Latrunculia (Latrunculia) yepayek Hajdu, Desqueyroux-Faúndez, Carvalho, Lôbo-Hajdu and Willenz sp. nov., 2013

Hajdu, Eduardo, Desqueyroux-Faúndez, Ruth, Carvalho, Mariana De Souza, Lôbo-Hajdu, Gisele & Willenz, Philippe, 2013, Twelve new Demospongiae (Porifera) from Chilean fjords, with remarks upon sponge-derived biogeographic compartments in the SE Pacific, Zootaxa 3744 (1), pp. 1-64 : 37-41

Latrunculia (Latrunculia) yepayek Hajdu, Desqueyroux-Faúndez, Carvalho, Lôbo-Hajdu and Willenz sp. nov.

( Figs. 6D , 7D , 8 BE–BW, 9D; Tab. 6 )

Latrunculia yepayek Willenz & Hajdu, 2009 (nomen nudum) in Willenz et al. (2009: 144)

Type material. Holotype. RBINSc-IG 32233-POR 9940 , Adalberto Channel (48º36’28.70’’S – 74º53’55.70’’W, Chilean Patagonia), 16 m depth, coll. Ph. Willenz and E. Atwood, 11 March 2006 —fragment from holotype MHNG 61467 .

Diagnosis. Latrunculia (L.) yepayek sp. nov. is the only species of Latrunculia (L.) in the Magellanic region with dark-brownish/purplish colour alive, irregularly outlined pore fields on top of papillae densely arranged side by side, a second category of megascleres (oxeas) slightly larger than the styles, anisodiscorhabds with three whorls of spines (basal whorl usually with only isolated thorns, subsidiary whorl 0–3 µm distant from median whorl), and sanidaster-like anisodiscorhabds as microscleres.

Description ( Fig. 6D ). Globular-ovoid sponge, 6 cm in maximum diameter (= height), with conspicuous, regularly distributed, slightly elevated, irregularly outlined papillae (up to 12 mm in maximum diameter) bearing pore fields. Oscula contracted at time of collection and barely visible (2–3 mm in diameter). Consistency firm and rubbery. Colour in life dark purplish-brown, in ethanol slightly faded.

Skeleton ( Fig. 7D ). Ectosome in two layers, intermingled at places, an external one (up to 90 µm thick) intercalating sectors of single layered palisade of mostly erect anisodiscorhabds, and long stretches devoid of any spicules; and a dense basal layer (ca. 60 µm thick) of megascleres mainly arranged parallel to the surface. The latter quite often changes to a criss-crossed disposition inseparable from subectosomal architecture and swallowing the outer layer too, with the consequence that the sponge appears to have at places no ectosomal specialisation of any sort. Choanosome, an irregular reticulation of pauci- to multispicular tracts of megascleres (25–60 µm thick), denser, when closer to the ectosome, forming criss-crossed patterns at their intersection points. Innumerable aquiferous channels (50–200 µm in diameter) perforate the choanosome, especially in its deeper portions, and very large lacunae (ca. 1.2 mm in diameter) occur just underneath the subectosomal region. No specialised architecture accompanies either the channels or the lacunae. A very few microscleres can be seen in the deeper parts of the choanosome

Spicules ( Figs. 8 BE–BW, 9D). Megascleres. Styles ( Figs. 8 BS–BV), mainly smooth, straight or slightly curved, isodiametric, variably sharp apices, occasionally bearing tyles at varied distances from base but mostly not polytylote, 268–360 µm long and 4–10 µm thick. Oxeas ( Fig. 8 BW), rare, mostly bent in the middle (sometimes irregularly crooked), thin, isodiametric, occasionally bearing a few short conical spines, with sharp endings, 280– 420 µm long and 2–8 µm thick. Microscleres of two kinds, with rare intermediate forms. Anisodiscorhabds ( Figs. 8 BE–BN, 9D): manubrium with large thorns, frequently with secondary spines, thorns simple ( Figs. 8 BE, BG, BL, 9D), biphid or polyfurcate ( Fig. 8F ); basal whorl adjacent to manubrium, composed of isolated (unfused) thorns; shaft generally smooth or with a few conical thorns; median whorl conspicuously segmented, each segment with thorns incised variably deep, frequently with secondary spines or serrated ( Fig. 9D ); subsidiary whorl, similar and variously adjacent to median one (0–3 µm distant from it), slightly narrower; apical whorl with thorns incised variably deep, frequently with secondary spines; 38–43 µm long and 19–26 µm across, shaft 4–7 µm wide. Sanidaster-like anisodiscorhabds ( Figs. 8 BO–BR), whorls of spines only barely seen in a few spicules ( Fig. 8 BP), these are substituted by large thorns spread all over the shaft, with or without a sparse secondary apical microspination, sometimes even the manubrium and apex being impossible to differentiate ( Figs. 8 BP–BR); 42–46 µm long and 13–20 µm across, shaft 6.5–9 µm wide.

Distribution and ecology. Known only from its type locality at Adalberto Channel (48º36'28.70"S – 74º53'55.70"W, Chilean Patagonia), where it was collected from 16 m depth. Several other sponges were collected from the same locality, including Amphimedon maresi Sarà, 1978 and Dragmacidon egregium ( Ridley, 1881) .

Etymology. “Yepayek” is used as a noun in apposition. It means “cypress of the Guaitecas” in Kaweskar language and it is the name of the boat owned by CONAF (Corporacion Nacional Forestal de Chile), on board of which an important expedition was held to Chilean fjords between 48 and 51°S in 2006, permitting collection of the species holotype and this far single known specimen.

Remarks. In the absence of the genetic and chemical data deemed essential for species recognition in New Zealand ( Alvarez et al. 2002), our taxonomic efforts on Chilean Latrunculia rested on morphology alone. At first glance, it appeared there might be several species of Latrunculia in the Chilean fjords region, as inferred from their distinct external morphology apparent from the series of underwater pictures taken during the field trips. Characters such as live colour, specimen diameter, and the shape and distribution of aquiferous openings permit recognition of a few morphotypes. Colour varies from purple and dark-purplish-brown to (dark) green. Openings can be spread, densely arranged side by side, flush with the surface or variably raised, round or irregularly outlined. However, microscopic analysis of 16 specimens revealed a rather more conservative spicule complement, where micrometries did not help sorting morphotypes into putative species, forcing a detailed assessment of micromorphological features, which has been accomplished via over 450 electromicrographies. This was necessary also because single specimens studied here appeared to present more disparity in microsclere morphology than several New Zealand species put together ( Alvarez et al. 2002). The strategy of providing a single illustration for the anisodiscorhabds, commonly encountered in the recent literature, appears rather inadequate. In Samaai et al. (2012), albeit mention is made to the fact that “in addition to the typical anisodiscorhabd, there may be [in some Latrunculia ] large spined metaster-like oxydiscorhabds, acanthomicroxeas, or aciculodiscorhabds in which the apex elongates to form a spine, or a form that has numerous whorls along the shaft”, from their comparative assessment of anisodiscorhabd morphology (their Fig. 6 ), one gets the impression that L. crenulata Lévi, 1993 is the only species where variation had previously been found, proportional to those we encountered in Chilean species. After all, it is the only species for which a second illustration of anisodiscorhabds was provided. Lévi (1993), nevertheless, had already mentioned two other species with simultaneous occurrence of “normal” and “elongate” anisodiscorhabds, viz. L. apicalis Ridley and Dendy, 1886 (sensu Boury-Esnault & Van Beveren 1982) and L. multirotalis Topsent, 1927 (sensu Topsent 1928), and we can add here L. (L.) biformis Kirkpatrick, 1908 . We are of the opinion that much intraspecific variation will prove to be hidden in superficial descriptions, or their summarisation in exceedingly concise tables and plates. This is not just a consequence of the usually small numbers of specimens known from each species, but also, and perhaps equally important, of the underestimation of the significance of such variation for the establishment of sharp species boundaries. For instance, where such variation has been spotted, as in the case of L. (L.) fiordensis Alvarez et al., 2002 , it has been deemed irrelevant for diagnostic purposes, in view of an alleged inconsistent distribution among different specimens ( Alvarez et al. 2002).

Given that concise diagnoses such as those based on micrometric values and/or presence/absence of spicule categories ( Table 6 ) do not clearly set Chilean Latrunculia spp. apart from each other, neither from many previously known species, we had to resort to more extensive descriptions of micromorphology in the hope of convincing the reader that species that look different on the exterior can also be set apart through their anatomical features. A similar strategy has been adopted for the study of South African species by Samaai et al. (2003, 2012), but failed to provide clear cut species boundaries for several New Zealand species ( Miller et al. 2001, Alvarez et al. 2002). In the case of the latter, species diagnoses based on morphological traits relied heavily on multivariate statistics, and authors have no option other than resort to live-colour and chemical profiles for keying those species out.

The four proposed new species are diagnosed against each other above, so that in these remarks we will focus on the additional Latrunculia reported in the world. Van Soest et al. (2013) recognise 37 species to be valid in Latrunculia , 15 classified in subgenus Latrunculia , 17 in subgenus Biannulata and 5 left unassigned as to subgenera. From the latter five, we consider L. rugosa ( Vacelet, 1969) and L. sceptrellifera ( Carter, 1887) to be unrecognisable, and only doubtful records of Latrunculia pending further revision, so that only 35 species will be remarked upon below. Although Miller et al. (2001) showed there is genetic back up for the recognition of taxa morphologically characterised by the occurrence of two vs. three whorled anisodiscorhabds, our observations, as well as fragmentary information available from the literature ( Lévi 1993, Alvarez et al. 2002), indicate that there may be variation in this character, thus suggesting that more comprehensive comparisons are needed in order to firmly establish the status of our newly proposed species. Accordingly, in spite of classifying all four new Chilean Latrunculia within the nominotypical subgenus, we also performed a thorough comparison with species of L. ( Biannulata ) as well as those incertae sedis.

From easiest to hardest to figure, we follow the order megasclere length-microsclere length (anisodiscorhabd or equivalent)-megasclere micromorphology-microsclere micromorphology in structuring this comparison. Accordingly, five Latrunculia spp. exhibiting megascleres reaching over 500 µm in length are at once deemed distinct from all four new species. These are L. (L.) bocagei Ridley & Dendy, 1886 , L. (L.) basalis Kirkpatrick, 1908 , L. (L.) brevis Ridley & Dendy, 1886 , L. tricincta Hentschel, 1929 (subgenus incertae sedis) and L. velera Lehnert et al., 2006 (subgenus incertae sedis). Conversely, L. (L.) cratera Bocage, 1869 , with megascleres seemingly always smaller than 200 µm is obviously quite different from the Chilean species. A series of species possesses microscleres (anisodiscorhabds or derived forms) consistently larger than 50 µm, and are also considered easily distinguished from all four new species. This group includes 13 species, among which L. (L.) bocagei , L. (L.) basalis , and L. (L.) brevis , already sorted above. The remaining 10 species are L. (L.) apicalis Ridley & Dendy, 1886 , L. (L.) biformis Kirkpatrick, 1908 , L. (L.) palmata Lévi, 1993 , L. (L.) multirotalis Topsent, 1927 , L. (L.) crenulata Lévi, 1993 , L. (L.) novaecaledoniae Samaai et al., 2006 , L. (B.) austini Samaai et al., 2006 , L. (B.) lunaviridis Samaai et al., 2003 , L. (B.) microacanthoxea Samaai et al., 2003 , and L. (L.) tetraverticillata Mothes et al., 2008 .

Micromorphological features of the megascleres are simply the consistent occurrence of spines on those of a very few species, or the oxeote or nearly so terminations in others. Species with acanthostyles include L. (L.) verenae sp. nov., and the previously known L. velera Lehnert et al. 2006 (subgenus incertae sedis), L. oparinae Samaai & Krasokhin, 2002 (subgenus incertae sedis) and L. (B.) kerwathi Samaai et al., 2012 . The latter three species not only have spines distributed differently, closer to their base as opposed to more densely arranged on both extremities as is observed in the new species, but also exhibit anisodiscorhabds of markedly distinct shape, an aspect that will be further detailed below. The consistent occurrence of these spines further differentiates L. (L.) verenae sp. nov. from all remaining Latrunculia spp. Two species are known with oxeas or oxeote styles as the main megascleres, L. (B.) purpurea Carter, 1881 and L. (B.) spinispiraefera Brøndsted, 1924b , respectively. Thus, they are easily distinguished from all four new species proposed here.

Microsclere micromorphology plays an essential role in separating species of Latrunculia , and is relevant here to differentiate L. (L.) verenae sp. nov. from additional species bearing acanthostyles, as exposed above. It also assures differentiation of the three Chilean species and 19 further species known worldwide bearing only smooth styles of comparable dimensions (usually under 400–450 µm long). All three previously known Latrunculia spp. with acanthostyles bear anisodiscorhabds markedly distinct from those of L. (L.) verenae sp. nov. In summary these are much stouter, but additional distinctive micromorphological characters are the regular blunt denticulated thorns of median whorls in L. (subgen. i.s.) velera , the exceedingly wide basal whorl in L. (subgen. i.s.) oparinae [mistakenly referred by Samaai et al. (2012) to their Fig. 6 AD, the correct spicule is shown in Fig. 6 AC], and the more conical and paucispinose (stouter), as opposed to flat and denticulated more slender), median and subsidiary whorls in L. (B.) kerwathi . L. (L.) verenae sp. nov. appears thus to obviously merit status as a new species.

Within the 19 species referred to above, ten are currently classified in Latrunculia (Biannulata) , and where, judging from SEM data provided in Samaai et al. (2012) and references therein, the basal whorl of spines and manubrium are completely merged in a cuff of spines, distinction from the Chilean species is not difficult to realise. Nevertheless, this complete merge is taken here to be exhibited by eight species only: the New Zealand L. (B.) kaakaariki Alvarez et al., 2002 , L. (B.) duckworthi Alvarez et al., 2002 , L. (B.) procumbens Alvarez et al., 2002 , L. (B.) kaikoura Alvarez et al., 2002 , L. (B.) wellingtonensis Alvarez et al., 2002 , and L. (B.) millerae Alvarez et al., 2002 , the South African L. (B.) algoaensis Samaai et al., 2012 , and the south-eastern Brazilian L. (B.) janeirensis Cordonis et al., 2012 . In these species basal spines frequently bi- or trifurcate parallel to the axis, while in L. (B.) gotzi Samaai et al., 2012 and L. (B.) kerwathi the spines can still be separated into those slanted up and those down, and one could argue that an incipient basal whorl exists. The marked stoutness of anisodiscorhabds, and the abundant occurrence of spines pointing up or down from the median and subsidiary whorls sets L. (B.) gotzi apart from the new species. Assignment of this South African species to subgenus Biannulata appears to us to be established on weak grounds, and should better be verified through an independent dataset. Latrunculia (B.) kerwathi cannot be mistaken for any of the remaining new species because of the conical and paucispinose (stouter), as opposed to flat and denticulated (more slender), median and subsidiary whorls of its anisodiscorhabds, as just pointed out in reference to L. (L.) verenae sp. nov.

The complete merge of basal whorl and manubrium exhibited by the eight Latrunculia (B.) listed above is unmatched by all four new species proposed here, at least as a standard morphology. Evidence for lack of basal whorl, or its possible merging with the manubrium is apparent in Figs. 8E–F, T, W , but these are taken to represent deviate morphologies. Fully developed anisodiscorhabds, or simply put, the extreme range of morphological plasticity in the new species proposed, always includes a clearly discernible, albeit not necessarily complete basal whorl of spines ( Fig. 9A–D ). This is strong evidence of closer affinity to the triverticillata- group ( Miller et al. 2001, Alvarez et al. 2002), currently recognised as subgenus Latrunculia . Whereas, it has been shown through integrative taxonomy ( Miller et al. 2001) that the six New Zealand species among the seven species in question (cf above) do not belong to this group. Accordingly, the remaining three new species proposed, viz. L. (L.) ciruela sp. nov., L. (L.) copihuensis sp. nov. and L. (L.) yepayek sp. nov. are considered to be clearly distinct from all other species of Latrunculia .

The presently proposed synonymisation of Chilean records of L. (L.) lendenfeldi [sensu Desqueyroux (1976)] with L. (L.) ciruela sp. nov. and L. (L.) verenae sp. nov. is mostly based on SEM study of their anisodiscorhabds. MHNG 62578 and 62579 were observed to have the basal whorl of spines adjacent to the manubrium, while MHNG 62580 has a clearly separate basal whorl, a feature only observed in L. (L.) verenae sp. nov. The first two were assigned to L. (L.) ciruela sp. nov. also for the apparent lack of sanidaster-like microscleres and of acanthose megascleres. Hence, the four new species described here are the only known species of Latrunculia from the entire SE Pacific, for which we propose the following identification key.

Figures

FIGURE 6. In situ photographs of Latrunculia ciruela sp. nov. (A, IZUA-POR 145, holotype), Latrunculia copihuensis sp. nov. (B, RBINSc-IG 32233-POR 9915, holotype), Latrunculia verenae sp. nov. (C, IZUA-POR 146, holotype), Latrunculia yepayek sp. nov. (D, RBINSc-IG 32233-POR 9940, holotype), Clathria mytilifila sp. nov. (E, MNRJ 8198, holotype), Neopodospongia tupecomareni sp. nov. (F, MNRJ 10787, paratype), and Myxilla araucana sp. nov. (G, MNRJ 8220, holotype). Scale bars: A–D, 5 cm; E–G, 2 cm.

FIGURE 7. Skeletal architectures in transverse sections of Latrunculia ciruela sp. nov. (A, IZUA-POR 145, holotype), Latrunculia copihuensis sp. nov. (B, RBINSc-IG 32233-POR 9915, holotype), Latrunculia verenae sp. nov. (C, IZUA-POR 146, holotype), and Latrunculia yepayek sp. nov. (D, RBINSc-IG 32233-POR 9940, holotype). Scale bars: A–D, 500 µm.

FIGURE 8. Latrunculia ciruela sp. nov. (A–R): A–N, anisodiscorhabds; O, (subtylo)style; P, oxea; Q–R, details of oxeas. Holotype (IZUA-POR 145): F, J. Paratypes: MNRJ 9186, M–N; MNRJ 10811, G, P–R; MNRJ 10844, C, E; MNRJ 10849, K; MNRJ 10855, D, I, O; MNRJ 10861, A, L; MNRJ 10921, H; MNRJ 12928, B. Latrunculia copihuensis sp. nov. (S–AL, RBINSc-IG 32233-POR 9915, holotype): S–AF, anisodiscorhabds; AG–AH, oxeas; AJ–AL, details of (subtylo)styles. Latrunculia verenae sp. nov. (AM–BD, type series): AM–AZ, anisodiscorhabds; BA–BB, acanthostyles; BC–BD, details of acanthostyles. Holotype (IZUA-POR 146): AM–AS, BA. Paratypes: MNRJ 10914, AT–AV; MNRJ 10942, AW–AZ, BB–BD. Latrunculia yepayek sp. nov. (BE–BW, RBINSc-IG 32233-POR 9940, holotype): BE–BR, anisodiscorhabds; BS–BU, styles; BV–BW, details of styles. Scale bars: A–N, Q–AF, AJ–AZ, BE–BR, 10 µm; O–P, AG–AI, BA–BB, BS–BW, 100 µm; BC–BD, BU–BV, 25 µm.

FIGURE 9. Anisodiscorhabds in detail: Latrunculia ciruela sp. nov. (A, IZUA-POR 145, holotype); Latrunculia copihuensis sp. nov. (B, RBINSc-IG 32233-POR 9915, holotype); Latrunculia verenae sp. nov. (C, IZUA-POR 146, holotype); Latrunculia yepayek sp. nov. (D, RBINSc-IG 32233-POR 9940, holotype). Scale bar: 20 µm.