Molecular study supports the position of the New Zealand endemic genus Lamellomorpha in the family Vulcanellidae ( Porifera , Demospongiae , Tetractinellida ) , with the description of three new species

Due to the possession of huge contort strongyles, and a lack of triaenes in an otherwise ‘astrophorine’ spicule complement, the phylogenetic position of the endemic, monospecific New Zealand sponge genus, Lamellomorpha Bergquist, 1968, has remained enigmatic. The genus was established within Jaspidae de Laubenfels, 1968 (in the abandoned order Epipolasida Sollas, 1888), but it was not until 2002 that the genus was transferred formally to Astrophorina Sollas, 1887, albeit incertae sedis, by Hooper & Maldonado (2002). In this study, we recognise specimens of Lamellomorpha from the Subantarctic New Zealand region and Chatham Rise, considered by Bergquist to be conspecific with the type species, L. strongylata Bergquist, 1968, first described from the Three Kings-Spirits Bay European Journal of Taxonomy 506: 1–25 (2019) 2 region of Northland, as the new species, L. australis Kelly & Cárdenas sp. nov. These two species of Lamellomorpha have differences in external morphology and colour, skeletal architecture and spicules, natural products, geographical distribution, and depth ranges. Sequencing of the COI Folmer barcode/ mini-barcode and of 28S (C1–C2 domains) of these two species suggests phylogenetic affinities of Lamellomorpha with the tetractinellid suborder Astrophorina and the family Vulcanellidae Cárdenas et al., 2011. Two Subantarctic New Zealand species of the vulcanellid genus Poecillastra Sollas, 1888, P. ducitriaena Kelly & Cárdenas sp. nov. and P. macquariensis Kelly & Cárdenas sp. nov., provide further support for the close relationship of Lamellomorpha and Poecillastra.

Neither Cárdenas et al. (2011) nor Cárdenas & Rapp (2012) sampled L. strongylata for molecular sequences. Here, we sequence, for the fi rst time, the COI Folmer barcode/mini-barcode and 28S (C1-C2 domains) of L. strongylata and the new species L. australis Kelly & Cárdenas sp. nov. We also describe a second, less common Subantarctic New Zealand species, initially identifi ed as a third species of Lamellomorpha with rare, calthrop-like triaenes, but now considered to be a species of Poecillastra Sollas, 1888, in the family Vulcanellidae: Poecillastra ducitriaena Kelly & Cárdenas sp. nov. The systematic and phylogenetic implications of these species are considered with respect to the broader phylogenetic position of Lamellomorpha, Poecillastra, and the wider Tetractinellida.

Collections and morphological systematics
Specimens were collected by rock and cone dredges as well as by beam and Agassiz medium trawls, from several research vessels between 1962 and 2010. The majority of specimens were collected onboard the National Institute of Water & Atmospheric Research (NIWA) research vessels RV Tangaroa and RV Kaharoa; numerical voyage identifi er and associated stations cited as NIWA Stn TAN(voyage number)/(station number) and NIWA Station KAH(voyage number)/(station number), respectively. Several specimens from the Three Kings Islands were collected by dredge from RV Kaharoa in 1999, on a voyage chartered by the CRRF.
Specimens were frozen immediately upon collection and then preserved in 70% ethanol or preserved immediately into 70% ethanol (CRRF). Histological sections of the sponges were prepared by embedding a small piece of the sponge in paraffi n wax and then sectioning with a microtome at 70 μm. Spicule slides and SEM spicule preparations were made following the methods of Kelly & Sim-Smith (2012). Clean spicules for SEM examination were spread on a plastic disc, air-dried, and coated with platinum for 600 s. Spicules were viewed on a Philips XL30S FEG SEM. Spicule dimensions were measured using a Meji MT5300L compound microscope fi tted with a Leica DFC420 microscope camera that was connected to Leica Application Suite imaging software (Leica Microsystems (Switzerland) Ltd.). Spicule measurements in the species descriptions are given as the mean length (range) × mean width (range) of twenty spicule measurements per specimen unless stated otherwise and are based on measurements from the holotype or paratypes and confi rmed through examination of all other specimens. Collection information on specimens examined in this study and previous records were gathered in a table made available in the PANGAEA data repository (https://doi.pangaea.de/10.1594/PANGAEA.895370).
Primary and secondary type material of the new species, and additional material, are accessed within the NIWA Invertebrate Collection (NIC) at NIWA, Greta Point, Wellington, using the prefi x 'NIWA-'. Pieces of the holotypes of L. strongylata, L. australis sp. nov., and Poecillastra ducitriaena sp. nov. are also stored at the Zoological Museum in Uppsala, Sweden (prefi x 'UPSZTY-').
Additional abbreviations used in the text include CRRF (Coral Reef Research Foundation, Palau). The taxonomic authority for the new taxa described in this paper is restricted to the authors Michelle Kelly and Paco Cárdenas.

Molecular systematics
DNA was extracted using a DNeasy Blood and Tissue kit (Qiagen). PCRs were carried out in 25 μl solutions using PuReTaq Ready-To-Go PCR beads (GE Healthcare). Due to poor preservation of the specimens for molecular work (storage in 70% ethanol, instead of the recommended 96% ethanol), the DNA quantities were very low, and it was very degraded (observation on a gel of 1 μl of DNA extract). The complete Folmer fragment could not be sequenced using the standard animal barcoding primer pair LCO1490/HCO2198 (Folmer et al. 1994) so it was sequenced in two parts: the universal minibarcode (130 bp, without primers) was obtained using the primer pair LCO1490/Tetract-minibarR1 (Cárdenas & Moore 2019). Then the second part of the Folmer fragment (539 bp, without primers) was amplifi ed using the primer pair VulcanF2/HCO2198. VulcanCOI-F2 (5'-GGGGATGACCAACTTTATAATG-3') is a new specifi c primer made to amplify COI in Vulcanellidae species. PCR conditions were (5 min/94 °C; 37 cycles (15 s/94 °C, 15 s/46 °C, 15 s/72 °C); 7 min/72 °C). The 28S fragment (C1-C2) of 308-369 bp, was obtained using the primer pair C1'/Ep3 (Chombard et al. 1998) and the same PCR program as for COI except that we used 50°C for the annealing temperature. We pruned the comprehensive Tetractinellida COI alignment from Kelly & Cárdenas (2016), to keep only species of Astrophorina. We added additional sequences of Astrophorina from the Galapagos (Schuster et al. 2018) along with the new sequences. The COI data matrix included 115 sequences (with eight Spirophorina Bergquist & Hogg, 1969 outgroups). For 28S, we built an alignment based upon the Astrophorina 28S (C1-D2) alignment from Cárdenas et al. (2011) and added Astrophorina 28S (C1-D2) sequences (Thacker et al. 2013;Schuster et al. 2015Schuster et al. , 2018. The 28S data matrix included 126 sequences (with seven Spirophorina outgroups) and was automatically aligned using MAFFT v.7 (Katoh & Standley 2013), L-INS-i option, implemented in AliView 1.18 (Larsson 2014). Phylogenetic analyses were conducted on the CIPRES science gateway v. 3.3 (http://www.phylo.org) (Miller et al. 2010): RAxML 8.2.10 (Stamatakis 2014) for maximum likelihood (ML) and MrBayes v. 3.2.6 (Ronquist et al. 2012) for Bayesian analyses. Bayesian analyses were run with BEAGLE, and consisted of two runs of four chains, each for 5 000 000 generations and sampled every 1000 tree after a 25% burn-in.
Lamellomorpha Bergquist, 1968Lamellomorpha Bergquist, 1968 Diagnosis (modifi ed from Hooper & Maldonado 2002) Massive, lamellate stalked-palmate, or paddle-shaped sponges, with a relatively smooth, granular, or fl eshy, slightly conulose surface. Ectosomal skeleton a skin-like membrane packed with microstrongyles. Choanosomal skeletal architecture a core of megascleres, which are straight, curved, sinuous, or contort oxeas, frequently modifi ed with one or both ends rounded as in strongyloxeas. These radiate through the stalk and fan. Straight oxeas arise as short subectosomal tracts that emerge oblique to the surface. Roughened microstrongyles or microxeas and streptasters (amphiasters, metasters, and spirasters) scattered throughout the body.

Description
The holotype was described by Bergquist (1968) as a "massive, thick, sometimes folded and incurved lamellate sponge", 130 mm high, 102 mm wide, and 18-22 mm thick, supported by a stout stalk 30 mm in diameter. The surface was described as smooth where the dermal membrane was intact, otherwise ragged due to projecting clumps of oxeas and strongyles. Oscules, 1-2.6 mm in diameter, were found on the convex surface of the lamella and lie fl ush with the surface (Bergquist 1968). Examination of the numerous preserved specimens in NIC reveal occasional membranous oscules, but it is diffi cult to tell whether they are restricted to one side of the sponge. However, in the holotype, pores were observed on the opposite side to the oscules, in cribriporal areas, separated by small ridges, or with no boundaries, making a continuous pore surface; each pore is 40-80 μm in diameter. The texture was described as, "fi rm but compressible, crisp, easily broken". The colour in life was described as "bright green" and the colour in spirit, "blue green or yellowish green" (Bergquist 1968). The most recent collection was by the Coral Reef Research Foundation in 1999 (NIWA 93474 leg.; Fig. 2A), who described a "dark, royal blue (not navy blue), (palmate) fan sponge with pointed tips, 20 cm high and about 1 cm thick, that tears easily, and which has a fl eshy surface".

Skeleton
The description by Bergquist (1968) of the choanosome as "lax and confused with slight traces of radiate construction discernible", is accurate, but in NIWA 93474 leg. the contort strongyles strongly radiate through the plane of the fan. Bergquist described a "subectosomal region", in which there were tracts of megascleres, variable in thickness, that curved outward and intersect with the surface at an acute angle; in NIWA 93474 leg. these are predominantly oxeas (Fig. 6A). The ectosome is densely packed with microstrongyles and streptasters, which also occur throughout the sponge, but in much less abundance.  Bergquist (1968) considered the megascleres of L. strongylata (Bergquist 1968: 31, 32 (table of spicule dimensions)) to be "strongyles, oxeas and strongyloxeas", all of similar range in length and width, varying only in relative frequency in the two specimens (presumably the Three Kings holotype and the NZOI Station B176 specimen from Campbell Plateau), with oxeas being dominant in the latter. Re-examination of the holotype megascleres, and those of more recent collections, indicate that there are probably two forms of megascleres: 1) straight to slightly curved oxeas that are common in the subectosomal tracts, ranging from about 1500-1750 μm long and up to 25 μm thick; and 2) massive sinuous or contort oxeas that are usually very thick and frequently modifi ed with one or both ends rounded as in strongyloxeas, rarely as in true strongyles, ranging from about 1600-2375 μm long and up to 40 μm thick. However, it is diffi cult to distinguish the various megascleres in some specimens, and in some the spicules are much less contort. Table 2; Fig. 2C-F) Microstrongyles are "squat, evenly rounded spicules, slightly roughened and occasionally centrotylote" (Bergquist 1968: 31, 32 (table of spicule dimensions)) and range from about 21-34 μm long (Table 2). Bergquist described the streptaster microscleres of L. strongylata (Bergquist 1968: 31, 32 (table of  spicule dimensions)) as "plesiasters, small spicules with 3-12 smooth, sharply pointed rays". A reexamination of the holotype (Fig. 2E-F) using scanning electron microscopy has revealed that the streptaster microscleres are metasters and occasionally amphiasters with relatively long microspined rays, all in one size category, following the defi nition of Sollas (1888), and as used in Cárdenas & Rapp (2012). We describe these spicules as metaster-to amphiaster-like streptasters with heavily spined, relatively long rays in one size category, ranging in length from about 7-15 μm long (Table 2).

Distribution
Northeast of New Zealand.

Substrate, depth range and ecology
Attached to rocky reefs and sediment and rubble-covered rocky platforms, depth 41-200 m.

Remarks
Lamellomorpha strongylata was originally described in considerable detail by Bergquist (1968), and the holotype was redescribed without re-examination more recently by Hooper & Maldonado (2002).
No further material was examined. Here, for the fi rst time, we illustrate the sponge as it appears upon collection, showing the beautiful royal blue colouration ( Fig. 2A), and illustrate the detail and ornamentation of the microscleres (Fig. 2C-F) using scanning electron microscopy. There is little to add to the original description, consequently the description and skeletal details are provided in comparative prose. Lamellomorpha strongylata is restricted to the northernmost tip of New Zealand and beyond to the Three Kings Rise, and is easily recognised in the fi eld by the palmate, tree-like shape and the deep blue to green colouration.

Description
Uni-to bilamellate fan sponge, table tennis bat-shaped (Fig. 3A) or club-shaped (Fig. 3B), 130-200 mm high with a short, broad stalk, 2-3 cm thick, and a relatively thick lamella (up to 2 cm thick in places), attenuating towards the margins, which are frequently incised. The specimen from NZOI Station B176 was described by Bergquist (1968) as being 160 mm high, 89 mm wide, and 32-58 mm thick, supported by a stalk that was broken and thus was not measured. Oscules were not visible in the holotype or any other specimen. Pores are inconspicuous and compressed (probably due to the fi xation) and were 40-80 μm in diameter (measured on the holotype). Surface relatively smooth with low ridges radiating from the stalk to the fan margins. Texture, relatively soft, compressible. Colour in life and preservative, tan.

Skeleton
Choanosome disorganised, with megascleres orientated more or less parallel with the axis of the fan and stalk (Fig. 6B-D), with single or a couple of megascleres extending beyond the surface from the subectosome. The ectosome is extremely thick and packed with microstrongyles and streptasters, which also occur in great density throughout the sponge. (Table 1; Fig. 3C) Bergquist (1968) considered the megascleres of the "subantarctic specimen" (presumably the NZOI Station B176 specimen from Campbell Plateau) to have, "predominantly oxeas, some of which are curved, but most are contort". Our examination of new material reveals that oxeas dominate the megasclere complement; these are rarely to never modifi ed; all have sharp attenuated tips. The majority are straight to slightly curved and contort, but not to the degree seen in L. strongylata. The megascleres reach their greatest length in L. australis sp. nov., up to 3575 μm long in the specimen NIWA 93486 leg. (paratype). Fig. 3D-F) Bergquist did not differentiate between the microscleres of the holotype of L. strongylata (from the Three Kings) and the subantarctic Campbell Plateau specimens (L. australis sp. nov.), calling them all "plesiasters" in the table of spicule dimensions. However, in pl. 11, fi gs E2 and F2-3, a clear difference is obvious between the illustrations of the streptasters: they are metasters in pl. 11, fi g. E2 (L. strongylata) and larger metasters (pl. 11, fi g. F3) and "abnormal microrhabds" in pl. 11, fi g. F2 (L. australis sp. nov.). The "abnormal microrhabds" of Bergquist (1968: pl. 11, fi g. F2) are most likely spirasters (as in our Fig. 3F), the ornamentation of which would not have been visible under light microscopy available at the time.

Substrate, depth range and ecology
Attached by a thick stalk to sediment covered rocky substrate, depth 71-549 m.

Remarks
Bergquist (1968) listed two specimens from Campbell Plateau in the Subantarctic region of New Zealand, from NZOI Station B176 (46 fathoms = 84.12 m) and NZOI Station B184 (103 fathoms = 188.4 m). Unfortunately, neither specimen was found in the NIWA collections and both are presumed lost. However, we did fi nd a specimen from NZOI Station B184 (NIWA 93499 leg.) from a depth of 344 m. Bergquist (1968) considered the two specimens she examined to be conspecifi c with L. strongylata, despite the obvious disjunct distribution, but noted that the subantarctic specimens had predominantly oxeas, an observation we agree with.
Examination of numerous specimens uncovered in NIC allows us to convincingly separate L. australis sp. nov. from the type species on geographic distribution, morphology, and skeletal details, despite the COI minibarcodes not differentiating them (Fig. 7). The most obvious difference that separates L. australis sp. nov. from L. strongylata is the markedly disjunct distribution and depth ranges: L. strongylata has only been recorded to the north of New Zealand, 41-200 m depth, while L. australis sp. nov. is only found on and south of Mernoo Bank on the Chatham Rise, ranging in depth from 71 m on Mernoo Bank, to 549 m in the Solander Trough. In terms of morphology and colouration in life, L. strongylata forms a relatively soft, dark royal blue, palmate sponge, supported by a relatively narrow stalk, while L. australis sp. nov. forms a distinctive, tan, paddle-shaped sponge, with thin, incised margins, on a thick, short stalk. In terms of skeletal architecture, the choanosome of L. australis sp. nov. is much more densely packed with microscleres than L. strongylata, and the former species lacks the relatively distinct subectosomal tracts of the latter. As noted by Bergquist (1968), the megascleres of L. australis sp. nov. differ from those of L. strongylata in being predominantly oxeas (straight and contort) with no modifi cations of the tips to strongyloxeas as in L. strongylata. In addition, we note that the contort forms are much longer on average, and have a greater size range, than in L. strongylata. Finally, microscleres also discriminate L. australis sp. nov. from L. strongylata. Lamellomorpha strongylata has stubby, often centrotylote roughened microstrongyle, while L. australis sp. nov. has a relatively fi ne, curved, slightly longer roughened microxea. Lamellomorpha strongylata has metaster-to amphiaster-like streptasters with heavily spined, relatively long rays in one size category, while L. australis sp. nov. has metaster-like streptasters with heavily spined, relatively long rays and spirasters with abundant, short, microspined rays that emanate from a long, spiral axis. Spirasters are absent in L. strongylata.
As part of their ongoing investigations into New Zealand marine natural products in sponges, professors Murray Munro and John Blunt and their group in the Department of Chemistry, University of Canterbury, Christchurch, collected what was identifi ed by the late professor Patricia Bergquist as L. strongylata, from Mernoo Bank on the Chatham Rise. Vouchers of these sponge specimens were donated to NIC for their preservation and future study and have been re-identifi ed here as L. australis sp. nov., extending the known distribution of L. australis sp. nov. north to the Chatham Rise. Thus, it is L. australis sp. nov., and not L. strongylata, from which biologically active secondary metabolites were isolated by the University of Canterbury group, including calyculins (A, B, E, and F), calyculinamides (A and B), swinholide H (Dumdei et al. 1997), and theonellapeptolides (Li et al. 1998;Hickford 2007); identical and related compounds are found in sponges in the genus Theonella Gray, 1868 and Discodermia du Bocage, 1869 (family Theonellidae).
It has been shown that calyculins and its derivatives (e.g., calyculinamides) could be produced by the fi lamentous bacteria 'Entotheonella' spp. in Discodermia . 'Entotheonella' spp. are especially abundant in Theonella swinhoei, as well as in many other demosponges (Wilson et al. 2014). In her PhD thesis, Hickford (2007) noticed that fi lamentous heterotrophic (Gram positive) bacteria were very abundant in L. australis sp. nov. and were associated with several theonellapeptolides. The producer of theonellapeptolides is currently unknown but the results of Hickford (2007) suggest that L. australis sp. nov. may be a host for theonellapeptolides-producing 'Entotheonella'-like bacteria. Hickford (2007) also isolated unicellular bacteria from the same specimens and showed these were associated with swinholide H. This result concurs with previous results from Bewley et al. (1996), who identifi ed swinholide A in unicellular bacteria isolates from Theonella swinhoei from Palau. However, it is an apparent contradiction with Ueoka et al. (2015) who convincingly show that misakinolide A (swinholide-like compound) from another Theonella swinhoei chemotype (chemotype WA from Japan) is produced by 'Entotheonella serta'. Therefore it seems that swinholide-type compounds may be produced by bacteria other than 'Entotheonella' in L. australis sp. nov. and Theonella swinhoei (chemotype Palau). Hickford (2007) further states that specimens from the "northern population of L. strongylata" (L. strongylata) not only had very low quantities of fi lamentous bacteria (apparently limited to the surface of the sponge), but also missed the biological activity, and therefore may not produce the above mentioned compounds. Thus, to conclude, L. strongylata and L. australis sp. nov. also clearly differ in terms of natural products and microbial communities.

Etymology
Named for the possession of triaenes in addition to the apparent spiculation of Lamellomorpha, and their guide to the phylogenetic origins of this species ('duci-' in the sense of a guide).

Type locality
Subantarctic region of New Zealand, Snares Island Platform, depth 125-213 m.

Description
Multilamellate, foliose, fan sponge (Fig. 4A), 160 mm high, 104 mm wide, with a short thick stalk about 2 cm thick. Lamella up to 2 cm thick in places, attenuating to curled margins. Oscules were not visible on the holotype. Cribriporal pore areas are widespread between parallel tangential tracts of oxeas; individual pores 80-160 μm in diameter, on both sides of the lamella. Texture fi rm, compressible, fl exible, granular and smooth to the touch. Colour in preservative tan.

Distribution
East of Snares Island Platform.

Substrate, depth range and ecology
Attached by a thick stalk to sediment covered rocky substrate, depth 125−213 m.

Remarks
This remarkable sponge was fi rst identifi ed as a third species of Lamellomorpha, as it appeared to have an almost identical form (stalked, multilamellar fan), a megasclere complement of straight and contort oxeas (more or less restricted to the stalk), small centrotylote microxeas, and metasters (albeit rare). Because the short-shafted triaenes were relatively uncommon, it was initially hypothesised that this was a species of Lamellomorpha with rudimentary triaenes that 'showed the way' to the true affi nity of the genus with other triaene-bearing Tetractinellida. However, molecular sequencing consistently linked Poecillastra ducitriaena sp. nov. with other Poecillastra species (Fig. 7). Despite its consistency with two independent markers, we note that this grouping is not supported (bootstrap of 60 for COI, of 10 with 28S). This may be due to the absence of other subantarctic Poecillastra species in our sampling which Poecillastra ducitriaena sp. nov. may be closer to (P. Cárdenas, unpublished results).
Although not fully documented (Kelly et al. 2009), our knowledge of Poecillastra in the New Zealand region is reasonable and includes what we consider to be Poecillastra laminaris (Sollas, 1886) (Zeng et al. 2016 and Poecillastra schulzei (Sollas, 1886). While several undescribed species are known from the New Zealand EEZ, no specimens are known that contain the characteristic contort oxeas of Poecillastra ducitriaena sp. nov.

Description
Solid stalk of sponge of unknown morphology, 15 mm in diameter, 20 mm high, expanding on the broken, upper surface, sides of stalk sculpted, attachment base contains patches of substrate (Fig. 5A). Surface hispid and scratchy to the touch; texture fi rm, incompressible. Colour in preservative tan.

Skeleton
Stalk composed of huge swathes of contort oxeas and triaenes between which are abundant microscleres.

Distribution
Macquarie Ridge.

Substrate, depth range and ecology
Attached to rock substrate; depth 462-524 m.

Remarks
The specimen is the attachment base of a sponge of unknown morphology, but it clearly differs from the holotype of Poecillastra ducitriaena sp. nov. in having a very hispid, crisp, scratchy surface, indicating a reduction of the ectosomal crust of microscleres, and the abundance of large megascleres. It is similar to Poecillastra ducitriaena sp. nov. in the possession of abundant contort oxeas in the stalk, but differs in the lack of straight oxeas in the stalk and the much larger dimensions of all the spicules: the contort oxeas are up to 2000 μm longer, on average, in Poecillastra macquariensis sp. nov., and the triaenes are about double the size of those in Poecillastra ducitriaena sp. nov., and much more abundant, the microxeas are about ten times larger, and the sponge contains plesiasters, absent in Poecillastra ducitriaena sp. nov.
Because our knowledge of Poecillastra in the New Zealand region is reasonable (see above), we have made the decision to record and name this second Poecillastra species, despite our lack of information on the body shape, and because surface texture, spicule types and dimensions are so different from those of Poecillastra ducitriaena sp. nov.

Discussion
We were not successful in obtaining sequences from the holotype of L. strongylata (NIWA 356 leg.) but we obtained one Folmer COI sequence from non-type specimen NIWA 51172 leg. and one COI mini-barcode (130 bp) from non-type specimen NIWA 51267 leg., both from the same type locality, the Three Kings Islands. The 28S fragment (C1-C2 domains) was also obtained for the latter two specimens (369 bp each). We obtained one COI mini-barcode from the holotype of L. australis sp. nov. (NIWA 89736) but failed to amplify 28S for L. australis sp. nov. We obtained the Folmer (661 bp) and the 28S fragments (308 bp) from the holotype of Poecillastra ducitriaena sp. nov. (NIWA 61944 leg.).
In the COI tree (Fig. 7), the two species of Lamellomorpha, which have identical sequences (at least identical minibarcodes) were a sister group to Vulcanella Sollas, 1886; this grouping with Vulcanella was poorly supported (bootstrap of 61). Lamellomorpha + Vulcanella were sister to a poorly supported Poecillastra clade (bootstrap of 60), which included Poecillastra ducitriaena sp. nov. Despite these poorly supported nodes, the Vulcanellidae was a very well supported clade suggesting that Lamellomorpha is clearly part of this family. The 28S tree confi rms the position of Lamellomorpha in the Vulcanellidae but with no true support. Vulcanella appeared paraphyletic (poorly supported), while the grouping of Lamellomorpha was uncertain with respect to Vulcanella or Poecillastra. This poor resolution may be due to the short sequences we obtained (308-369 bp), which are in fairly conserved domains of 28S and so with few bp differences to differentiate species. The grouping of Lamellomorpha with the Vulcanellidae suggested by the molecular markers is in accordance and supported by the morphological revision of this study. Indeed, the lamellate external morphology, the oscule/pore morphologies and distributions, as well as the skeleton organization in Poecillastra and Lamellomorpha are similar. Some spicules are also clearly homologous, such as the metasters to spirasters. The key discovery in this work is the confi rmation that Lamellomorpha has integrity as a genus within the family Vulcanellidae, separate from Vulcanella and Poecillastra. This result generates new hypotheses about the evolution of spicules within the Vulcanellidae. The microrhabds in Lamellomorpha which earlier confused taxonomists, can now be considered as reduced Poecillastra/Vulcanella microxeas. Secondary losses of spicules have already been shown to be quite common in the Astrophorina (Cárdenas et al. 2011). Here, the short-shafted triaenes, which can be scarce (e.g., Poecillastra compressa) to quite rare (e.g., Poecillastra ducitriaena sp. nov.) in Poecillastra, would have been completely lost in Lamellomorpha. As for plesiasters (the largest streptaster category), they are absent in Lamellomorpha, as in some Poecillastra (e.g., Poecillastra ducitriaena sp. nov.) or Vulcanella (e.g., Vulcanella horrida (Schmidt, 1870)).