Lissodendoryx (Ectyodoryx) Lundbeck, 1909 (Coelosphaeridae, Poecilosclerida, Demospongiae) from Southern Chile: new species and a discussion of morphologic characters in the subgenus Author Fernandez, Julio C. C. Author Cárdenas, César A. Author Bravo, Alejandro Author Lôbo-Hajdu, Gisele Author Willenz, Philippe Author Hajdu, Eduardo text Zootaxa 2016 4092 1 journal volume 10.11646/zootaxa.4092.1.4 f448a9b2-e7ab-4dbf-87a1-64d0338b8b95 1175-5326 266115 51F44763-E817-4E58-A4AC-525E63B6D27B Genus Lissodendoryx Topsent, 1892 Massive, lobate, or flabelliform sponges, with irregular or clathrate surface. Ectosomal skeleton with smooth ectosomal tornotes, tylotes or strongyles forming tangential tracts and surface brushes; choanosomal skeleton an isodictyal reticulation of single spicules, an anisotropic reticulation, or a plumo-reticulation, composed of smooth or acanthose choanosomal styles, sometimes oxeas or strongyles, alone or in combination ; echinating spicules may be present; microscleres are arcuate isochelae, sigmas, and raphides in trichodragmata. Cosmopolitan distribution, many species (emended from van Soest, 2002a). In the molecular phylogeny proposed by Redmond et al. (2013), Lissodendoryx is a polyphyletic taxon. More specifically, L. ( Lissodendoryx ) forms a relatively well established clade (Bootstrap = 89). Then, one only finds 80% Bootstrap again for a large clade including Chondropsis (Chondropsidae) , Desmapsamma (Desmacididae) , Forcepia (Coelosphaeridae) , Myxilla (Myxillidae) and Tedania (Tedaniidae) . Lissodendoryx ( Ectyodoryx ) , with only two species evaluated in their analysis did not cluster together. But, as Bootstrap support for the inner branches of the larger clade including both species is too weak (below or barely above 50%), nothing can be said so far about the status of the subgenus, and of its most likely phylogenetic relations to other Poecilosclerida . Species of Lissodendoryx ( Acanthodoryx ) clustered together with 100% Bootstrap support, but the clade might contain only a single species, if the included L. ( A. ) sp. ever proves to belong in L. ( A. ) fibrosa (100% 18S rDNA similarity). Again, Bootstrap support at more inclusive clades is less than 50%, so affinities of the subgenus must remain obscure. A combined analysis of 18S and 28S rDNA fragments permitted Morrow et al. (2013) to recover with 100% Bootstrap support the following relationships for a polyphyletic Lissodendoryx : ( Lissodendoryx , Forcepia , Myxilla , Tedania, Trachytedania ) and ( Lissodendoryx , Crella , Higginsia , Hymedesmia , Plocamionida ). The first of these clades clusters Lissodendoryx with an additional taxon in the Coelosphaeridae ( Forcepia ) , and the poeciloclerid families Myxillidae and Tedaniidae , thus including taxa bearing arcuate chelae ( Lissodendoryx and Forcepia ) and taxa bearing anchorate chelae ( Myxilla ). The second clade with 100% support, aside from the interordinal relationship to the Axinellida ( Higginsia ), clusters Lissodendoryx together with Crellidae and Hymedesmiidae , and only includes taxa bearing arcuate chelae. But, if a more inclusive clade with 98% support is considered instead, than Myxilla , Phorbas (as Stylostichon ) and Spanioplon are included too, anchorate chelae come into play again. Similar relations had already been envisaged based on morphological analysis alone by van Soest (1984) and Desqueyroux-Faúndez & van Soest (1996). In part, our taxonomic results suggest a closer relationship between Lissodendoryx and Hymedesmiidae . Lissodendoryx ( E. ) ballena sp. nov. has a plumo-reticulated skeleton and ectosomal tornotes, rather than a reticulate skeleton and ectosomal tylotes as observed in other species of Lissodendoryx , such as L. ( E. ) corrugata sp. nov. and L . ( E .) coloanensis sp. nov. described here, and a few additional species reported by van Soest (2002a). In addition, the morphology of the two largest categories of arcuate isochelae in L. ( E. ) ballena sp. nov. is quite similar to the morphology of the isochelae in some species of Hymedesmiidae ; cf. Phorbas areolatus (Thiele, 1905) in Fernandez et al . ( in prep. ). It is also important to mention that a plumo-reticulate architecture is typical of Phorbas (see review in van Soest, 2002b). Thus, Lissodendoryx ( E. ) ballena sp. nov. appears indeed closer to Hymedesmiidae rather than to the additional Lissodendoryx spp. described here. Interestingly, Lissodendoryx ( E. ) jenjonesae Picton & Goodwin, 2007 recorded from the Northern Ireland and L. ( E. ) arenaria Burton, 1936 recorded from the South Africa , the two species in the subgenus analysed by Morrow et al. (2013) in their molecular phylogeny, also share features found in Hymedesmiidae . Specifically, a few images observed in Picton & Goodwin (2007: 1449, Figs 6 A–E) record of L. ( E .) jenjonesae acanthostyles with slightly swollen bases, as commonly observed in various species of Hymedesmia , arcuate isochelae slightly bent backwards (cf. Phorbas areolatus ), and a surface with structures that resemble Hymedesmiidae pore fields (cf. van Soest, 2002b). Burton (1936) recorded a skeleton with embedded grains echinated by acanthostyles for L. ( E. ) arenaria , which resembles the embedded polychaete tubes echinated by acanthostyles found in L. ( E. ) ballena sp. nov. ( Fig. 3 G). As in L. ( E. ) jenjonesae , L. ( E. ) ballena sp. nov. has small holes scattered on the surface, which are reminiscent of pore fields. In Hymedesmiidae , pore fields are shallow and have walls supported by diactines (van Soest, 2002a), while the small holes in L. ( E. ) ballena sp. nov. are narrow, single, deep, and their walls are surrounded/supported mainly by microscleres ( Fig. 3 D). If we think of a hypothetical ancestor with holes similar to those in L. ( E. ) ballena sp. nov. , where such structures have become shallower and their ectosomal diactines have invaded the choanosome, we would be very close to sponges currently classified in Hymedesmiidae (cf. van Soest, 2002b). In this way, the phylogenetic relations retrieved by Morrow et al. (2013) might make sense in a morphological context after all. A reticulate choanosomal architecture and ectosomal tylotes with microspined ends are rather common characters of species of Myxilla Schmidt, 1862 (van Soest, 2002c). Based on these observations, L . ( E .) corrugata sp. nov . and L. ( E. ) coloanensis sp. nov. seem more closely related to Myxilla , than L. ( E. ) ballena sp. nov. , and its Hymedesmiidae characters (cf. above). Lissodendoryx ( E .) diegoramirezensis sp. nov. , however, has an architecture of intermediate morphology between reticulate and plumose. Additionally, the latter species’ isochelae are quite distinct from those found in the other species described, and to the best of our knowledge, rather similar only to those of L. ( E. ) anacantha . The shape of tylotes and the spination pattern of the extremities of these ectosomal diactines occur in congruence in part of the new species. These characteristics are most likely reflecting phylogenetic signal in a more inclusive level within Lissodendoryx . For example, thorns arranged as a crown ( L . ( E .) corrugata sp. nov and L . ( E .) coloanensis sp. nov. ) and randomly arranged thorns ( L . ( E .) diegoramirezensis sp. nov. and L. ( E .) anacantha ) might be suggesting these may be evolutionarily species-pairs. The latter pair of species might still be closer to L . ( E .) corrugata sp. nov . and L . ( E .) coloanensis sp. nov. , which share similar ectosomal tylotes, rather than to L. ( E. ) ballena sp. nov. , which has tornotes. Widening the comparison, one cannot dismiss the possibility that the tylotes of L. ( E .) diegoramirezensis sp. nov. and L . ( E. ) anacantha might be close relatives of those found in some Myxilla spp., e.g. Myxilla ( Burtonanchora ) araucana Hajdu, Desqueyroux-Faúndez, Carvalho, Lôbo-Hajdu & Willenz, 2013. Accordingly, it appears clear to us that the set of morphologic characters commonly assessed to describe species in the Poecilosclerida , is likely to permit various phylogenetic reinterpretations of boundaries for higher taxa, among which are those building up in the form of clades in the molecular phylogenetics literature ( e.g. Redmond et al. 2013). In the absence of pre-established phylogenetic frameworks, still, these characters can and should be used to propose primary hypothesis of relationships, as we have done above for L. ( E. ) ballena sp. nov. This species seems to share a series of characters with Hymedesmiidae sponges: small openings on the surface possibly homologous to porefields, as well as arcuate isochelae and acanthostyles with morphology commonly found in this family, thus suggesting that alternative classification scenarios at the family level might be considered for the Poecilosclerida .