatively poorly studied area of the Mediterranean Sea, drawing great interest by the scientific community in the last years. This zone represents a connection between the shallow water and the deep-sea communities, in which photophilic framework builders (e.g. coralline red algae) are gradually replaced by heterotrophic ones, such as ahermatypic corals and the bivalve Neopycnodonte cochlear. In this habitat the framework-forming organisms produce a hard substrate with a high topographic complexity, hosting a great biodiversity of secondary structuring taxa like bryozoans. During a survey on coralligenous banks in the mesophotic zone in c. 60 m depth off Gallipoli (southern Apulia), epibiotic aggregations of N. cochlear were found on the fans of the hexacoral Savalia savaglia. In the present paper the diversity of cheilostomatid bryozoans hosted by these bivalve aggregations is described and compared with published information on similar nearby habitats. A total of 48 taxa were found, six of which are newly described: Crassimarginatella matildae sp. nov., Micropora biopesiula sp. nov., Haplopoma celeste sp. nov., Schizomavella (Schizomavella) cerranoi sp. nov., Schizomavella (Calvetomavella) biancae sp. nov., and Schizoporella adelaide sp. nov. The species richness known from the southern Apulian shelf at this depth (47 species) is hereby raised to 83 cheilostomatid bryozoans. Moreover, only 12 species are shared with the other localities studied previously, while 36 are restricted to Gallipoli, supporting the hypothesis of a high rate of exclusivity among Apulian sites in terms of species composition. The differences in faunal composition, and particularly the presence of several new species discovered at Gallipoli, show once more that our knowledge of the bryozoan fauna in certain Mediterranean habitats is still incomplete and warrants further studies. https://doi.org/urn:lsid:zoobank.org:pub:DDC82039-EF44-4169-8198-C67F60B14BA0


Introduction
In the Mediterranean Sea the bryozoan diversity is relatively high considering the dimensions of the basin: the c. 556 described species represent about 10% of the global bryodiversity (Rosso & Di Martino 2016). This number is expected to increase since many habitats and areas in the Mediterranean are still understudied, such as the eastern and southern regions (D'Onghia et al. 2015;Rosso & Di Martino 2016). Each year, several new species and genera are also being described in the Mediterranean thanks to revision works (e.g. Souto et al. 2010a;Vieira et al. 2014;Reverter-Gil et al. 2016;Berning et al. 2019) as well as owing to newly collected material from hardly accessible habitats like dark caves, the mesophotic zone and deep-sea areas (Rosso et al. 2018;2020a, 2020b. The mesophothic zone is a relatively unexplored area of the Mediterranean Sea, and growing interest by the scientific community in the exploration of this zone is evident in the last years (e.g. Cerrano et al. 2019 for a review; Albano et al. 2020;Giampaoletti et al. 2020). Infact, this zone comprises many different habitats (e.g. part of the coralligenous, various kinds of animal forests and maërl beds) and, due to the limitation of light, the light-dependent, shallow water bioconstructors are gradually replaced by heterotrophic deep-sea communities (see Gori et al. 2017). In the mesophotic zone of the southern Italian Apulian region, two peculiar habitats were recently discovered in which invertebrates are the main bioconstructors (Corriero et al. 2019;Cardone et al. 2020). One of these habitats is constructed by the bivalve Neopycnodonte cochlear (Poli, 1795) and the second is made by two species of scleractinan corals (Phyllangia americana mouchezii (Lacaze-Duthiers, 1897) and Polycyathus muellerae (Abel, 1959)). In these two different habitats the bioconstructions are composed of a multilayered aggregation of dead animals in the centre and an outer layer of presently living organisms, forming a hard substrate with a high topographic complexity, and hosting a huge biodiversity (Corriero et al. 2019;Angeletti & Taviani 2020;Cardone et al. 2020;Giampaoletti et al. 2020). In this habitat the bryozoans, as well as other calcified organisms, play an important secondary role in bioconstruction . In a previous study, the analysis of blocks from these peculiar habitats along the Apulian coasts showed the presence of 50 bryozoan species, 46 of which were Cheilostomatida (Giampaoletti et al. 2020).
In the frame of the Mediterreanean Mesophotic research project (MESOMED) expeditions, which aim at exploring the Mediterranean mesophotic zone, coralligenous banks were found off Gallipoli (Apulia, Southern Italy). On the banks the outer layer was dominated by the octocoral Paramuricea clavata Risso, 1826 or by the gold coral Savalia savaglia (Bertoloni, 1819). Epibiontic organisms occasionally settle on injured branches of S. savaglia. Particularly the bivalve Neopycnodonte cochlear may form large epibiotic aggregations. The aim of the present paper is to describe the

372
D. Pica et al. diversity of cheilostomatid bryozoans hosted by these aggregations and compare the results with the bryofauna present at the same depth in the adjoining Apulian areas.

Material and methods
The coralligenous banks at 60 m depth off Gallipoli (Lecce) (40°0ʹ56.20″N; 17°55ʹ17.84″E) along the Ionian Apulian coast (Figure 1(a)) were investigated by technical scuba diving on 16 July 2015. The coralligenous banks are up to 2 m in height, towering over a horizontal substrate. The coralligenous is characterised by an outer layer that is dominated by the octocoral Paramuricea clavata Risso, 1826 and the gold coral Savalia savaglia. Epibiontic organisms settle on injured branches of S. savaglia, in particular the bivalve Neopycnodonte cochlear (Figure 1(b)). This species starts to colonise dead or damaged portions of the colony (e.g. those inflicted by fishing lines), triggering the development of a very diverse epibiotic community throughout its lifetime as well as post mortem. The bryozoan species studied in the present work live on dead shells of N. cochlear (Figure 1(c)) that were still attached to the S. savaglia skeleton. Fourteen groups of shells cemented to each other (about 5-6 shells each) were collected from different S. savaglia colonies and preserved dry. The shell specimens were observed using a Nikon SMZ18 stereomicroscope in order to detect and preliminarly identify the living bryozoan specimens. For all the preliminarily identified bryozoan species, small portions were detached from the shell for detailed analysis of the morphological characters under a scanning electron microscope (SEM). Both untreated and bleached colony portions (using sodium hypochlorite) were mounted on stubs, coated with goldpalladium in a Balzer Union evaporator, and examined with a Philips XL20 SEM. Morphometrics were made on the micrographs using the image software ImageJ (Schneider et al. 2012), and are given in the descriptions as mean ± standard deviation, minimummaximum values, and number of measurements (whenever >2 measurements were taken). Values are in µm unless otherwise noted; abbreviations: L -length, W -width. The type specimens are deposited in the Museo di Storia Naturale di Genova (Italy) labelled with the acronym MSNG. All other examined material is in the first author's personal collection. Comparative material was examined from the bryozoan collection of the Natural History Museum London (NHMUK). The synonymy lists of the respective species are not exhaustive. In the light of recent taxonomic revisions, which revealed that many of the previously accepted bryozoan species are actually species complexes, we here provide only those references in which the species is figured (besides the original publication in which the species was introduced), which allowed verifying their synonymy.

Remarks
The colonies are too poorly preserved to permit identification to species level, while it is clear (from the budding pattern of the encrusting colony portion) that at least two species are present.

Remarks
The auto-and ovicellate zooids in the colonies studied here usually have six oral spines while in other populations most zooids have only four spines (cf. Hayward & Ryland 1998: 160;Hayward & McKinney 2002: 18). While Callopora dumerilii usually occurs in shallow waters in the Mediterranean Sea (Hayward & McKinney 2002;Chimenz Gusso et al. 2014), the present specimens from 65 m presumably mark the deepest record of the species in this region.

Remarks
The present specimens are in morphological accordance with the other Mediterranean populations referred to Copidozoum planum.

Diagnosis
Crassimarginatella with subhexagonal zooids, gymnocyst moderately well developed proximally and narrowing laterally; cryptocyst evenly developed proximally and laterally, narrowing and disappearing only at the very distal end, surface granular to beaded, encircling an oval opesia; zero, two or occasionally three spines on distal zooid margin, two spines in ovicellate zooids. Cystid of vicarious avicularium usually smaller than autozooid but may reach the same size, avicularium with an overall oval appearance, rostrum and mandible distally producing a narrow downcurved mucro; crossbar complete, thick and without columella. Ovicells hyperstomial, ooecium hemispherical; ectooecium smooth, entirely calcified apart from a narrow proximal window of variable shape.

Etymology
Named after the first author's nephew, Matilda Colapaoli.

Material examined
Holotype
Avicularium vicarious, surrounded by five autozooids, usually slightly smaller than an autozooid while some may be as large as an autozooid (L: 342 ± 43, 269-399, 15; W: 208 ± 30, 167-267, 14); cystid with variably developed gymnocyst (Figure 6(e)). Rostrum (L: 196 ± 26,14) at an acute angle to frontal plane, distally with a relatively long and narrow mucro that is downcurved and incised into the transverse wall of the avicularian cystid; avicularian surface almost equally divided into a distal rostral and a proximal opesial half by a robust and relatively smooth crossbar, semicircular proximal opesia usually slightly wider than the similarly shaped palatal opesia, both framed by a cryptocystal rim similar to that of autozooids (Figure 6(e)).

Remarks
Hincks (1880a) described Crassimarginatella crassimarginata from Madeira, which is characterised by avicularia with a distally rounded rostrum and mandible, while oral spines are apparently absent. The nominal species was subsequently recorded also in the Mediterranean Sea (e.g. Gautier 1962;Prenant & Bobin 1966;Zabala & Maluquer 1988). However, although Harmelin (1973) as well as Harmelin and d'Hondt (1993) explicitly noted the differences in avicularium morphology between the Atlantic and Mediterranean forms, the populations have until now been treated as synonymous. As the mucro in the distal rostrum, in which the pointed, sclerotised tip of the mandible comes to rest, is a constant and exclusive character in the Mediterranean population, we here treat the two taxa as specifically distinct, and introduce Crassimarginatella matildae as new species. It is endemic to the Mediterranean Sea while the Alboran Sea is an ecotonal area where both species co-occur (Harmelin & d'Hondt 1993

Remarks
The present specimens morphologically agree with the other Mediterranean populations referred to Ellisina gautieri.

Remarks
The small colony studied is possibly the early astogenetic, encrusting phase of an erect flustrid, which cannot be determined even to genus level owing to the absence of significant characters and comparative material as usually only the erect colony portion of flustrids is described and figured. The zooids are very lightly calcified, c. 0.7 mm long and 0.5 mm wide, and a cryptocyst is not developed. Spines or avicularia were not observed. As Giampaoletti et al. (2020) reported Chartella papyrea (Pallas, 1766) from the mesophotic site of Monopoli in the Adriatic Sea, we doubtfully assign the present specimens to this genus.

Material examined
Pica coll: one unbleached colony.

Remarks
Only a tiny colony was obtained, which was observed using an optical stereomicroscope in order to confirm the genus as it does not allow to determine the specimen to species level.

Remarks
The specimens from Apulia showed no differences to Beania mediterranea, which was recently described by Souto et al. (2018).

Material examined
Pica coll: one unbleached colony.

Remarks
As only a single colony with few zooids was obtained, the specimen was not examined using SEM.

Diagnosis
Micropora with zooids that are somewhat longer than wide (L/W = 1.41); cryptocyst granular, perforated by 25-50 pseudopores and two pairs of opesiules: a larger, elongated and funnel-shaped one in the distolateral corners, and a smaller pair of round opesiules proximal to the former; autozooidal opesia slightly wider than long (L/W = 0.61). Gymnocystal bosses lateral to proximal orifice small and usually inconspicuous. Avicularia absent.
Ovicell semi-immersed in distal zooid, ooecium hemispherical, about as wide as long (L/W = 1.12), cryptocystal surface granular, proximal margin arched, composed of a broad flat band of smooth gymnocyst with a central suture above which an apex is formed; ooecial orifice dimorphic, slightly larger than in autozooids but with similar length/ width ratio (0.63).

Etymology
The name refers to the presence of two pairs of opesiules.
First asexually produced zooid budded from distal end of ancestrula, the two second-generation autozooids are produced by both the ancestrula and the 1st generation zooid, oriented at a c. 90° angle; the 3rd-generation zooids grow in the interspaces of this cross-shaped formation, establishing growth of the colony in an opposite direction respect to the polarity of the ancestrula (Figure 10(h)).

Remarks
Three Micropora species are known from the modern Mediterranean Sea and eastern Atlantic. Micropora coriacea (Esper, 1791) is the most frequently cited species with an alleged worldwide distribution, which, however, most likely represents a species complex. The species differs from Micropora biopesiula sp. nov. in having only the single distal pair of funnel-shaped opesiules, and usually more pronounced bosses lateral to the proximal orifice margin (e.g. Hayward & Ryland 1998: 288, figs 97, 99C-D). Besides having interzooidal avicularia, Micropora normani Levinsen, 1909 from the Atlantic coast of Europe has smaller zooids, fewer pores in the cryptocyst, and also only a single pair of opesiules (Hayward & Ryland 1998: 290, figs. 98, 99B). Finally, the West African Micropora robusta Cook, 1985 also has interzooidal avicularia and just a single pair of opesiules.
Three other Micropora species may exist in the region (see Bock & Gordon 2021): Micropora depressa (Moll, 1803) and Micropora aculeata (d'Orbigny, 1852), which are both considered as taxa inquirenda, as well as Micropora africana (d'Orbigny, 1852). To our knowledge, none of these species have ever been cited or redescribed since their original description, and their status, as well as their generic assignments, are unclear at present.
As it will be difficult to clarify the status and systematic placement of these three species any time soon, we have decided to describe the present species as new to science in order to raise awareness among biologists that the oft-cited M. coriacea is not the only Micropora species present in the Mediterranean Sea.

382
D. Pica et al. Micropora biopesiula sp. nov. has so far only been reported from Italian coasts (Rosso 1996;Chimenz Gusso et al. 2014). Whether the new species is also present in the eastern Mediterranean cannot be decided at present as images and descriptions are wanting (e.g. Hayward 1974).

Material examined
Pica coll: SEM stub n°49, one bleached colony; two additional unbleached colonies.

Remarks
Mollia circumcincta [for a recent description based on type material from the Adriatic see Hayward and McKinney (2002: 34)] has been reported only a few times since its discovery, and all records from beyond the central Mediterranean Sea have to be regarded as doubtful. For instance, in the colony from the western Mediterranean figured by Zabala and Maluquer (1988: pl. 2, figs. C, D), the ooecium differs in producing a relatively large area of smoothly calcified ectooecium proximally and laterally that encircles a distally positioned area of exposed nodular endooecium. In Adriatic and NE Ionian M. circumcincta, in contrast, the ectooecium is extremely reduced and restricted to the proximolateral corners of the ooecium. Moreover, the salient distolateral rim framing the zooids is thicker and the nodules covering the cryptocystal surface larger in central Mediterranean populations. The peculiar calcified ridge on the interior basal wall that was noted in fossil specimens (Berning 2006: figs. 32 and 33) are here shown to be present also in the Recent species (Figure 11

Material examined
Pica coll: SEM stub n°44, one bleached colony; one additional unbleached colony.

Remarks
While Taylor et al. (2018) have recently summarised the problems existing around the type species of the genus Onychocella, and discussed the potential synonymy of the fossil Onychocella angulosa (Reuss, 1848) and the present-day Onychocella marioni Jullien, 1882, Rosso et al. (2020b) thoroughly redescribed and figured the Recent species, to which we assign the Apulian specimens.

Remarks
There is quite a confusion concerning the morphology of Cribrilaria hincksi. When introducing the species, Friedl (1917) mentioned (without figuring the material) that his specimens from the northern Adriatic Sea are characterised by "sehr langen dolchförmigen Avicularien" (transl.: very long, dagger-shaped avicularia) that he considered similar to specimens reported from Madeira by Hincks (1880a: 74, pl. 10, fig. 1) as Cribrilina radiata var. (Moll 1803). According to Hincks' figure, the Madeiran species has indeed very long rostra that are apparently positioned between zooids (which is, however, not quite clear from the drawing). In subsequent works (e.g. Prenant & Bobin 1966;Harmelin 1988;Zabala & Maluquer 1988), C. hincksi was regarded as having an extremely elongate avicularium and a serrated rostrum that is recumbent on the distal zooid. The present specimens comply with these characters, but neither ours nor any of the colonies recorded in the above-mentioned publications were from the type locality, the northern Adriatic. Recently, Hayward and McKinney (2002: 38, figs. 17A-C)) did identify specimens from the type locality as P. hincksi, which have, however, comparatively short avicularia that are wedged between zooids, and are therefore unlikely to be conspecific. Thus, without having seen the type material, which is beyond the scope of this paper, it is impossible to define the exact morphology of C. hincksi.

Remarks
The present specimens are within the morphological range observed in other Mediterranean populations referred to Cribrilaria venusta. The original description and image of the species by; Canu and Bassler (1925), however, suggest that both the orifice and ovicell are distinctly smaller in relation to zooid size. A revision of the type material is therefore necessary before the Mediterranean specimens can unequivocally be synonymised with Canu & Bassler's species from the southern Gulf of Cádiz.

Remarks
Several slight morphological differences extist between the British Cribrilaria innominata (Couch, 1844), as redescribed by Bishop (1986), and Mediterranean specimens attributed to this species. For instance, the apophyses that occur at the base of (some of) the oral spines in Mediterranean populations (Harmelin 1988;Chimenz Gusso et al. 2014) have not been reported in British specimens. Also, in Mediterranean colonies only a single costal pair contributes to the formation of the suboral umbo that proximally frames the lacuna, and there is no median ridge on the costate frontal shield, whereas in British specimens the two distal pairs of costae form the umbo that proximally turns into a short median carina (Bishop 1986;Hayward & Ryland 1998). Additional comparative SEM studies and genetic analyses are needed before a conclusion as to the status of the Mediterranean specimens can be drawn.

Remarks
The pedunculate avicularia and the kenozooidal ovicell are typical for the genus, and the present specimens fully conform with previous records of Glabrilaria pedunculata.

Material examined
Pica coll: SEM stub n°64, one bleached dead colony; one additional unbleached colony.

Remarks
The colonies encrust a sponge and were, post mortem, covered by a thin veneer of sponge tissue and numerous spicules in turn, giving the colonies a peculiar appearance.

Material examined
Pica coll: SEM stub n°CHO1 one bleached colony; two additional unbleached colonies.

Remarks
A variety of morphotypes from around the world have been attributed to Chorizopora brongniartii

Diagnosis
Haplopoma with an evenly perforated frontal shield except for an area around the ascopore, pseudopores bevelled but simple, shield surface smooth but transversely wrinkled and often with a distinct median crest of variable length produced by jointed knobs; orifice broader than long, with a pair of small round condyles; ovicellate zooids of similar size as autozooids; ooecial surface evenly perforated by pseudopores, occasionally with a low median ridge.

Etymology
Named after the first author's niece, Celeste Beatrix Pica; used as a noun in apposition.    (Figure 20(a,b)). Gymnocystal frontal shield translucent, slightly convex, surface smooth but with transverse wrinkles and uniformly perforated by some 28-36 bevelled round pseudopores except for an area around the ascopore (Figure 20(b)), a short slit is visible distal to each pore using an optical microscope; ascopore circular (23 µm), only slightly larger than the frontal pores ( Figure 20(b)); proximal to the ascopore a distinct narrow crest of variable length formed by a linear series of more or less jointed knobs is present in most zooids (Figure 20(b)). Five oval basal pore chambers on each zooid side, communication pores round and  Ovicellate zooids about the same size as autozooids, ovicells kenozooidal, ooecium hemispherical, partly immersed below colony surface, slightly wider than long (L: 238 ± 8, 229-247, 5; W: 245 ± 15, 230-266, 5), closure type cleithral; smooth ectooecial calcification resembling the frontal shield of the zooid with scattered pores, occasionally with a low central ridge (Figure 20(d)).

Material examined Holotype
Ancestrula not observed.
Together with H. bimucronatum, H. planum and H. sciaphilum, Haplopoma celeste sp. nov. is among the species that are recorded from waters below 50 m depth (or from caves in shallower waters), while the other species are mostly restricted to in the shallow subtidal. To our knowledge, H. celeste has not been recorded before in previous works and is thus only known from the central Mediterranean Sea at depths of 60 m.

Material examined
Pica coll: two unbleached colony fragments.

Remarks
Only two poorly preserved colony fragments were found, which makes it difficult to precisely identify the species.

Remarks
For the sake of consistency, we here continue to follow the most recent works and refer the modern species to the fossil Prenantia ligulata (Manzoni, 1870), which would have priority if it should turn out to be synonymous with Prenantia inerma (Calvet, 1907), as suggested by Poluzzi (1975). However, the types of both species need to be consulted before this decision can be finalised. The ancestrula has been observed before but was only incompletely described owing to the periancestrular zooids covering the proximal and lateral gymnocyst (Rosso 2004). Here we provide additional information on the morphology of the ancestrula (Figure 21(e,f)). It is about 420 µm long and 290 µm wide, with steep marginal walls rising to an elevated oval area that comprises about the distal two-thirds of the ancestrula. The distinctly raised margin demarcating this area is tightly framed and indented by nine spines that are more or less arching over the frontal surface, the six distal ones of which are relatively closely spaced. The area comprises a broad proximal shelf of smooth cryptocyst that gently slopes towards the centre and narrows distolaterally, and a relatively small, distal, suborbicular opesia. The proximal cryptocystal shelf exhibits a central longitudinal suture.

Remarks
The present specimens fully conform with the characters of Schizomavella (Schizomavella) cornuta.  fig. 11.

Remarks
The Apulian specimens belong to the Schizomavella linearis species complex but differ from all populations hitherto described. Most closely related is the form referred to by Reverter- Gil et al. (2016) as the "pseudolinearis" morphotype from the Adriatic Sea, which has a relatively narrow sinus and avicularia that were reported to be often directed distally (Reverter-Gil et al. 2016: 308), although their fig. 11f exclusively shows medially directed avicularia. In the present specimens, the sinus is even narrower, and the avicularia are oriented distomedially to distolaterally. However, despite the large morphological differences between some of the Mediterranean populations and those from the type locality, the British Isles (see Hayward & Thorpe 1995;Hayward & Ryland 1999), Reverter-Gil et al. (2016) were reluctant to introduce new species owing to the transitional nature of some of the characters, and also because of an absence of a clear geographic pattern in the morphotypes. We have, accordingly, decided to add the present specimens to the species complex and wait for genetic analyses to tackle the issue of their relatedness. (Hayward & McKinney, 2002)

Remarks
The present specimen slightly differs from those from the type locality, Rovinj, in the northern Adriatic Sea (Hayward & McKinney 2002), in having a less rugose frontal shield with somewhat smaller pores. As all other characters are identical, we consider these differences to have been caused by environmental conditions at greater depth in which the Apulian population was found (60 m vs. 5-20 m off Rovinj).

Schizomavella (Schizomavella) cerranoi
Pica & Berning sp. nov. Fig. 25a-f Diagnosis Schizomavella with unilaminar colonies. Zooids rectangular in outline, frontal shield perforated by numerous, relatively small pores. Orifice proximally and laterally surrounded by a thin peristome; primary orifice suborbicular, with a narrowly U-shaped sinus occupying about one-fourth of the proximal margin, condyles relatively long and narrow, with sloping shoulders or running parallel to proximal orifice margin and stopping short of the sinus, no additional structures; distal orifice margin usually with 4 spines persisting thoughout ontogeny in non-ovicellate zooids. Suboral avicularium forning a tall umbo, oriented almost perpendicular to frontal plane, rostrum elongate triangular, crossbar with a small rounded columella. Ovicell marginally covered by secondary calcification of the distal zooid, forming a rugged crest around a small central area of exposed ectooecium that is regularly perforated by small pseudopores. Etymology Named after the person who collected the material studied in this paper, Prof. Carlo Cerrano.

Remarks
The new species is very closely related to Schizomavella (Schizomavella) subsolana (see above) but differs from it in having a distinctly narrower and deeper sinus. Moreover, Schizomavella (Schizomavella) cerranoi sp. nov. is characterised by larger zooids (particularly their width) and orifices, by a smaller area of exposed ectooecium as well as by having a maximum number of four oral spines. As the two morphotypes apparently occur sympatrically, and as intermediate morphologies were not found in any of the colonies, we have decided to introduce a new species for the present material.

Remarks
Only two small immature colonies were obtained. The specimens are very similar to the recently introduced species Schizomavella (Schizomavella) tubulata Reverter-Gil, Souto, Novosel & Tilbrook, 2015 from the Adriatic Sea. Both species have similarly shaped orifice -ith four to five distal spines and a spatulate suboral avicularium oriented almost perpendicular to the frontal plane. In our specimens the early astogenetic zooids show more spines, up to seven, decreasing in number toward the zone of astogenetic repetition. The present colonies differ slightly in lacking a finely denticulate rostrum (cf. Reverter-Gil et al. 2016: fig. 8E), in the absence of a peristome encircling the entire orifice, and in the frontal shields with far fewer and smaller pores than in the holotype. However, our colonies are relatively young and immature, in contrast to the originally described mature colonies that showed multilaminar growth. Reference of the Apulian colonies to S. tubulata is thus somewhat doubtful.

Diagnosis
Schizomavella (Calvetomavella) with encrusting unilaminar colonies. Autozooidal frontal shield with a coarsely granular to nodular surface and few small pores around a central imperforate area. Orifice suborbicular in shape and with a narrow and deeply U-shaped sinus as well as broad condyles sloping towards sinus, six or seven oral spines. Avicularia dimorphic: large avicularia often paired and situated along distolateral margins though single latero-suboral ones also occur; oblong in outline, i.e. rostrum distally rounded with slightly raised and serrated edges, palate mostly calcified by a pair of cryptocystal shelves forming central ridges and an elongate foramen at about mid-distance, crossbar with tiny columella. Small avicularia single and usually (latero)suboral though distolateral ones also occur, oval to oblong in outline, slightly raised distal part of rostrum serrated, distal foramen Y-shaped, crossbar with tiny columella. Ovicells relatively large, often covering over half the frontal area of the distal zooid, ooecia flattened globular, ectooecium almost entirely exposed apart from a narrow peripheral rim and with c. 25 rimmed tubaeform pores, ovicellate zooids forming a peristome a with a large U-shaped suboral notch.

Etymology
Named after one of the first author's nieces, Bianca Colapaoli.

Material examined
Holotype
Ancestrula not well preserved, oval in shape, longer than wide, opesia mushroom-shaped surrounded by nine or ten mural spines (Figure 27(f)).

Remarks
As noted by Reverter- Gil et al. (2015: 40), who figured material of Schizomavella (Calvetomavella) discoidea (Busk, 1859) from its Madeiran type locality using SEM for the first time, the species is specifically distinct from eastern Atlantic continental shelf and Mediterranean Sea populations, which were historically assigned to that species. Schizomavella (C.) discoidea differs from the present colonies in having larger pores in the frontal shield, a differently shaped peristome in ovicellate zooids (with sloping lateral margins, i.e. the U-shaped proximal notch is not developed), a broader orificial sinus, and in the shape of the avicularian rostrum, which is tapering distally (cf. Reverter- Gil et al. 2015: fig. 2). At the present state of knowledge, Schizomavella  (Calvetomavella) biancae sp. nov. occurs in the central and western Mediterranean Sea, and along the Atlantic continental shelf of Europe, with the Shetland Islands as its northernmost limit of distribution (cf. Hayward & Ryland 1999;Reverter-Gil et al. 2015). However, in the absence of SEM images from most publications, the distinction between the new species and the morphologically closely related Schizomavella (Schizomavella) halimedae (Gautier, 1955) Gil et al. 2016: 313). Also, the ooecium is covered by nodular secondary calcification in the former species, whereas the pseudoporous ectooecium is entirely exposed in the latter. As the depth ranges of these species are also largely overlapping, with S.

Remarks
Schizoporella magnifica (Hincks, 1886) occurs throughout the Mediterranean Sea and north to the SW British Isles (Hayward & Ryland 1999: 216). The sinus in the present colonies seems to be relatively wide but nevertheless within the range observed in other populations (e.g. Hayward & McKinney 2002: fig. 31F, H).
Another similar yet specifically distinct Schizoporella species is also present on the Neopycnodonte shells (see below).

Paratypes
Same data as for holotype. MSNG 62432: one bleached colony.

Remarks
Schizoporella adelaide sp. nov. is morphologically similar to Schizoporella magnifica (see above) but can be distinguished from that species owing to a more elongated orifice with a distinctly shorter sinus and condyles, and a crossbar in the avicularium that has a prominent columella. Similarities also exist with the British Schizoporella cornualis Hayward & Ryland, 1995, the only other European species that has an avicularium with a columella. While its orifice is rather similar as well (about as long as wide; see Hayward & Ryland, 1995: 46), it has distinctly larger condyles than the new species, and also produces a suboral umbo on its frontal shield. Two undescribed species, Schizoporella sp. 2 and sp. 3, which were recorded by Chimenz Gusso et al. (2014) from the southern Italian Island of Ustica are also similar to S. adelaide sp. nov. with regards to the distal pair of avicularia (see also Rosso et al. 2019b : fig, 5I). Both species differ from the new species, however, in having a broader sinus, non-crenellate and shorter condyles, and in the absence of a columella in the avicularian crossbar, among other characters.
As no other records could confidently be assigned to Schizoporella adelaide sp. nov., the species must be regarded as endemic to the central Mediterranean Sea at present.

Remarks
The number and identity of Hagiosynodos species present in the Mediterranean Sea is unclear and a matter of debate, and genetic analyses are certainly needed in order to solve this problem. Most of the interspecific differences given to justify the distinction between Hagiosynodos latus (Busk, 1856), H. kirchenpaueri (Heller, 1867) and H. hadros Hayward & McKinney, 2002 are the dimensions of certain characters. However, the more (fossil) material is looked at, the more difficult it is to differentiate between interspecific differences and intraspecific variability in response to (micro)environmental conditions, and to draw a clear line between these species (Schmid 1989;Berning 2006). Accordingly, the present material is adding to this problem (see Table I kirchenpaueri and H. hadros (distance between orifices, orifice length, ovicell width), between H. kirchenpaueri and H. latus (width of anter, distance between frontal shield pores), or in between all of the three (ovicell length). Zooid width, on the other hand, is distinctly greater in the present material than in all of the species reported by Hayward and McKinney (2002). Concerning qualitative character differences, the present colonies resemble H. latus in the presence of a distal lip (or occasionally a pair of distolateral teeth connected by a narrow shelf), while this character is apparently absent in H. kirchenpaueri and H. hadros (Hayward & McKinney, 2002: 79). Thus, whereas skeletal dimensions and relative proportions would vaguely support an assignment of the present colonies to either H. kirchenpaueri or H. hadros, the only distinct character difference argues against this decision. Thus, while we do not reject the likelihood of three distinct species being present in the NE Atlantic and Mediterranean Sea as such, we here stress the inability of distinguishing the species based on morphometry and morphology, and hope that genetic analyses may aid in solving the issue. Accordingly, we here assign the specimens to H. lata and treat the other species as synonymous for now.

Remarks
The colony recovered is presumably a part of the encrusting base of an erect colony. Although the zooids are chaotically arranged, possibly as a result of damage  to the colony and subsequent reparative budding, the zooids look similar to Phoceana tubulifera (Heller, 1867), which has been described from the Adriatic Sea though seldom reported thereafter. The frontal shield is perforated proximally by pseudopores and forms a prominent peristome around the orifice, while the surface is finely granular. Differences to P. tubulifera exist, however, in characters of the orifice. While a lyrula is also present in our material, it is distinctly longer and broader than that figured by Hayward and McKinney (2002: fig. 23c) from the erect portion of the colony. Moreover, condyles were not figured nor reported to be present by these authors, while short and rounded ones exist in our material (Figure 31(b)).
We have therefore decided to merely confer our colonies to P. tubulifera.

Remarks
The present specimens, though mostly small colonies, seem to agree with previous records of M. appendiculata from the Mediterranean Sea (see Di Martino & Rosso 2021). (Peach, 1868)

Remarks
The fragments here reported are exclusively from the encrusting base of the erect M. verrucosa and lack ovicells.

Remarks
Escharina dutertrei Audouin, 1826 was originally described from the Red Sea. As its types are lost, and a neotype has not been selected yet, the species remains ill-defined. This problem led; Zabala et al. (1993) to introduce three new subspecies characterised by a pair of small distal avicularia: the nominotypical Escharina dutertrei dutertrei, Escharina dutertrei haywardi from the boreal Atlantic, and Escharina dutertrei protecta from the Mediterranean Sea and the Azores. At least the Atlanto-Mediterranean subspecies only occur offshore, which, together with the contrast in water temperatures between the boreal Atlantic and the tropical Red Sea, suggests that they are not Lessepsian species. This, in turn, argues against their status as subspecies as they must be regarded as evolutionary entities that are geographically separated from the Red Sea population. Therefore, and in concert with the modern species taxon concept in Bryozoa, in which subtle differences are often considered species-specific, we here regard the three taxa as

406
D. Pica et al. distinct at species level. Accordingly, while E. dutertrei still needs to be redefined, Escharina haywardi; Zabala, Maluquer & Harmelin, 1993 comb. nov. from W Britain is distinguished from the Mediterranean Escharina protecta Zabala, Maluquer & Harmelin, 1993 comb. nov. in lacking protective structures around the orifice and in having a different orifice morphology. Based on the image of a specimen of E. protecta from the Azores provided by; Zabala et al. (1993: fig. 14), and owing to the sheer distance between the localities that inhibit a genetic exchange, the central Atlantic population is also likely to be specifically distinct. The present specimens are very similar to E. protecta but may not be conspecific for the following reasons: there is a distinct area of gymnocystal calcification around the proximolateral orifice of autozooids that forms the outer rim of the protective structure, which is little elevated in the centre while rising distally towards the proximal pair of spines where it forms a pair of small flaps. The flaps are much larger and thicker as well as more distinctly raised and irregularly conical in the autozooids of E. protecta, and apparently not connected by a peristomial rim proximally, and the gymnocystal area also seems to be missing. Only in ovicellate zooids is the peristome similarly developed in both species, although the lateral flaps also appear to be larger in E. protecta. Moreover, the condyles are even narrower and the avicularian mandibles are distinctly longer the present specimens when compared with the images provided by Zabala and Maluquer (1988: pl. 15, fig. C) and Zabala et al. (1993: fig.  12b), respectively. We have therefore decided to describe this morphotype in detail, while an analysis of additional material from different habitats and regions is necessary to test where the boundaries between intra-and interspecific differences lie in this species complex.

Remarks
There are no apparent differences between the colonies recovered here and other records of Escharina vulgaris from the Mediterranean Sea.

Remarks
Only early astogenetic colonies were recovered, in which the characters of the adult zooids described by Berning et al. (2008) are not fully developed yet.
Based on the small U-shaped sinus as well as on the absence of an areolar pore between the vibraculum and orifice, however, we consider the colonies to belong to H. majae. The ancestrula and early astogeny has not been described before, and we here figure these stages for the first time. The ancestrula is, as in the type species of the genus, Herentia hyndmanni (Johnston, 1847), a dome-shaped kenozooid with a completely calcified frontal shield, apart from a central pore and a pair of distolateral spines, the bases of which can vaguely be observed (Figure 36(c,d)). As in H. hyndmanni, the first-generation autozooid is budded distally, which then buds another zooid laterally, while the third-generation zooid is formed in between the two. Even at this stage, differences between H. hyndmanni and H. majae are evident as the early astogenetic zooids in the former already produce the drop-shaped sinus (Berning et al. 2008: fig. 1B).

Remarks
The present colonies can unequivocally be assigned to Therenia rosei Berning Tilbrook & Rosso, 2008. Their occurrence on shells at 60 m depth falls within the previously reported habitat range.

Remarks
The present specimens are morphologically indistinguishable from Hippomenella mucronelliformis (Waters, 1899) from Madeira, the types of which have recently been imaged and described by Berning (2013). As only little and unbleached material was available for that study, and owing to the few published records and images, we here present additional characters of the species. The ancestrula is tatiform, oval in outline (L: 465; W: 322), with the gymnocyst well-developed and gently sloping proximally while narrowing and steepening distally. The opesia is also oval (L: 262; W: 208) and the mural rim carries some 12 spines. The 1st-generation autozooid is budded distally, which then gives rise to another zooid laterally. The avicularian rostrum is serrated along its entire length in both the long and short forms.

Remarks
The present specimens, all of which small colonies, are identical with the species recorded as Celleporina canariensis Arístegui, 1989 from the Adriatic Sea by Hayward and McKinney (2002

Remarks
The only recovered specimen is a delicate branch fragment composed of 4 alternating series of zooids, and is identical with Buskea nitida Heller, 1867 from the Adriatic as reported by Hayward and McKinney (2002), which is arguably regarded as conspecific with Buskea quincuncialis Norman, 1867 from Great Britain.

Material examined
SEM stubs n°78, 79, each with a fragments of the same bleached ovicellate colony.

Remarks
Only a single, large yet immature colony was recovered, which, based on zooidal and avicularian characters, nevertheless allows us to assign it to Turbicellepora avicularis (Hincks, 1860) as reported by Hayward and Ryland (1999) and Hayward and McKinney (2002).

Remarks
As the species is not well defined, it is difficult to unambiguously assign the present specimens to Turbicellepora coronopus Wood, 1844. The type of the nominal species is a Pliocene fossil from Great Britain, which has never been examined using SEM, while the alleged Recent records of the species are confined to the Mediterranean region (Hayward 1978). Lagaaij (1952: 137, pl. 15, fig. 8) designated a lectotype and provided an optical image of it, from which details of the orifice and ovicell cannot be observed. At least one character given in the description, however, differs from Recent specimens assigned to T. coronopus: the "normal", oval, suboral avicularium is reported to be occasionally replaced by a small spatulate one in the type; these avicularia have not yet been recorded in Mediterranean specimens. Moreover, the only Recent specimens that have been imaged before using SEM (Chimenz Gusso et al. 2014) apparently differ in having the suboral avicularium in the centre of the peristome (if single), i.e. directly proximal to the sinus, or on the lateral margin (if paired). In our specimens, as well as in the type of T. coronopus, they are always single (occasionally even absent) and usually slightly offset to the right or left from the proximal peristomial centre. We believe that, besides the lectotype, more material needs to be studied to determine what is to be regarded as intraspecific variability in this

Remarks
While lacking ovicells, the single colony fragment found suggests it may be specifically distinct from Reteporella couchii (Hincks, 1878), which was originally described from Great Britain. Although the distinct peristome with its lateral flaps forming a median suture as well as the typical suboral avicularium, which is positioned on an erect cylindrical cystid, are present, both are distinctly shorter than in R. couchii. Whereas these differences may be related to secondary thickening of the frontal surface during ontogeny, the SEM images of additional specimens from the Mediterranean Sea provided by Zabala and Maluquer (1988) and Chimenz Gusso et al. (2014) show the same features. It may thus be possible that a sister species of R. couchii inhabits the Mediterranean Sea. The species reported as R. couchii by Reverter- Gil et al. (2019) may represent yet another species as the frontal plane of the suboral avicularium is not facing distally but frontally, among other differences.

Remarks
The only three specimens available, a fragment from close to the colony growth margin as well as two colony bases, are lacking ovicells and are difficult to assign to a species. The species comes closest to the drawings of specimens referred to Reteporella harmeri (Hass, 1948) by Zabala and Maluquer (1988). As in Reteporella cf. couchii above, the peristome is welldeveloped, forming two lateral flaps that merge in the centre, producing a long suture leading to a proximal labial pore. One of the flaps either develops a pointed umbo during early ontogeny or produces a large, elongate triangular and distally hooked avicularium that is facing rather distally and pointing in a frontal to lateral direction. Both the umbo and the avicularian cystid form a slightly protruding and relatively straight edge with three knobs along the proximal peristomial rim. The lateral orifice margins initially carry six spines, while these are lost during ontogeny and the bases gradually covered by secondary calcification. Two additional types of dimorphic avicularia are present and formed during later ontogeny: oblong avicularia on the frontal as well as elongate-triangular ones on the abfrontal surface. Both types are medium-sized, budded anywhere on the surface during ontogeny, and point in various directions. None of the crossbars in the three avicularium types bear a columella. The abfrontal surface is granular. Whereas most of these characters are shared with the species figured by Zabala and Maluquer (1988), R. harmeri needs revision. As Hass' types are apparently lost, a neotype needs to be selected, and the species redescribed and imaged using SEM.

Remarks
The present specimens conform with the characters of the lectotype of Dentiporella sardonica, which was recently imaged by Souto et al. (2010b).   fig. 6i.

Remarks
The present specimens conform in all aspects with Schizotheca fissa (Busk, 1856), which is widespread in the Mediterranean Sea and NE Atlantic.

Conclusive remarks
The shell aggregations of Neopycnodonte cochlear collected in the mesophotic zone of the Apulian coast display a relatively large associated biodiversity. We identified 48 cheilostomatid bryozoan species, six of which are newly described in the present paper: Crassimarginatella matildae sp. nov., Micropora biopesiula sp. nov., Haplopoma celeste sp. nov., Schizomavella (Schizomavella) cerranoi sp. nov., Schizomavella (Calvetomavella) biancae sp. nov., and Schizoporella adelaide sp. nov. This number is particularly remarkable when compared with the species richness of other mesophotic habitats from southern Apulia, which were recently presented by Giampaoletti et al. (2020) and Cardone et al. (2020). These authors reported 22 bryozoan species from the site of Monopoli (30-55 m depth) and 18 species from Otranto (45-64 m), both located in the southern Adriatic Sea, as well as 26 species from Santa Maria di Leuca (45-70 m), which is a little south of the locality Gallipoli (60 m) in the Ionian Sea studied herein. The number of species from Gallipoli is even slightly greater than the total number of the three localitites (47 spp.) studied until now (Table SI). Giampaoletti et al. (2020) already noted a high rate of exclusivity among sites in terms of species composition, which is corroborated by our study as only 12 shared species were reported from Gallipoli, while 35 are restricted to this site and were not recorded by Giampaoletti et al. (2020). Thus, a total of 83 cheilostomatid species have been reported from the southern Apulian mesophotic habitats to date.
A comparison of the data needs to be done with caution, however, as Giampaoletti et al. (2020) did not provide images of the species, and the taxa were apparently identified by optical means only. Without SEM images it is difficult to judge whether morphologically closely related taxa were correctly identified by them. For instance, Microporella marsupiata (Busk, 1860) may be confused with M. appendiculata, and it cannot be ruled out that the colonies they assigned to Schizoporella magnifica belong to that species or to the very similar S. adelaide sp. nov. Moreover, the species Giampaoletti et al. (2020) reported as C. crassimarginata (Hincks, 1880) from Santa Maria di Leuca is likely synonymous with C. matildae sp. nov., and their Schizomavella discoidea (Busk, 1859), reported from all three stations, is probably identical with S. (C.) biancae sp. nov.
The differences in faunal composition between similar and geographically relatively proximate sites in the mesophotic zone, and particularly the presence of several new species discovered at Gallipoli, show once more that our knowledge of the bryozoan fauna in certain Mediterranean habitats, such as the outer shelf and upper slope, is still uncomplete and warrants further studies (Rosso & Di Martino 2016).