Seek and ye shall find: new species and new records of Microporella (Bryozoa, Cheilostomatida) in the Mediterranean

Abstract The Mediterranean specimens of the genus Microporella collected from shallow water habitats during several surveys and cruises undertaken mostly off the Italian coast are revised. As a result of the disentanglement of the M.ciliata complex and the examination of new material, three new species, M.bicollarissp. nov., M.ichnusaesp. nov., and M.pachyspinasp. nov., are described from submarine caves or associated with seagrasses and algae. An additional species Microporella sp. A, distinct by its finely reticulate ascopore, is described but left in open nomenclature owing to the limitations of a single infertile fragment. After examination of all available material, based on their identical zooidal morphology, the genus Diporula is regarded as junior synonym of Microporella and the combination Microporellaverrucosa is resurrected as first suggested by Neviani in 1896. Fenestrulinajoannae is also reassigned to Microporella. The availability of a large number of colonies of the above-mentioned and other species already well known from the area (i.e., M.appendiculata, M.ciliata, and M.modesta), allowed the assessment of their high intraspecific variability as well as the observation, for the first time, of some morphological characters including ancestrulae, early astogeny, and kenozooids. Finally, M.modesta, in spite of M.ciliata as defined by the neotype selected by Kukliński & Taylor in 2008, appears to be the commonest species in the basin.

In the Mediterranean, Microporella is represented to date by nine species (Rosso and Di Martino 2016;Di Martino et al. 2020a), and knowledge of the diversity of the genus in the basin has been stable during the last decade. The last thorough revision was by Harmelin et al. (2011), but was limited to the south-eastern sectors of the Mediterranean along the coast of Lebanon. This work increased the number of Microporella species recorded from the Mediterranean, but also showed how the stock of Microporella species found in this area was composed only of tropical alien species.
Here, we examine large volumes of material, collected along the Italian coast in several shallow water habitats, to assess the diversity of the genus in these sectors of the Mediterranean, the morphological intraspecific variability of the Microporella species found, as well as their distribution and commonness. We also describe three new species and, for those previously known, illustrate for the first time some key morphological features. An additional species is described and left unnamed. Although some of its morphological features are distinct from other congeners (e.g., ascopore), the lack of some diagnostic characters, such as ovicells, and the scarcity of material available prevents the description of a new species.
Scanning electron microscopy (SEM) was conducted on uncoated specimens using a TESCAN VEGA 2 LMU in backscattered-electron/low-vacuum mode at the Microscopical Laboratory of the University of Catania. Measurements were taken from SEM micrographs using the image processing program ImageJ (available from https://  imagej.nih.gov/) and are given in the species descriptions and/or in the remarks as ranges and, in parentheses, mean ± standard deviation plus the number of zooids measured (N). Abbreviations for the measurements are: AvL avicularium length; AvW avicularium width; OL orifice length; OW orifice width; OvL ovicell length; OvW ovicell width; ZL autozooid length; ZW autozooid width.
Type material of the new species and figured specimens form part of the Rosso Collection deposited at the Museum of Palaeontology of the University of Catania (PMC) under the catalogue numbers reported for each species.
Microporella flabelligera Levinsen, 1909 described from the vicinity of Siracusa, at depths (28-46 m) comparable to those of the PS/81 sites in the Gulf of Noto, is likely to be conspecific with M. appendiculata based on the original drawings (Levinsen 1909: 331, pl. 24 cited as 23 in the text, fig. 6A-C). However, a formal synonymy requires the examination of the type specimen.
Microporella appendiculata differs from other Mediterranean congeners in having paired avicularia, a character shared only with M. coronata (Audouin & Savigny, 1826). However, in M. coronata the avicularia are located proximally to the ascopore, the ovicell is personate, the oral spines are greater in number (6-8) and unbranched.
Here, we document the regeneration of the ancestrula as an autozooid for the first time (Fig. 2F). Trifurcated proximal spines have already been figured on some zooids, though not described as such by Zabala (1986: fig. 181B, as M. marsupiata) and by Chimenz and Faraglia (1995: pl. 2C, as M. pseudomarsupiata). Gautier (1962) also recorded autozooids with up to seven oral spines.

Distribution and ecology.
In the Mediterranean, M. appendiculata has been reported from several localities and usually in shelf habitats, associated with coarse detritic bottoms, often encrusting shells and calcareous algae (Gautier 1962;Rosso 1996a, b;Madurell et al. 2013;Chimenz Gusso et al. 2014), as well as from shadowed microhabitats associated with Cystoseira s. l. communities (e.g., Campisi 1973;Rosso et al. 2019a). It also occurs in submarine caves of NW Sardinia (Fraschetti et al. 2010) and near Monaco (J.-G. Harmelin, pers. comm., May 2021) in the Iberian-Provençal basin, at the transition between coralligenous and semi-dark cave habitats. Microporella appendiculata has been also collected in dark sectors of Agios Vasilios cave in Lesvos, Greece (Rosso et al. 2019b) and (as M. umbracula) in the Mitigliano cave in the Sorrento peninsula, Italy (Balduzzi et al. 1989;Balduzzi and Rosso 2003).
Autozooids hexagonal, 460-522 (494±31, N = 3) × 411-476 (433±37, N = 3) μm (mean L/W = 1.16), boundaries marked by narrow, sinuous grooves and/or a raised rim. Frontal shield slightly convex, densely and evenly pustulose, with 11-25 circular  B29b2. 20.11.2020 E, G Gulf of Taranto, Porto Cesareo MPA, sample PCI 10, Paratype, PMC. B29b1. 20.11.2020 A two colonies nearly completely covering subsequent segments of the green alga Halimeda tuna B peripheral colony portion, showing the transition from the older autozooids (ovicellate, bottom right) to younger autozooids (non-ovicellate, top left) C group of autozooids, one with six oral spine bases (centre left) D close-up of two ovicellate zooids with the typical personate ovicell forming a collar bridge between the orifice and the ascopore E close-up of the orifice with five spine bases and smooth hinge-line with two rectangular condyles at corners F unbleached group of ovicellate autozooids with well-developed collars and avicularian mandibles G group of bleached autozooids, some ovicellate, fouled by algae. Note the collar developing from converging lateral lappets, the vizor-like band of gymnocystal calcification leaving visible the proximalmost pair of oral spines. Scale bars: 1 mm (A); 500 μm (B, C, F); 200 μm (D, G); 100 μm (E).
Ancestrula tatiform partially overgrown (four spines still visible) and regenerated as an autozooid without avicularium.
Etymology. From the Latin prefix bi-, two/double, and the adjective collaris, pertaining to the neck, referring to the bridging structure between the orifice and the ascopore appearing as a double collar.
Remarks. Four species with personate ovicells are known to date from the Mediterranean. Microporella coronata (Audouin & Savigny, 1826) differs from the new species in having paired avicularia and a greater number of oral spines, always hidden in ovicellate zooids. Microporella browni Harmelin, Ostrovsky, Cáceres-Chamizo & Sanner, 2011, M. genisii (Audouin & Savigny, 1826, and M. orientalis Harmer, 1957 differ in having personate ovicell structures not enclosing the ascopore, and by the denticulation either on the distal or the proximal margin of the orifice. Among all Microporella species known worldwide, the most similar to M. bicollaris sp. nov. is the eastern Pacific M. pontifica Osburn, 1952 reported from Clarion Island, Galapagos and the Gulf of California. Unfortunately, SEM images are not available for this species, but the original drawing (Osburn 1952: pl. 44, fig. 5) shows the same personate structure of the ovicell observed in M. bicollaris sp. nov. However, the new species differs in having a larger avicularium placed more terminally relative to the lateral margin of the zooid, and by the presence of condyles in the orifice. The specimen drawn in Hayward and Ryland (1999: fig. 136D) as Microporella ciliata "personate" form of Hincks (1880), also appears similar to M. bicollaris sp. nov. However, the illustration in Hincks (1880) appears different, but it is unclear whether Hayward and Ryland (1999) examined any additional material. The north-eastern Atlantic specimens need to be revised to assess their conspecificity with the Mediterranean colonies.
Distribution and ecology. Microporella bicollaris sp. nov. is presently known only from Porto Cesareo MPA (Gulf of Taranto, southwestern Apulia, NE Ionian Sea), and the Mazzere submarine cave in the Plemmirio MPA (western Ionian Sea). All colonies are from shallow waters, collected in photophilic algae or found in a semi-dark submarine cave. Remarks. The neotype chosen by Kukliński and Taylor (2008) was selected from material collected in 1984 and 2005 at Penta Palummo Bank, a submerged volcanic edifice in the Bay of Naples rising up to about 40 m depth from the surrounding 80 m shelf. Kukliński and Taylor (2008) already suggested that the Mediterranean material overall assigned to M. ciliata, might actually correspond to more than one closely related species. This was observed, for instance, in SEM images of north Adriatic specimens in Hayward and McKinney (2002), which differ from the neotype. Following Kukliński and Taylor (2008), , Chimenz Gusso et al. (2014), and Rosso and Di Martino (2016) remarked on the need for a revision of the material assigned to this species, whose accorded great variability (e.g., Gautier 1962;Hayward and Ryland 1999) is very likely to correspond to a species complex. The effort to disentangle the ciliata complex began with the recent description of Microporella modesta Di Martino, Taylor & Gordon, 2020, and is pursued in the present paper with the description of two new species sharing with M. ciliata the general appearance and the presence of a single, lateral avicularium as well as non-personate ovicells. Examination of all material available to us, as well as images available in the literature, suggest that M. ciliata, as defined by the neotype, is possibly a rare species in the Mediterranean Sea, or at least a species with a very restricted geographical and/or ecological distribution. A similar conclusion was reached by Chimenz Gusso et al. (2014) after revising their material identified as M. cf. ciliata, found associated to Posidonia and algae between 5 and 88 m depth in several localities of the Tyrrhenian Sea (i.e., off Latium; Spargiotto Island, NE Sardinia; Ustica and Volcano Islands) and the Sicily Strait (Lampedusa Island). The revision of all specimens originally assigned to M. ciliata or M. gr. ciliata in the collections of one of us (AR), originating from several regions and habitats of the Mediterranean Sea, led to the detection of only one colony corresponding with the neotype (Fig. 4). Although the type locality is unknown, it is likely that the original material described by Pallas (1766) was collected in shallow waters from easily accessible substrates, which is not the case for the selected neotype. Colonies of M. ciliata recorded from different western Mediterranean localities (Gautier 1962;Zabala 1986, and references therein), Greece (Harmelin 1969) and Turkey (Ünsal 1975), all show great variability and need to be revised to ascertain their identity.

Microporella ciliata (Pallas, 1766)
Distribution. Awaiting a general revision of specimens reported from all over the world, focusing on the Mediterranean Sea, the occurrence of Microporella ciliata to date is only confirmed in the Gulf of Naples (SE Tyrrhenian Sea) and the Gulf of Noto (W Ionian Sea). Diagnosis. Colony encrusting, multiserial. Autozooid frontal shield densely pustulose and sparsely pseudoporous. Orifice transversely D-shaped; hinge-line smooth with blunt condyles close to corners; four thin oral spines, hidden in ovicellate zooids. Ascopore field semi-elliptical; ascopore opening an arched fissure marked by a distal tongue with radial spines. Avicularium usually single, same level as or proximal to the ascopore, occasionally paired, directed distolaterally; crossbar complete; rostrum lanceolate, channelled. Ovicell non-personate.
Autozooids usually hexagonal to rhomboidal but sometimes irregularly shaped, 307-587 (434±73, N = 20) × 284-439 (357±59, N = 20 μm) (mean L/W = 1.21), boundaries marked by narrow grooves and raised rims of lateral walls (Figs 5F, 6B, G). Frontal shield slightly convex with polygonal and flat-topped pustules giving a tessellate appearance, and pierced by circular (diameter 6-12 μm), irregularly distributed pseudopores, usually numbering 10-30 but more numerous in some colonies; 2-4 marginal areolae, elliptical to fissure-like, barely visible or distinguishable from pseudopores (Figs 5F, 6B). The straight, short, setiform mandible has a hook at about one-third of its length which clamps it to the rostrum tip C bleached autozooids showing size and shape variability. Note that the majority of autozooids has a single avicularium, while few autozooids have paired avicularia (white asterisks) or none (black asterisks) D irregularly-shaped autozooids and kenozooids along the contact zone of lobes E autozooids and kenozooid (asterisk) near the colony margin with porechamber windows visible along exposed lateral walls F autozooid at the growing edge showing the morphology of the orifice and four, thin spine bases. Scale bars: 500 μm (A, C, D, E); 50 μm (B); 100 μm (F).
Kenozooids smaller than or nearly as large as autozooids, lacking openings such as orifices and ascopores but sometimes equipped with avicularium ( Fig. 6A, F).
Ancestrula not observed. Etymology. From Ichnusa the Latinized form of the ancient Greek name for Sardinia. Remarks. Size and shape of autozooids vary remarkably within and between colonies, including dwarf-like autozooids, about half the size of the more regular ones, as well as extremely large and irregularly shaped autozooids, appearing as the result of the fusion of contiguous autozooids (Fig. 5C, E). In this latter case the avicularium can be placed much more proximally than in regular autozooids. Irregularly-shaped autozooids also occur in M. browni (Harmelin et al. 2011: fig. 3b). These unusual autozooids and the kenozooids observed in this species seem to be particularly common in zones of contact between colonies or lobes of the same colony, and in damaged areas, also associated with evidences of reparation (Figs 5D, 6A, D), such as regeneration of broken autozooids, patches of calcification to close holes in the frontal shield, or orifices of presumably not functional autozooids (Fig. 6G). Intrazooidal budding, a feature that is common in bryozoans from submarine caves (e.g., Rosso et al. 2020aRosso et al. , 2020b, has been more commonly observed in avicularia (Fig. 6C, D) than autozooids (Fig. 6F, G). The occurrence of ovicells seems rare, observed only on the colony selected as the holotype.
The general appearance of this new species is very similar to M. ciliata. However, the orifice in M. ciliata, although of comparable size (0.06-0.08 mm long by 0.11-0.15 mm wide), is proportionately shorter, the hinge-line shows a series of median denticles .2020 A colony portion with broken ovicells, autozooids showing evidences of reparation, and kenozooids with (white asterisk) and without (black asterisk) avicularia B autozooids with complete or broken ovicells C paired autozooids, seemingly repaired, one lacking avicularium D close-up of a colony portion with evidence of zooidal repair. Note the occluded orifice indicated by the partly protruding spines (see arrow) E group of zooids (two ovicellate), some with open or closed mandibles and one lacking an avicularium F close-up of two zooids. Owing to the absence of an ascopore, which is usually placed at the same level as the avicularium, the upper one is more likely to be a kenozooid equipped with an avicularium than an autozooid with obliterated orifice. Note also the different frontal shield texture of its proximal margin, likely due to ovicell resorption G close-up of an autozooid with sealed orifice (central zooid), and an orifice with a thin secondary rim, indicating intramural budding presumably as a result of predation. Scale bars: 500 μm (A); 200 μm (B-G). and the two lateral condyles are more prominent and more laterally placed ( Fig. 4E; see also Kukliński and Taylor 2008: fig. 1G). The type and position of the oral spines are similar but the number of spines is 4-6 (more commonly four) in M. ichnusae sp. nov. and 1-4 (and occasionally lacking in the zone of astogenetic repetition) in M. ciliata (Kukliński and Taylor 2008). In M. ciliata the frontal avicularium is constantly single, only lacking in the first autozooid budded from the ancestrula, and no kenozooids were reported (Kukliński and Taylor 2008). Furthermore, the ovicells in M. ciliata have length comparable with those of M. ichnusae sp. nov. but are much narrower.
Distribution and ecology. Microporella ichnusae sp. nov. is presently known only from submarine caves in the Capo Caccia-Isola Piana MPA, in NW Sardinia. However, it is possible that some previous records of M. ciliata, to date the only Microporella species with a single avicularium considered as widespread in the Mediterranean, belong to this species. Fig. 7 Microporella modesta Di Martino et al. 2020a: 23, fig. 11  A bilaminar branch with elongate autozooids along branch edge and a cluster of ovicells B apparently cylindrical branch resulting from a colony encrusting an internode of Cellaria. Note the formation of irregularly elongate autozooids C unbleached colony encrusting a soft algal frond. Prominent avicularian rostra and mucros associated with ascopore contribute to the spiky appearance of the colony. Note the co-occurrence of ovicells with and without frontal tubercles D bleached colony on algae with ovicells with different degrees of frontal tubercle development E ovicellate zooids of a tubular "pseudovinculariiform" colony originally developed on a thin ephemeral substratum F originally tatiform ancestrula regenerated as a kenozooid, budding two distolateral autozooids lacking avicularia. Scale bars: 1 mm (A); 200 μm (B, E, F); 500 μm (C, D). Remarks. Microporella modesta has been recently established ) for some Mediterranean specimens collected off Algeria and growing as erect, narrow, bilaminar fronds. However, SEM images in Hayward and McKinney (2002) and Chimenz Gusso et al. (2014) reveal that a large part of the figured Mediterranean specimens seem to be conspecific with M. modesta, although none of them has been described as erect bilaminar. This is also the case for most colonies and colony fragments in the collection of one us (AR), mostly previously assigned to M. ciliata, which allowed us to assess the intraspecific variability of M. modesta.

Microporella modesta Di Martino, Taylor & Gordon, 2020
Strap-like branches including up to 11 or 12 longitudinal series of zooids occur only occasionally (Fig. 7A), while bilaminar arched fronds narrowing proximally are rare, and encrusting sheets are the most common. Branches are always broken off so that the way they form from the putative encrusting base remains unknown. Lateral edges often consist of extremely elongate autozooids covering the length of three regular autozooids, with the proximal two thirds about half the width of the distal part (Fig. 7A). Apparently cylindrical branches result from colonies encrusting thin, cylindrical substrata (Fig. 7B, E), either mineralised and persistent (mostly internodes of Cellaria in deep-water samples, and geniculate coralline algae in shallow-water sam-ples), or organic and ephemeral. In this latter case, the resulting pseudovinculariiform colonies are found in the sediment as hollow tubular fragments after substratum decay. A great number of colonies were observed encrusting soft algae, small Peissonnelia spp., and leaves of Posidonia oceanica (Linnaeus) Delile 1813, usually developing relatively small colonies, occasionally reaching 1-2 cm 2 .
In addition to colony morphology, variability includes also autozooidal characters. The orificial condyles were less prominent and more laterally placed in specimens from the Apollo Bank; the number of spines is usually five in encrusting colonies from the Ionian Sea (e.g., Amendolara and Ognina), but up to six or seven on some autozooids in colonies from the Sicily Strait (e.g., Madonna cave); spines also tend to be thicker, and the proximal pair more visible on ovicellate autozooids (e.g., in colonies from caves of the Ionian Sea and Sicily Strait) than in the holotype; a prominent central umbo can develop on the ovicell of some zooids (Fig. 7C, D). As in Microporella ichnusae sp. nov., irregularly shaped elongated autozooids develop when lobes of a colony join, mostly in colonies on cylindrical substrata (Fig. 7B). Intrazooidal budding frequently occurs in avicularia, while repair of autozooids was rarely observed.
The ancestrula was observed for the first time only in two colonies (sample PSI1 and ECE5), seemingly because it is soon overgrown. It is tatiform, with 10-12 spines surrounding a very thin, raised rim without apparent cryptocyst. It is rebudded as a kenozooid with a row of at least eight pores in one case (Fig. 7F). A comparable structure was figured in Kukliński and Taylor (2008) for M. ciliata and can be also observed in M. browni from the Maldive Islands (https://www.univie.ac.at/Palaeontologie/ Sammlung/Bryozoa/Maldive_Islands/Cheilostomata/Microporellidae/Microporella/ Microporella-browni.html). The ancestrula buds two distolateral autozooids with 5-7 oral spines.
Ovicells are very numerous in this species, occurring in the majority of autozooids in some colonies (Fig. 7A, C, E). Ovicell tubercles when developed, along with the mucro associated to the ascopore, and the raised rostrum of the avicularium give to the colony a characteristic knobby or spiky appearance.
The majority of specimens recorded from the Mediterranean previously assigned to M. ciliata, as well as some isolated colony fragments attributed to M. orientalis seem to belong to M. modesta.
Distribution and ecology. Microporella modesta was previously reported exclusively from off Algeria ), but the material studied here as well as specimens figured in recent literature (Hayward and McKinney 2002;Chimenz Gusso et al. 2014) widen its geographical distribution. The species appears to be widely distributed in the Mediterranean where it occurs in the Iberian-Provençal basin, the Tyrrhenian, Ionian, and Adriatic seas as well as in the Sicily Strait. The species seems widely distributed in shallow shelf habitats with preference for the vegetate bottoms of the infralittoral zone and the Coralligenous Biocoenosis, with the possibility to extend into semi-dark caves. Findings from deeper (mid-shelf ) particulate bottoms are largely represented by dead colonies/fragments possibly indicating a displacement. Diagnosis. Colony encrusting, multiserial. Autozooid frontal shield granular and centrally pseudoporous. Orifice transversely D-shaped; hinge-line smooth with rectangular condyles at corners; five (more commonly) to eight oral spines, the proximalmost pair placed slightly below the orifice hinge-line and very large in diameter. Ascopore field reniform to elliptical, developing a mucro proximally; ascopore opening transversely C-shaped, with tongue and radial spines. Avicularium single, located at half zooidal length, directed distolaterally; crossbar complete; rostrum triangular, channelled. Ovicell produced by distal autozooid, non-personate.
Ascopore field a reniform to elliptical area of smooth gymnocystal calcification (39-78 × 45-102 μm), more extensive proximally, developing a pointed, upward directed mucro not concealing the ascopore opening, placed 30-60 μm below the orifice, slightly depressed B31a.3.12.2020 F, G paratype PMC. B31b. 3.12.2020 A colony general view B close-up of orifice with proximolateral rectangular condyles, smooth hinge-line, and five oral spine bases. Note the giant proximolateral pair of spine bases located at hinge-line level and at some distance from the thinner, distal ones C close-up of avicularium with truncated, channelled rostrum projecting laterally outside its edge affecting the shape of the adjacent zooid margin D close-up of ascopore with proximal mucro E ovicells smooth and imperforate centrally, finely granular and with a row of large pores peripherally. Radial buttresses between marginal pores converge towards the median umbo F Ancestrula budding one distal and two distolateral autozooids. Two larger, proximolateral, curved autozooids complete the periancestrular region G autozooids from the colony growing margin with six or seven oral spines. Scale bars: 1 mm (A); 50 μm (B-D); 200 μm (E-G).
Etymology. From the Greek pachys, meaning thick, and the Latin spina meaning spine, referring to the robust proximalmost pair of oral spines.
Remarks. The main diagnostic character of Microporella pachyspina sp. nov. is the great size of the proximalmost pair of oral spines, as well as their position, halfway below the level of the orifice hinge-line. Among Microporella species known worldwide, M. alaskana Dick & Ross, 1988 from the eastern Pacific, M. echinata Androsova, 1958, andM. trigonellata Suwa &Mawatari, 1998, both from off Japan, share similar features. In M. alaskana the proximalmost pair of spines are larger in diameter compared to the remaining spines but they are placed more distally compared to the new species, approximately at orifice mid-length (Dick and Ross 1988); in addition, this species has paired avicularia, the ascopore is placed very close to the orifice hinge-line, and there is a prominent umbo centrally on the frontal shield (Dick and Ross 1988). Microporella echinata differs in having an evenly pseudoporous frontal shield, with pseudopores visibly larger, and tubercular (Mawatari et al. 1991). Microporella trigonellata shows the same number, relative size, arrangement and position of spines but differs from the new species in having the avicularium placed distolaterally, directed distally, with a pointed, non-channelled rostrum; also, the ridges and grooves on the ovicells are distinctly defined and more marked, and the ancestrula has a narrower proximal gymnocyst (Suwa and Mawatari 1998).
The general appearance of those zooids lacking avicularia in M. pachyspina sp. nov. reminds those of Fenestrulina joannae (Calvet, 1902), which are also similar in having the proximalmost pair of spines long, robust and rounded, non-stellate pseudopores sparse on the frontal shield, centrally smooth ovicells, sometimes with peripheral radial ridges, developing a mucro (Chimenz Gusso et al. 2014: 165, fig. 82a-c). This latter species, originally described as Microporella by Calvet (1902), was reassigned to Fenestrulina by Gautier (1962: 171) apparently based on a suggestion made by Hastings without any supporting statement, instead highlighting the different type of pseudopores (non-stellate) compared with those of the type species of the genus Fenestrulina malusii Audouin, 1826. Subsequent authors followed Gautier (1962). Fenestrulina species have large, stellate pseudopores mostly occupying the area of the frontal shield between the ascopore and the orifice, a sector that is usually imperforate in Microporella. Based on these observations, Fenestrulina joannae seems to have more affinities with Microporella and here we suggest its displacement.
Dry specimens on organic substrates (i.e., Posidonia leaves) appear with the zooids disconnected or almost disconnected, giving to the colony a slightly disjunct appearance because the zooids were less packed hence exposing a more extensive, smooth gymnocyst laterally (Fig. 8F). This loose packing is a common adaptation in species growing on flexible substrates to reduce the potential breakage of the zooidal skeletons.
Remarks. This species differs from its Mediterranean congeners in having a finely reticulate ascopore but it is left in open nomenclature owing to the availability of a single, infertile colony fragment. Similar ascopores can be found in M. arctica Norman, 1903 from Norway, M. ketchikanensis Dick, Grischenko & Mawatari, 2005 from Alaska, M. santabarbarensis Soule, Chaney & Morris, 2004 from southern California, and M. stellata (Verril, 1879) from Maine, USA. Microporella arctica differs from Microporella sp. A in having a finely granular frontal shield pierced by a greater number of marginal areolae that are always very distinct from pseudopores, in the lack of oral spines, and in having a smooth gymnocystal area laterally and proximally to the orifice that is continuous with the gymnocyst of the ascopore field (Kukliński and Taylor 2008). The ascopores of both M. ketchikanensis and M. santabarbarensis have a similar, delicate cribrate aspect but lack the distal tongue extending from the distal edge (Dick et al. 2005;Soule et al. 2004). Microporella stellata differs in having only two oral spines and a proximal orifice margin with broad, rectangular condyles (Winston et al. 2000).
Distribution and ecology. Presently known only from the Palinuro Peninsula, along the Tyrrhenian coast of Campania (southern Italy). A dead colony was collected from the biogenic muddy sediment covering the floor of a completely dark sector of the Scaletta submarine cave, at 46 m depth where the colony presumably lived.
Branches cylindrical, often flattened at bifurcations (Fig. 10C, E), consisting of 9-16 longitudinal rows of zooids, alternating very regularly to simulate an helicoidal pattern; branch diameter 10-15 mm (exceptionally up to 20 mm), becoming thicker and stouter in older portions of the colony, near the encrusting base, owing to secondary calcification.
Older colony parts thickly calcified owing to secondary calcification progressively obliterating zooidal openings including orifices, ascopores and avicularia (Fig. 10D), making difficult the distinction between old autozooids and genuine kenozooids that probably also develop.
Remarks. First assigned to Eschara (Peach 1868), Eschara verrucosa served as the type species of the genus Diporula Hincks, 1879 in which it has been included since then with the exception of Neviani (1896a, b). Characters used to distinguish Diporula from Microporella included the "dendroid zoarium with cylindrical branches" and the morphology of the orifice described as "expanded above, contracted below, and slightly constricted by lateral projections (horseshoe-shaped)" (Hincks 1880: 220;and similar description in Gautier 1962: 176). However, both characters seem feeble to justify the distinction between the two genera. At least nine species of Microporella possess erect colony-growth, starting with a more or less developed encrusting portion as does Diporula. Also the shape of the orifice in Microporella species is highly variable , with several examples of species having orifices with the proximal margins narrower than the orifice maximum width [e.g., Microporella curta Almeida, Souza, Mengola & Vieira, 2017 from Brazil, Microporella franklini (Soule, Chaney & Morris, 2003) from California, the Mediterranean Microporella genisii (Audouin & Savigny, 1826), Microporella hastingsae Harmelin, Ostrovsky, Cáceres-Chamizo & Sanner, 2011 from the Red Sea, and the Arctic Microporella klugei Kukliński & Taylor, 2008].
Further differences between Microporella and Diporula were highlighted by Hayward and Ryland (1999: 292, 312), including interzooidal communications via basal pore-chambers in the former genus and multiporous septula in the latter, and the presence of pseudopores in the ovicells of Diporula. However, multiporous septula were observed, for example, in Microporella ordo (see Di Martino et al. 2020a: fig. 7D), and basal pore-chambers were observed in the encrusting portions of M. verrucosa, while pseudopores occur in the ooecium of many Microporella species including the type M. ciliata (see Kukliński and Taylor 2008: fig. 1c). A further presumed difference relates to the ooecium porosity, with Diporula reported as having a fully perforated endooecium and Microporella species usually described as having only pits in the endooecium (Harmelin et al. 2011: 2;Ostrovsky 2013: figs 2.43B-D, 2.44A). However, pores clearly perforate the endooecium also in Microporella as seen in broken ooecia of M. ichnusae sp. nov. (Fig. 6B) Based on these observations, here we propose Diporula as junior synonym of Microporella and resurrect the combination Microporella verrucosa first proposed by Neviani (1896a, b). Specimens of a second species of Diporula, D. coronula Ortmann, 1890 need re-examination. Based on the original description and illustration (Ort-mann 1890: 39, pl. 3, fig. 7), this species has a lepralioid orifice with condyles at about one-third of the orifice length, a single avicularium with spathulate mandible, and up to two frontal foramina, characters reminiscent of other cheilostome genera such as, for example, Poricella Canu, 1904.
Specimens originally described as Eschara lunaris Waters, 1878, from Pleistocene sediment of eastern Sicily and synonymised with M. verrucosa by Hincks (1880) need to be re-examined as well to confirm their conspecificity, but this is out of the scope of the present paper.
The rugose appearance observed by Peach (1868) and Hincks (1880), which inspired the species name, was not observed in our material, although secondary calcification is always very common in older parts of the colony. Intramural budding is rare and restricted to avicularia, while branch regeneration is common, apparently following breakage as indicated by broken autozooids with sharp edges. We also observed zooids with reverse polarity, sometimes budded from old stems with autozooids obliterated by secondary calcification. However, in these instances few whorls of autozooids usually develop from the regeneration surface, with only few tips appearing actively growing. Colony fragments longer than 2 cm can appear twisted, a morphology observed in some cyclostomes (Harmelin 1976) and other erect cheilostomes from the Gulf of Noto and the Ciclopi MPA area (Rosso 1989). This twisted branch morphology and the ability to regenerate after breakage might represent the adaptation of this species to colonize soft sediment bottoms. Basal, encrusting colony portions are relatively common in our samples and show that the ancestrulae settled on clasts ranging from a few mm to 1-2 cm in size. The ability of this species to encrust small particles, in addition to large substrates in rocky habitats, was suggested by Gautier (1962) after finding only colony fragments in dredges from sandy-muddy bottoms.
The diagnostic characters of this species seem constant in the Mediterranean specimens, except for the size of the ascopore related to the development of the distal tongue sometimes leaving only a fissure-like opening. Paired avicularia were observed only in one autozooid (Fig. 11E). Higher variability is observed when comparing the Mediterranean specimens with those from the Atlantic (e.g., Hayward and Ryland 1999: 302, figs 138C, D, 139A, B; unpublished SEM images provided by P.D. Taylor from Mauritania and Madeira) related to the ascopore shape, the size of the spines, and the distribution and size of pseudopores on the frontal shield, suggesting the existence of a species complex.
Distribution and ecology. Microporella verrucosa is a warm-temperate species with Atlanto-Mediterranean distribution. In the Atlantic, it has been reported from West Africa to the southwest of the British Isles (Hayward and Ryland 1999); in the Mediterranean, it occurs preferentially in mid-and outer-shelf habitats below 50-60 m depth, with an optimum at 70-120 m (Gautier 1962;Zabala 1986;Rosso 1989Rosso , 1996aRosso and Di Martino 2016), but it was also observed at shallower depths (20 m) in a shadowed open cave in Catalonia (André et al. 2014). It is associated with shadowed rocky habitats, including the Coralligenous and Semi-Dark and Dark Cave Biocoenoses, and detritic habitats, such as the Coastal Detritic and the Offshore Detritic Biocoenoses (Table 1; Gautier 1962;Harmelin 1969Harmelin , 1976Zabala 1986;Rosso 1989Rosso , 1996aRosso , 1996bDi Geronimo et al. 1990;Madurell et al. 2013;Rosso et al. 2014Rosso et al. , 2019bGerovasileiou and Rosso 2016). However, it is never very common or dominant at sample or habitat scale, occurring only with a few colonies per sample and/or in one out of four or five sampling stations (Table 1; see also Harmelin 1976: tables 1, 3).

Discussion
The present study increases the number of Microporella species known from the Mediterranean from nine (Rosso and Di Martino 2016;Di Martino et al. 2020a) to 14 (see Table 2 for a summary of species characters). The presence of M. orientalis and M. coronata remains based only on the records of Hayward (1974) from Chios and Harmelin et al. (2011) from Lebanon. Based on the description and illustrations of the ovicell as globose but non-personate (e.g., Zabala 1986: 513, fig. 180), further records of these two species are more likely to correspond to M. appendiculata. Microporella marsupiata, although listed in Rosso and Di Martino (2016), can be considered restricted to the Atlantic. Mediterranean specimens identified as M. marsupiata (Zabala 1986) belong to M. appendiculata. This also applies to some specimens in our collection. A. This is the result either of the re-examination of previously studied material assigned to the M. ciliata species complex or the examination of new material. Microporella bicollaris sp. nov. is clearly distinct from M. ciliata, given the presence of a personate ovicell. Microporella ichnusae sp. nov., M. pachyspina sp. nov., and the recently described M. modesta, on the other hand, have major affinities and share a certain number of features with M. ciliata. Furthermore, these species show high intracolonial and intraspecific variability common to other Microporella species and groups of species (e.g., Harmelin et al. 2011).
Most often, variability concerns the number of oral spines (e.g., 4-6 in M. ichnusae sp. nov.), 5-7 in M. modesta, 5-8 in M. pachyspina sp. nov., and 0-4 in M. ciliata), and the development of a tubercle on ovicells with the co-occurrence of tuberculate and non-tuberculate ooecia as in M. modesta. In this latter species, the mucro associated with the ascopore, considered as a diagnostic character by Di Martino et al. (2020a), also varies remarkably in the encrusting colonies examined here. In contrast, the size, density and distribution pattern of pseudopores on both autozooids and ovicells, as well as characters of the orifice, including the shape and size of condyles and denticulation of the hinge-line, are more stable. Harmelin et al. (2011) considered the variability of the denticulation of the orifice hinge-line in M. harmeri Hayward, 1988 (associated with differences in the morphology of the ascopore and in the shape of the avicularium and its mandible) as the indication that this widespread species, actually corresponds to a species complex. Periancestrular autozooids can lack avicularia in some species (e.g., M. luellae Grischenko, Dick & Mawatari, 2007, see Grischenko et al. 2007: fig. 34F), but the number of avicularia also seems constant within species, although sporadically autozooids may lack avicularia (as in M. ichnusae sp. nov.) or bear a second avicularium (as in M. ichnusae sp. nov. and M. verrucosa). Paired avicularia were also described in "M. ciliata" by Zabala (1986) and appear in Microporella ketchikanensis Dick, Grischenko & Mawatari, 2005(Dick et al. 2005 fig. 20F) and M. stellata (Winston et al. 2000: fig. 19). Autozooids lacking avicularia have been occasionally reported in Microporella species, such as the Pliocene Microporella sarasotaensis Di Martino, Taylor & Portell, 2019 from Florida (Di Martino et al. 2019: fig. 37), the Pleistocene M. rusti from New Zealand (Di Martino et al. 2017: fig. 4), and the present-day M. maldiviensis Harmelin, Ostrovsky, Cáceres-Chamizo & Sanner, 2011 from the Maldive Islands (Harmelin et al. 2011: fig . 12I), and M. ketchikanensis (Dick et al. 2005: fig. 20A). The three latter species also show autozooids developing paired avicularia, thus paralleling M. ichnusae sp. nov. for the co-occurrence of autozooids with 0, 1, or 2 avicularia. Kenozooids, with or without an avicularium, are here mentioned for the first time for Mediterranean species. Kenozooids were observed in M. ichnusae sp. nov., M. modesta and M. verrucosa, often along the contact zone between two different colonies or lobes of the same colony, as already seen in fossil species of Microporella involved in competitive interactions .
Of the Mediterranean Microporella species examined here, M. verrucosa and M. modesta are the most frequently recorded (Fig. 1). Microporella verrucosa is also the only Mediterranean erect species with cylindrical branches forming relatively large colonies, therefore easily recognisable. The identical zooidal morphology of erect species of Microporella (M. hastigera, see Di Martino et al. 2020a) with those originally assigned to Diporula, such as M. verrucosa, led to the suggested synonymy between these two genera. The main difference between the two genera related to the porosity of the ooecium does not hold (see also Remarks for M. verrucosa): Microporella species can have either evenly pseudoporous ooecia with pores penetrating the endooecium (e.g., M. ichnusae sp. nov., Fig. 6B), or ooecia with a single row of marginal pores imperforate centrally (e.g., M. pachyspina sp. nov., Fig. 8E), as well as imperforate ooecia (e.g., M. appendiculata, Fig. 2E). The lack of pseudopores in the ooecium is also a feature of the Langhian M. berningi Zagorsek, 2010 from the Czech Republic, which also shares with M. appendiculata the formation of ooecia by kenozooids that are therefore independent of the distal autozooids, which is likely the reason for the lack of peripheral areolar pores (Fig. 2C, D). Although based on a single species, "M. ciliata", the ooecia in Microporella have been defined as microporelliform (i.e., an outgrowth of the proximal part of the frontal shield of the distal autozooid) (Ostrovsky 2013: 141, fig. 2.43A-D), which does, therefore, not apply to at least some of the species.
Another Microporella-like genus, Flustramorpha Gray, 1872, recorded off the coast of South Africa, is likely to be also indistinguishable from Microporella. However, species assigned to Flustramorpha need to be revised before the synonymy can be considered.
Microporella modesta has high growth plasticity and, although first described as erect bifoliate, encrusting colonies are more common. No habitat information is available for the type locality off the coast of Algeria, while colonies studied here are from shelf habitats. At shallow depths, M. modesta preferentially lives as an epibiont of soft and ephemeral substrates such as seagrasses and algae, and also on geniculate corallines, while in deeper environments it is mainly associated with other bryozoans. Despite the large number of colonies examined, no evidence was found of bilaminar branches starting from the encrusting base. The revision of large bryozoan collections from the Mediterranean (i.e., Rosso's collection, this study; Chimenz Gusso's collection in Chimenz Gusso et al. 2014), has shown that those species, including M. ciliata, previously considered widespread in the Mediterranean, are instead restricted to some specific areas and habitats. For instance, Microporella ichnusae sp. nov. and M. sp. A were found only in submarine cave habitats, with the former species inhabiting relatively lit caves of NW Sardinia (E Iberian-Provençal basin) shallower than 10 m, and the latter species collected dead at 46 m depth in a completely dark cave of the Palinuro Peninsula (E Tyrrhenian Sea). Microporella bicollaris sp. nov. and M. pachyspina sp. nov. seem to be southern species, inhabiting shallow-water habitats from the Sicily Strait, mostly associated with the Infralittoral Algae and the Posidonia meadow Biocoenoses.
In summary, based on the above discussion, some characters were observed for the first time in Microporella: tatiform ancestrula regenerated as an autozooid or kenozooid; presence of kenozooids; interzooidal communication through basal pore cham-ber windows in encrusting species or encrusting base of erect species and multiporous septula in erect branches; ovicells of kenozooidal origin; erect species with cylindrical branches (owing to the displacement of M. verrucosa).

Conclusions
With the description of three new species (i.e., M. bicollaris sp. nov., M. ichnusae sp. nov., M. pachyspina sp. nov.), and the inclusion of M. verrucosa and M. joannae, this study confirms Microporella as one of the most species-rich genera of the Mediterranean, after Schizomavella and Reteporella, with 22 and 15 known species, respectively (see Rosso and Di Martino 2016). Three or even four Microporella species, such as those recorded or described from the Mediterranean eastern sectors, are considered non-indigenous and possibly introduced in recent times, mostly via the Suez Canal (Harmelin et al. 2011) as suggested by the fact that they are all shallow-water species (3-43 m depth) found in harbours, marinas, and nearby natural environments. For the new species described here, their status as native or alien species cannot be assessed with certainty, although two of them, M. bicollaris sp. nov., M. pachyspina sp. nov., but also M. sp. A, show their highest affinities with extra-Mediterranean species, specifically with species from the Pacific Ocean, including the Galapagos Islands and the Arctic, areas that are geographically distant and characterised by decidedly different environmental parameters. An anthropogenically-mediated transport might be hypothesised for the first two species that have a shallow-water distribution, often in association with algal fronds. However, a similar widely disjunct distribution was observed for the calescharid species Tretosina arculifera (Canu & Bassler, 1927), previously only known from deep-waters of Hawaii and recently discovered in submarine caves from Lesvos Island (Rosso et al. 2020b). To date, there is no evidence of the occurrence of Microporella species in highly human-impacted areas (Ferrario et al. 2018), with the only exception of colonies of M. modesta found on plastic at Ognina marina, north of Catania.
Like for some other genera (e.g., Setosella Hincks, 1877; see Rosso et al. 2020a), the description of such a significant number of new species shows how incomplete our knowledge of the bryozoan biodiversity in the Mediterranean basin still is, despite being one of the most investigated areas during the past few centuries (see Rosso and Di Martino 2016, and references therein). 2000 project funded by the Italian Ministry of Environment and Capitaneria di Porto di Catania (PI A. Rosso); the CoNISMa projects MATTM 3AMP "Study of submarine cave environments-CODICE HABITAT 8330-in the Marine Protected Areas of Pelagie, Plemmirio and Capo Caccia-Isola Piana" (PI S. Fraschetti) and MATTM 4AMP "Analisi e valutazione dello stato degli ecosistemi marini delle zone A e B in 4 Aree Marine Protette anche al fine di valutare l'efficacia delle misure di gestione delle stesse"; the projects Noto 1996 and Piattaforma Siciliana (PS) funded by the Regione Sicilia and led by I. Di Geronimo and G. Cantone (University of Catania), respectively; the ENEA-CNR Minerva cruise 1996 at Amendolara Bank, Gulf of Taranto.