Geodia arma n. sp.

(Figs. 4 & 5; Table 2)

Material examined. Holotype MABIK IV00158487, specimen dried, collected by Chano Hwang with a bottom trawl aboard the FV Oyang 96, 6 April 2014, 298 m depth, Koko Guyot, Emperor Seamounts, North Pacific Ocean (35° 20.202´N, 171° 32.598´E) . Paratypes MABIK IV00158489, specimen dried, collected by Chano Hwang with a bottom trawl aboard the FV Oyang 96, 14 April 2014, 297 m depth, Koko Guyot, Emperor Seamounts, North Pacific Ocean (35° 28.800´N, 171° 23.202´E) ; MABIK IV00158492, specimen dried, collected by Chano Hwang with a bottom trawl aboard the FV Oyang 96, 5 May 2014, 300 m depth, Koko Guyot, Emperor Seamounts, North Pacific Ocean (35° 18.300´N, 171° 29.502´E) .

Description. Three specimens of this semispherical sponge were collected and have diameters of 3.4–12 cm (Figs. 4A, B, & C). The specimens examined are half spheres with one side covered by a thick cortex and the other side with the choanosome exposed (Figs. 4B & C); probably the sponges were broken when trawling them off the substrate, we suspect we only got the upper parts of the specimens. The cortex of sterrasters is 1.5–6 mm in diameter and consequently hard and inelastic (Fig. 4C). The surface is optically smooth but rough to the touch. One specimen has the surface partially covered with wart-like, slightly elevated apertures 2.5–4 mm in diameter (Fig. 4A). A second specimen has smaller apertures, 0.5–1mm in diameter. There are no other apertures visible, even when examined under a dissecting microscope.

Skeletal architecture. Choanosomal confused arrangement becomes radial towards the surface, the long oxeas reaching into the thick cortex of aspidasters (Fig. 4D).

Spicules. Megascleres are thin, sinuous (Fig. 4F) to thicker, almost-straight oxeas to strongyloxeas (Fig. 4E) in a wide size range, 280 to over 6000 x 16–42 µm. Long oxeas are mostly broken in spicule preparations so the lengths are hard to measure as one has to follow the course of the spicules in the slide preparations. Microscleres are oval sterrasters (Figs. 4D, 5A, B, C), 118–158–171 x 202–221– 243 µm, oxyasters with microacanthose rays (Figs. 5D & E), 18–20– 23 µm in diameter and, tylasters (Figs. 5D & F), 8–9– 11 µm in diameter.

Discussion. We compared the new species with all other species of Geodia lacking triaenes, a genus formerly referred to as Geodinella . The genus Geodinella was erected by Lendenfeld (1903, p. 117) for Geodia cylindrica Thiele, 1898, a species with reduced triaenes, with the clads of triaenes reduced to two or less so that these spicules resemble diaenes, monaenes, styles or tylostyles. Uriz (2002) did not consider Geodinella as a valid genus and listed it among the synonyms of Geodia (p. 138) which is likely the reason why Van Soest et al. (2018) also regard Geodinella as a synonym of Geodia . However, Uriz (2002) in the same publication remarked regarding the genus Sidonops (p. 140) “Aborted triaenes as those in Geodinella have also been described in some species of Isops, Geo- dia, and Sidonops and, consequently, this character has no diagnostic value. As the inhalant orifices of Geodinella are uniporal and the oscula are cribriporal, this genus becomes synonymous of Sidonops .” So Uriz (2002) argued that Geodinella is synonymous with Sidonops but listed Geodinella among the synonyms of Geodia . The question regarding assignment of Geodinella to Geodia or Sidonops may be irrelevant when taking into consideration the publications of Cárdenas et al. (2010, 2011). Based on genetic sequences and morphological characters they suggest to synonymize the genera Geodia, Isops and Sidonops under Geodia . Consequently the characters uniporal/cribripo- ral apertures would lose their importance and Geodinella would either have to be regarded as a genus of its own or fall into synonymy with Geodia . It is not within the scope of this publication to rule on this issue so we follow the current scheme outlined in Van Soest et al. (2018). For recent reviews of the Geodiidae we refer to Cárdenas et al. (2010) and Sim-Smith & Kelly (2015). For the present study the “ Geodinella group” was helpful to detect species with reduced or no triaenes which we used to compare with the species described here.

Geodia arma is the only species which has oxeas which are over 6 mm in length and the width of the oxeas also includes thin, sinuous oxeas. It has the thickest cortex of all described species. The new species differs from compared congeners in the following characters (Table 2):

G. cylindrica has only one category of oxeas (only half the length of the shorter oxeas of G. arma), plagiomon- and -diaenes, smaller sterrasters, a considerably thinner cortex, and no oxyasters.

*: measurements from Lévi & Lévi, 1989

G. hyotania has only one category of oxeas, additional styles and subtylostyles, sterrasters are much smaller, the cortex is thinner, smaller spherasters than the oxeas of G. arma, and no tylasters.

G. isabella has two categories of oxeas (both approximately 1/10 th the size in G. arma), sterrasters (approximately 1/7 th the size in G. arma), a thinner cortex, two categories of oxyasters (one larger and the other smaller than those in G. arma), and no tylasters.

G. lendenfeldi has oxeas which are blunt, additional styles, strongyles, tylotes and tylostyles, spicules which do not occur and which never reach the length of the oxeas and styles in G. arma . The width of monaxons of G. lendenfeldi are between 40–80 µm and are not sinuous while the monaxons of G. arma can be sinuous and are considerably thinner G. lendenfeldi has plagiomonaenes, ortho- and promonaenes which are completely missing in G. arma . G. Lendenfeldi has a thinner cortex.

G. spherastrosa has only one category of oxeas and additional reduced triaenes resembling tylostyles, smaller sterrasters, a thinner cortex, and larger oxyasters.

G. vestigifera has three categories of oxeas (none reaching the length of the thin, sinuous oxeas of G. arma), triaenes resembling styles, smaller sterrasters, a thinner cortex, two categories of oxyasters (the larger category larger than those in G. arma), strongylasters and strongylospherasters, and no tylasters.

Etymology. Named from the Latin arma—armour, referring to the unusual thick cortex of the new species.