Adriatic calcarean sponges (Porifera, Calcarea), with the description of six new species and a richness analysis

In this study we analyze the calcarean sponge diversity of the Adriatic Sea, the type locality of some of the first described species of calcarean sponges. Morphological and molecular approaches are combined for the taxonomic identification. Our results reveal six species new to science and provisionally endemic to the Adriatic Sea (Ascandra spalatensis sp. nov., Borojevia croatica sp. nov., Leucandra falakra sp. nov., L. spinifera sp. nov., Paraleucilla dalmatica sp. nov., and Sycon ancora sp. nov.), one species previously known only from the Southwestern Atlantic (Clathrina conifera), and three already known from the Adriatic Sea (Ascaltis reticulum, Borojevia cerebrum, and Clathrina primordialis). We confirm the presence of the alien species Paraleucilla magna in the Adriatic and again record Clathrina blanca, C. clathrus, and C. rubra. We emend the description of the genus Ascaltis, propose a lectotype for Borojevia cerebrum and synonymise B. decipiens with B. cerebrum. A checklist of all calcarean species previously and currently known from the Adriatic Sea (39 species) is given. The Central Adriatic is indicated as the richest calcarean sponge fauna sector; however, the biodiversity of this class is underestimated in the whole Adriatic Sea and new systematic surveys are desirable.


Introduction
Although the biodiversity of sponges of the Mediterranean Sea has been studied for a long time (e.g., Schmidt 1862Schmidt , 1864Haeckel 1872), some regions and sponge taxa have remained rather neglected. The Adriatic Sea is one of the seven eco-regions of the Mediterranean Province (Spalding et al. 2007) and is considered a biodiversity hotspot (Bianchi et al. 2012) of major ecological importance. It forms a very narrow, semi-enclosed basin in the northernmost part of the Mediterranean Sea, subdivided into three sectors: Northern Adriatic, Central Adriatic and Southern Adriatic (Bianchi & Morri 2000). The karst limestone is known for forming unique habitats such as caves, overhangs and pits, which are rather inaccessible and often inhabited by a number of invertebrate groups, including calcarean sponges. They are generally less investigated, mainly because of a smaller number of experts involved in their taxonomy, and consequently a large number of species is still unknown. Several calcarean species are known from the Mediterranean, including the Adriatic (see Pansini & Longo 2008), but literature data on these Adriatic species is very scarce or difficult to access, moreover lacking a comprehensive and detailed morphological and molecular descriptions.  Table 1. Specimens included in the phylogenetic analyses with collection sites, voucher numbers and GenBank accession numbers. *Specimens with newly generated DNA sequences.
different methods were applied for phylogenetic reconstruction: maximum likelihood (ML) and Bayesian inference (BI). The Akaike Information Criterion (AIC) implemented in jModeltest 3.7 (Guindon & Gascuel 2003;Darriba et al. 2012) was used to select the best-fit models of sequence evolution. The models were chosen for each dataset as follows: for 28S analysis, GTR+I+G and TrN+I+G models were chosen for Calcinea and Calcaronea, respectively; for ITS analysis, the TrN+G model was chosen for both datasets. Phylogenetic analyses were performed in PhyML 3.0 (Guindon et al. 2010), where datasets were analysed by the ML method. Bootstrap tests of phylogeny were performed with 1000 replicates. Bayesian MCMC analyses were performed in MrBayes v. 3.1.2. (Huelsenbeck & Ronquist 2001), considering the same models for given datasets. Two parallel runs each comprising four Markov chains were run for 1,000,000 generations with a sampling frequency of one in every 100 trees; a consensus tree was constructed based on the trees sampled after burn-in of 100,000. Phylogenetic trees were generated separately for each dataset, rooted at midpoint and displayed in FigTree v. Leucosolenia reticulata -Dendy & Row 1913: 723. -Breitfuss 1932: 243. Leucosolenia reticulum -Dendy & Row 1913: 723. -Breitfuss 19301932: 243;1935: 14. -Topsent 19341936: 22. -Hôzawa 1940: 32. -Arndt 1941: 4. -Tanita 19421943: 386. -Burton 1963

Colour
White in life and white in ethanol.

Description
Cormus is composed of regular and tightly anastomosed tubes. Water-collecting tubes are present ( Fig. 2A). As the specimen was fragmented, it was not possible to observe the pseudoatrium. The skeleton is composed of one category of triactines, one of tetractines and diactines. Diactines are organised in tufts of two to five spicules, perpendicularly disposed in the tubes (Fig. 2B). Triactines are the most abundant spicules.
DiacTines. Slightly curved. The tip that protrudes through the surface is lanceolated (Fig. 2F). Trichoxeas are also present on the surface of the tubes. Size: 106.3/4.9 µm.

Ecology
Specimens were collected on a vertical, shaded hard limestone bottom. Klautau et al. (2013) proposed to transfer this species to the genus Ascaltis based mainly on morphological, but also on molecular data. Although the type species of this genus (A. lamarcki Haeckel, 1870) was not included in the molecular dataset, A. reticulum did not group with any of the included genera (Fig. 16).

Remarks
Besides, morphologically it is more similar to Ascaltis than to any other genus. Therefore, although the classification of A. reticulum in the genus Ascaltis must still be verified regarding the type species of the  (2016) genus, it was morphologically and molecularly proved that it cannot be included in the genus Clathrina. Hence, we keep the proposition of Klautau et al. (2013) and name this species A. reticulum.
This is the first time that spines were observed on the apical actine of the tetractines of A. reticulum. For that reason, we examined the neotype of this species and detected spines as well. They are abundant and very small. We also observed a great variation in the size of the diactines, which are much larger in the neotype (102.0-212.2 (±54.1)-306.0 /14.3 (±5.1) µm).

Ecology
The specimen was collected on a shaded, vertical hard bottom.

Colour
Light yellow in life and in ethanol.

Description
Cormus is composed of regular and tightly anastomosed tubes (Fig. 4A). Large water-collecting tubes are present. The skeleton consists of triactines, a few tetractines and tripods, which in fact are large triactines. It has no special organisation (Fig. 4B).
TeTracTines. Regular (equiangular and equiradiate). Actines are slightly conical to conical, straight, with sharp tips. Sometimes they are slightly undulated near the tips. It is possible to recognise two types of tetractines: small (Fig. 4E) and large (Fig. 4F). Large tetractines are the same size as tripods.
The apical actine of the tetractines is shorter than the basal ones, slightly conical, sharp and frequently curved only at the tip. It is ornamented with few (ca. six) spines, which are large, conical and cover only the last third of the apical actine. (Fig. 4G). Size: 81.9/8.5 µm (basal actine); 46.8/5.4 µm (apical actine).

Ecology
The specimen was collected on a semi-vertical hard limestone bottom.

Remarks
Similar to other species of Borojevia, B. cerebrum has thin, regular and tightly anastomosed tubes forming the cormus. The oscula are present at the end of water-collecting tubes. The skeleton is composed of tripods (with the characteristic elevated centre or similar to large triactines), triactines and tetractines. Individuals of B. cerebrum always have spines on the apical actine of their tetractines; however, in the same individual some tetractines may be smooth. In B. cerebrum, the spines are not very abundant; they are large and scattered, only near the tip of the apical actine. The Adriatic and Mediterranean specimens of B. cerebrum formed a well supported clade in the ITS tree ( Fig. 16), separated from the clade comprising B. brasiliensis (Solé-Cava, Klautau, Boury-Esnault, Borojević & Thorpe, 1991).
Borojevia cerebrum is the type species of the genus. Its type locality is Lesina (Island of Hvar) and it commonly occurs in the Mediterranean and the Adriatic Sea. The type specimen of B. cerebrum (PMJ-Inv. Nr. Porif. 156) is not very well preserved (Klautau & Valentine 2003); thus, we got a great opportunity to redescribe this species from near its type locality.
Analyses of other individuals of B. cerebrum from several sites in the Adriatic and Mediterranean Seas verify that the shape of the tripods is very variable. It varies from the characteristic shape of tripods, with stout actines and elevated centre, to only large triactines. This kind of variability may be assigned to polymorphism or plasticity. Indeed, Haeckel (1872) proposed two varieties of B. cerebrum (as Ascaltis cerebrum), based on the presence of either characteristic tripods or large triactines. The first variety he called B. cerebrum var. gyrosa, while the other one he considered B. cerebrum var. decipiens. Dendy & Row (1913) elevated B. cerebrum var. decipiens to species level (as Leucosolenia decipiens) and kept B. cerebrum (as L. cerebrum) as a valid species. The variety gyrosa had not been oficially elevated to the status of species; however, it was mentioned as Ascaltis gyrosa in a synonym list of B. cerebrum made by Burton (1963: 186).
Considering that both varieties were proposed only to differentiate specimens with characteristic tripods from those with only large triactines and that we found this morphological variation inside individuals and among specimens placed within the same species, we propose here the synonymisation of B. decipiens with B. cerebrum.

Description
Cormus composed of regular and tightly anastomosed tubes (Fig. 5A). Water-collecting tubes are present and form a single apical osculum. The skeleton is composed of tripods, triactines and rare tetractines. It has no special organisation (Fig. 5B).

Ecology
Specimens were collected on a shaded, vertical, hard limestone bottom.

Remarks
The genus Borojevia is currently composed of five species: B. aspina (Klautau, Solé-Cava & Borojević, 1994), B. brasiliensis, B. cerebrum, B. paracerebrum (Austin, 1996 and B. tetrapodifera (Klautau & Valentine, 2003). All of them show a very well defined cormus, with regular and tightly anastomosed tubes and water-collecting tubes. The skeleton is always composed of tripods, triactines and tetractines with spines on the apical actines. Tetrapods may also be present (B. tetrapodifera). The sixth species of the genus, B. croatica sp. nov., is closer to B. cf. aspina in our ITS tree (Fig. 16). Both species have short spines; however, B. croatica sp. nov. has numerous spines, while in B. cf. aspina there are few.
Given that B. cerebrum is also present in the Adriatic Sea, the best way to differentiate it from B. croatica sp. nov. is by the shape and location of spines. They are shorter, more abundant and distributed along most of the actine length in B. croatica sp. nov., and larger, fewer and scattered only near the tip of the apical actine in B. cerebrum.

Colour
White in life and white or brown in ethanol.

Description
Cormus composed of irregular and loosely anastomosed tubes (Fig. 6A). Water-collecting tubes are not present. The skeleton consists of triactines without organisation (Fig. 6B).

Ecology
Specimens were collected on a semi-shaded, vertical hard limestone bottom under overhangs. They were often found in association with the macroalga Ellisolandia elongata (J. Ellis & Solander, 1786).

Description
Cormus is formed by large and loosely anastomosed tubes. Water-collecting tubes are absent (Fig. 7A). The skeleton is composed of one category of triactines (Fig. 7B). The size of the spicules is very variable and it is therefore not possible to categorize them. (Table 7) TriacTines. Regular (equiangular and equiradiate). Actines are conical to slightly conical with sharp tips (Fig. 7C). Their size is very variable. Size: 121.5/12.2 µm.

Ecology
The specimen was collected on a shaded, vertical hard limestone bottom.

Remarks
Haeckel (1872) assigned the name Ascetta primordialis to a group of different species, and even genera, whose skeleton comprised only triactines, but, unfortunately, did not select a holotype. In 2003, Klautau & Valentine revised the genus Clathrina and analysed two specimens of C. primordialis, one from the Adriatic Sea (PMJ 154) and another one from Naples (ZMB 1306). Both specimens clearly represented different species and the authors suggested the specimen ZMB 1306 was the true C. primordialis, because C. primordialis (originally Prosycum primordiale Haeckel, 1870) was first described from Naples.
However, analysing the present specimen and re-analysing the slides of the specimens PMJ 154 and ZMB 1306 and the catalogue from the ZMB, we now have a different opinion. On the specimen's label and in the catalogue of the ZMB it is not noted that ZMB 1306 is a syntype of C. primordialis. Consequently, Klautau & Valentine (2003) should not have designated the specimen ZMB 1306 as a lectotype of C. primordialis. On the other hand, the label of the specimen PMJ 154 mentions it is a syntype of C. primordialis. Therefore, in our opinion, the specimen PMJ 154 is more reliably a true representative of this species and should be considered the lectotype of C. primordialis.
Considering the morphology of PMJ 154, the specimen IRB-CLB3 = UFRJPOR 6863 represents C. primordialis, as well as the specimen PMR 14305, recently published as C. cf. hondurensis Klautau & Valentine, 2003(Imešek et al. 2014. The similarities between C. primordialis and C. hondurensis made us ponder on the possibility of synonymy between these two species. However, as we could not obtain DNA sequences of C. hondurensis from the type locality (Honduras) to verify this, we prefer to keep C. hondurensis as a valid species restricted to the Caribbean Sea, until further analyses are done.

Description
The sponge is massive and vase-shaped, with one apical osculum without crown. The atrium is central and large. The aquiferous system is leuconoid (Fig. 8A). The sponge surface is smooth, but harsh. The cortical skeleton is composed of small, tangentially arranged triactines. The choanosomal skeleton has no organisation (Fig. 8B). It is composed of two categories of triactines (giant triactines and triactines larger than those of the cortex) (Fig. 8C). There are also tetractines and some triactines surrounding the canals (Fig. 8D). The atrial skeleton is smooth, composed mainly of triactines, with a few tetractines also present (Fig. 8E).

Ecology
The specimen was collected on a shaded, semi-vertical, hard limestone bottom.

Description
The body has the shape of a vase (0.8 × 0.4 cm), with a single apical osculum surrounded by a membrane and a crown of a few, or even no trichoxeas (Fig. 10A). The osculum is supported by sagittal tetractines, but a few triactines are also present. They are organised in parallel and point their apical actines to the osculum. They become disorganized, smaller, thinner and less sagittal farther from the osculum. They are also substituted by triactines. Numerous diactines on the surface make it very hispid. The aquiferous system is leuconoid and the atrium is large (Fig. 10A). The cortical skeleton is composed of tangential triactines, perpendicular giant diactines, microdiactines and rare trichoxeas (Fig. 10B-E). The giant diactines frequently cross the entire choanosome (Fig. 10B). The choanosomal skeleton has no organisation. It is composed mainly of subregular triactines, with curved paired actines. Tetractines are also present, but only surrounding canals. The atrial skeleton has triactines and a few tetractines that project their apical actines into the atrium (Fig. 10F). Microdiactines are also present in the atrium. (Table 9) oscular TriacTines (very few) anD TeTracTines (abundant). Sagittal. Actines are cylindrical and blunt to sharp. The unpaired actine is thinner than the paired ones. The apical actine of the tetractines is conical, sharp, smooth and strongly curved towards the osculum aperture.

Spicules
Trichoxeas. Very thin, long and straight. They are frequently broken. These spicules are rare, but can be found in the cortex and atrium.
choanosomal TeTracTines. Sagittal. The paired actines are curved, consequently the unpaired angle is smaller than the paired angles. Actines are slightly conical with blunt tips. The apical actine is straight or curved, conical, smooth and sharp ( Fig. 11H-I). These spicules are present only surrounding the canals.

Ecology
Specimens were collected on a cliff in a shaded area.

Remarks
This species differs from all other species of Leucandra mainly by the composition of the skeleton, particularly by the presence of mainly triactines in the atrial skeleton, with very long and slender paired actines and few spiny microdiactines in the cortex. The most similar species is the Californian L. heathi Urban, 1906. However, this species has no tetractines, while L. spinifera sp. nov. has a few tetractines. Besides, microdiactines are not abundant in L. spinifera sp. nov., while in L. heathi they form a continuous palisade in the cortex.

Description
The body has the shape of a vase with a single apical osculum surrounded by a crown of trichoxeas (Fig. 12A). Surface is very hispid. The aquiferous system is leuconoid (Fig. 12B). The cortical skeleton is composed of the basal system of large tangential tetractines and few triactines (Fig. 12C). Giant diactines cross the surface, penetrating deeply into the choanosome. They are present from the osculum to the base of the sponge. Among these giant diactines there are also very thin and long trichoxeas, organised in tufts, and very few microdiactines (Fig. 12D). The choanosomal skeleton is characteristic of Paraleucilla, with an inarticulate region (outer region) and a zone without organisation (inner region) (Fig. 12E). The outer region is formed by the apical actine of the cortical tetractines, the unpaired actine of subatrial tetractines and very few triactines. The paired actines of these subatrial spicules are frequently curved, resembling a hook. The inner region is formed by scattered subatrial tetractines and very few triactines. The atrial skeleton is composed of tetractines only (Fig. 12F). In some parts of the sponge the inarticulate skeleton seems not to exist and it becomes more similar to Leucandrilla. (Table 10) oscular TriacTines. Strongly sagittal. Actines are conical and sharp. The unpaired actine is longer and thinner than the paired ones and basipetally directed.

Spicules
DiacTines. Giant. They are present in the oscular crown and cortex. They are almost fusiform but slightly curved, with a thicker tip outside the sponge (Fig. 13A). The size is very variable. Many diatoms are attached to the diactines surrounding the osculum. Size: 1000.0/25.0-50.0 µm.
Trichoxeas. Present in the oscular crown and cortex. They are thin, straight and most of them are broken. Size: > 330.0/2.5 -5.0 µm. microDiacTines. Very rare, fusiform or arrow-headed. Sometimes one of the tips has small spines while the other one is thicker (Fig. 13B). They are present in the cortex. Size: 95.0/2.5 µm.

Ecology
Specimens were collected on a cliff in a shaded area.

Remarks
Currently there are 11 known species of Paraleucilla, and P. magna Klautau et al., 2004 is the only one that has been recorded in the Mediterranean Sea up to now. Both the external morphology and spicule composition differ in these two species. The most similar species to P. dalmatica sp. nov. are P. perlucida Azevedo &Klautau, 2007, from Brazil, andP. princeps (Row &Hôzawa, 1931), from Australia. Nonetheless, P. dalmatica sp. nov. can be differentiated from P. perlucida mainly by the absence of diactine I and trichoxea in the latter. Paraleucilla princeps also differs by the absence of diactine I and microdiactines. Therefore, P. dalmatica sp. nov. is the second species of Paraleucilla recorded from the Mediterranean Sea.

Description
The body is vase-shaped (1.1 × 0.8 cm), with a single apical osculum surrounded by a crown of trichoxeas (Fig. 14A) and diactines supported by sagittal tetractines. These tetractines are arranged parallel to each other and their unpaired actines are basipetally directed. The unpaired actine is longer and thinner than the paired ones and the apical actine is curved towards the osculum aperture. The paired actines are slightly curved. There is no suboscular region. The aquiferous system is syconoid and the atrium is central. The radial tubes are coalescent (Fig. 14B). Diactines and trichoxeas protrude through the distal cones; consequently, the surface is very hispid. These diactines (ca 10 to 15) penetrate only a little into the sponge surface (Fig. 14C). The unpaired actine of some triactines also protrudes through the cones.
The tubar skeleton is articulated, but not so well organised as in most sycons (Fig. 14D). It is composed of rows of sagittal triactines that point their unpaired actines to the surface. These tubar triactines are larger than those of the distal cones and the paired actines are frequently curved. The subatrial skeleton is composed of sagittal triactines and tetractines (Fig. 14E) with very thin actines. The unpaired actine is much longer than the paired ones and the longest ones are frequently localized among the choanocyte chambers. They point their unpaired actines towards the distal cones. Some of the subatrial triactines are similar to pseudosagittal spicules. The atrial skeleton is composed of two categories of tetractines tangentially organized (Fig. 14E). They frequently have long, unpaired and short, paired actines. One of the paired actines is commonly shorter than the other; however, the three basal actines can have the same size (Fig. 14F). When one of the paired actines is shorter than the other, it frequently penetrates an exhalant canal. The main difference between the two categories of atrial tetractines is in the apical actine. Tetractines with thinner apical actines project these actines mainly into the canals, while thicker and curved apical actines penetrate into the atrium (Fig. 14E). Few anchor-like tetractines are present at the sponge base and project their basal actines into the substrate. (Table 11) DiacTines. Almost fusiform, but the tip outside the sponge is a little thicker (Fig. 15A). Size: 537.8/16.1 µm.

Spicules
Trichoxeas. Very thin, long and straight. They were always broken.
anchor-like TeTracTines. The basal actines are very short and curved, while the apical one is very long. Frequently there are spines on the apical actine, but near the basal ones. They vary from four to seven, but seven spines are more common (Fig. 15B). Size: > 1000.0/25.0 µm. TriacTines of The cones. They are smaller than the tubar triactines. The unpaired actine protrudes through the cones and it is shorter than the paired ones, which are curved. Actines are slightly conical and sharp ( Fig. 15C-D). Size: 112.3/6.9 µm (paired actine); 78.6/7.0 µm (unpaired actine).
subaTrial TriacTines anD TeTracTines. The subatrial spicules are very thin. They are sagittal or, sometimes, similar to pseudosagittal spicules. Actines are slightly conical and sharp. The unpaired actine is longer than the paired ones (Fig. 15H). The apical actine of the tetractines is conical, sharp, smooth, shorter than the basal ones and curved in the direction of the atrium. Size: 97.9/5.4 µm (paired actine); 212.4/6.0 µm (unpaired actine).

Ecology
Specimens were collected on a semi-vertical hard limestone bottom. They were found among Cystoseira macroalgae.

Remarks
Currently there are 12 accepted species of Sycon in the Mediterranean Sea, 10 of which have already been reported for the Adriatic. We compared our specimens to all known species of Sycon and even more carefully to the Mediterranean ones, yet we could not find a perfect match.
The main characteristic discerning Sycon ancora sp. nov. from other species is the shape of the atrial triactines and the presence of anchor-like tetractines at the base. If we exclude these characteristics, this species would be mostly comparable to S. raphanus; however, there are several important differences between them.
Sycon raphanus was originally described from the Adriatic Sea by Schmidt (1862). Unfortunately, his description was not detailed enough. According to him, S. raphanus has a bulb shape and a peduncle. He even considered these characteristics to distinguish S. raphanus from S. ciliatum (Fabricius, 1780), a species from the English Channel which he believed to be present in the Adriatic Sea.
Although we believe the entire genus Sycon is in urgent need of revision, the characteristics we found in our specimens strongly indicate the presence of a new species.

Other calcarean species from the Adriatic Sea
Apart from the species described here, we also recorded and molecularly analyzed specimens of Clathrina blanca (Miklucho-Maclay, 1868), C. clathrus (Schmidt, 1864), C. rubra Sarà, 1958 and Paraleucilla magna Klautau, Monteiro & Borojević, 2004. These species are not redescribed here, since specimens from the Adriatic Sea have already been recorded and described in earlier works (Cvitković et al. 2013;Imešek et al. 2014). In the present study, C. blanca was recorded near Selce (45°09'07.8" N, 14°43'15.0" E ), about 1 m deep and C. rubra was recorded near the Island of Čiovo (43°28'58.5" N, 16°21'25.6" E), about 5m deep on a shaded hard bottom. In August and November 2010 they were quite abundant, always only a few millimeters in size and often found on bryozoans. C. clathrus was found in numerous locations along the coast (e.g., Prapratno Cove, 42°48'36.8" N, 17°40'38.4" E; near the Island of Čiovo, 43°28'58.5" N, 16°21'25.6" E) and the cryptogenic species P. magna was found in large numbers in on aquaculture installations in Grška Cove on the Island of Brač and in the Port of Ploče.

Molecular analysis
The number of sites used for the final alignments (gaps included) was as follows: 513 for ITS Calcinea, 1434 for 28S Calcinea, 734 for ITS Calcaronea and 846 for 28S Calcaronea. Both markers revealed the same tree topology in both analyses (but see Fig. 19), yet the Bayesian analysis rendered much better support values than ML in all cases. However, the Adriatic species nested within the respective genera with high bootstrap (BS) and posterior probability (PP) values, thereby confirming the results of morphological analysis (Figs 16-19).
Once more the presence of diactines did not show any phylogenetic signal (Rossi et al. 2011;Klautau et al. 2013). Furthermore, we found former guanchas with only triactines reunited in a monophyletic clade in the ITS analysis, with high support values inside the Clathrina group (0.99 PP and 0.84 BS; Fig. 16). In the 28S calcinean tree (Fig. 17) we recovered a clade where Levinella represents a sister group to Ascandra with high support values (1.00 PP and 0.99 BS), which confirms the results of Voigt et al. (2012). We also recovered a clade comprising the genera Murrayona and Ascaltis in both analyses; however, the support values were less good (0.71 PP and 0.54 BS). The molecular analyses also confirmed the presence of P. magna in the Adriatic Sea (Figs 18-19). Besides, we recovered a calcaronean clade with high support (1.00 PP and 0.99 BS in ITS analysis; 0.95 PP and 0.64 BS in 28S analysis) formed only by Paraleucilla species. The genus Paraleucilla formed a highly supported clade with Leucandra nicolae, while Leucandra spinifera sp. nov. is a sister species of L. aspera (Fig. 19). Sycon ancora sp. nov. represents a sister species of S. raphanus (Fig. 19). We confirmed the paraphyly of the genera Sycon and Leucandra (Voigt et al. 2012).

Species richness
Considering previous data, together with our present results based on morphological and molecular analyses, we found a total of 13 species of Calcinea (Table 12) and 26 of Calcaronea in the Adriatic Sea (Table 13). Taking into account the species richness by sectors (Fig. 20), the richest sector is the Central Adriatic, where 34 species were found, followed by the Northern Adriatic with 18, and the Southern Adriatic with only 5 species. Most of the species present in the Adriatic Sea are also present in other Mediterranean areas, yet, altogether we recorded six species provisionally endemic for the Adriatic, two calcinean and four calcaronean.

Discussion
Since some of the first studies on the class Calcarea were mainly done along the Dalmatian coast by Schmidt and Haeckel in the 19 th century (e.g., Schmidt 1862Schmidt , 1864Haeckel 1870Haeckel , 1872, the knowledge of the current species diversity and distribution certainly awakes taxonomic interest. Analysing previous results with ours, we found a total of 39 species of calcarean sponges in the Adriatic Sea (Tables 12-13).
In this species list we do not consider the records of Clathrina coriacea (Montagu, 1814), Sycon ciliatum (Fabricius, 1780) or S. proboscideum (Haeckel, 1870). The occurrence of C. coriacea was not considered because, analysing the descriptions of this species for the Adriatic Sea, we think that they most probably represent C. conifera or C. primordialis. Sycon ciliatum seems to be restricted to the North Atlantic and was probably mistaken for S. raphanus (Haeckel, 1872). Sycon proboscideum is a species from the Red Sea and its occurrence in the Adriatic Sea was mentioned only by Breitfuss (1935), which suggested that some specimens previously identified as S. raphanus could in fact be S. proboscideum. He did not give a description or any further clues. Therefore, the occurrence of this species in the Adriatic Sea has to be verified (Burton 1963;Longo & Pronzato 2011). Adriatic specimens are written in bold. Adriatic specimens obtained in this study are marked with an asterisk. The detached tree shows the only difference in the topology of the ML and Bayesian analyses.   Our results indicate Sycon to be the most diverse genus, with nine species, followed by Clathrina with six species. However, it is important to consider that Sycon is not a monophyletic genus. It is very difficult to identify Sycon species unequivocally, as most of them have a similar spicule composition -diactines, trichoxeas and triactines in the distal cones, tubar triactines, subatrial triactines and tetractines, and atrial tetractines. To date there have been no studies on the intraspecific morphological variability of Sycon. In addition, most species were poorly described and insufficiently analyzed on the molecular level, which also applies to the calcaronean genera Paraleucilla, Leucandrilla, Leucandra and Leucilla. Molecular phylogenetic studies including as many species as possible would be very desirable to evaluate the limits between these genera. Hence, new calcaronean species are welcome to facilitate more thorough revision of their systematics and to link the molecular traits to the phylogenetically important morphological traits.
We have also confirmed the presence of a few species known so far only from the Atlantic. It was unexpected to find Clathrina conifera in the Adriatic Sea, as this species was first described along the Brazilian coast and was considered endemic (Klautau et al. 1994). Our finding raises the question whether this species was ever truly endemic for Brazil. Since Adriatic calcarean sponges are vastly unexplored and C. conifera is part of the C. primordialis species complex, it is possible that it has been recorded previously as C. primordialis (or C. coriacea). In 2010, a specimen of Clathrina conifera was observed for the first time in the Southern Adriatic, near the Island of Lokrum, and a year later, more than 20 specimens were recorded near the city of Dubrovnik. As both locations are close to the area in Dubrovnik frequently visited by cruise ships, it is possible that this species has been introduced into the Adriatic. However, if this species arrived by anthropogenic means, we cannot state whether it arrived from the Western Atlantic to the Adriatic or vice-versa. It is important to mention that Paraleucilla magna is also present in the Southern Adriatic (Cvitković et al. 2013). It was first recorded in Brazil in the 1980's; however, the origin of this species is unknown. It seems to have been introduced by anthropogenic means into the Mediterranean (Longo et al. 2007) and to have spread into the Eastern Mediterranean, including the Adriatic Sea. Here, we molecularly confirm the presence of P. magna near the Port of Ploče and at a new location, near the Island of Brač (Table 13).
The molecular analyses revealed some interesting taxonomic traits. At the generic level, the monophyletic clade of former guanchas indicates that the development of a peduncle and of parasagittal spicules probably appeared only once in the evolution of Clathrina. Clathrina hispanica was nested within this group, although in the original description of this species neither peduncle nor parasagittal spicules were mentioned (Klautau & Valentine 2003). The type specimen of this species is fragmented, resulting in the impossibility of confirming if a peduncle was present or not; however, we re-analysed the slides of the holotype and found some parasagittal spicules. Another interesting result indicated the close relationship among Ascandra, Soleneiscus and Levinella revealed in the 28S analysis. Voigt et al. (2012) showed that the genus Ascandra is closely related to Soleneiscus and Levinella, which is now confirmed by our results. This implies that in the future the genera Levinella and Soleneiscus might be synonymised with the genus Ascandra; nonetheless, more detailed molecular and morphological analyses on a larger number of specimens and species are needed to confirm this action. At the species level, the molecularly confirmed presence of Clathrina conifera in the Adriatic raises a doubt of the earlier identification of the C. primordialis syntype, allowing the selection of a true lectotype of this species. Additionally, the re-description of Borojevia cerebrum, based on a molecular analysis of specimens discovered near its type locality (Lesina -Island of Hvar), confirmed the presence of this species in the Mediterranean Sea (Table 1; Fig. 16). Observing the morphological variations within a single, molecularly verified species, enabled the synonymization of two "cerebrum" varieties.
Step by step, the "cerebrum complex" is being solved. All this again confirms that molecular verification of morphological traits is very important for a proper species assignment. It goes hand in hand with morphological confirmation relying on the type specimens, which often become deteriorated or even lost, without detailed descriptions. This review of some of the first species of calcarean sponges, that were last recorded and described by Haeckel in the 19 th century, allowed validation of their taxonomic status at the molecular and morphological levels. It bears a significant weight in reviving museum collections, which would be of a great help for systematics research of calcarean sponges in the future.