Hexactinellida of the genus Rossella, of ANT XXIV / 2 (SYSTCO I) Expedition — Antarctic Eastern Weddell Sea Author Göcke, Christian Author Janussen, Dorte text Zootaxa 2013 3692 1 102 122 journal article 10.11646/zootaxa.3692.1.6 52a51f5e-bc73-4d57-8200-f198a12a905d 1175-5326 224084 E86E41ED-D12B-4E3D-9FA3-25C8B2923183 Rossella racovitzae Topsent, 1901 ( Figs. 2 F, 7, Tab. 6) Synonymy: Rossella racovitzae Topsent, 1901a: 5 –6; 1901b: 33–35, pl. 1, fig. 5, pl. 4, figs. 1–7; 1916: 3; 1917: 9–13, pl. 4, figs. 7–8, 10, pl. 5, fig. 5. Kirkpatrick 1907: 14. Schulze & Kirkpatrick 1910a: 296–297; 1910b: 19–23, pl. 3, figs. 1– 1m , 2. Burton 1929: 407–409, fig. 1, pl. 1; 1932: 256–257; 1934:7. Koltun 1969: map 2; 1976:165 (pars). Barthel & Tendal 1994: 92–95, figs. 35–36, pl. 3–4. Rossella racovitzae minuta Schulze & Kirkpatrick, 1910 a: p. 297; 1910b: p. 23, pl. 4, figs. 1–2. Rossella racovitzae microdiscina Topsent, 1916: 3 ; 1917: 12, pl. 4, fig. 10, pl. 5, fig. 5. Rossella podagrosa Kirkpatrick, 1907: 11 , pl. 3, figs. 2–3, pl. 5, fig.1. Topsent 1917: 14. Rossella podagrosa tenuis Topsent, 1916: 4 ; 1917: 15. Rossella hexactinophila Kirkpatrick, 1907: 12 , pl. 3, fig. 4, pl. 6, fig. 1. Rossella lynchnophora Schulze & Kirkpatrick, 1910a: 298 ; 1910b: 25, pl. 3, fig. 4. Rossella gaussi Schulze & Kirkpatrick, 1910a: 297 . 1910b: 23–25, pl. 4, fig. 1–2d. Rossella mixta Schulze & Kirkpatrick, 1910a: 299 ; 1910b: 29, pl. 3, fig. 3. Rossella vitiosa Hooper & Wiedenmayer, 1994: 527 . (new name for R. gaussi ) Material examined. Several mainly fragmentary sponges from SYSTCO-stations 16-1 (SMF 11870) and 48-1 which originate from approximately 20 specimens (SMF 11729, 11733, 11736, 11871-11900, 11902). Other material examined : IRSNB POR. 0 0 6 ( paratype ), BMNH 1908.2.5.16, BMNH 1926.10.26.89, BMNH 1926.10.26.208. ZMH S2933 ( R. racovitzae minuta ), S2939 ( R. racovitzae racovitzae ), S 2932 ( R. vanhoeffeni armata ). Description. This species shows an extreme variability in its outer shape. Like all Rossella spp., it is normally barrel-shaped. It usually bears conules, which are commonly pointed and may in large specimens be up to 2 cm high and 3 cm in diameter. In smaller specimens the conules are mostly very small and indistinct, while most large specimens are almost smooth with faintly developed conules. Large specimens usually have no distinct protruding spicules apart from those forming the root tuft. Small and medium-sized specimens commonly bear many protruding diactines and often also several protruding pentactins, which sometimes form a veil, that is somewhat like that of R. antarctica , but less distinct. The smallest specimens sampled during the SYSTCO I-Expedition were 2 cm in height. All large specimens were torn apart due to the process of dredging. Thus, we can only guess about the maximal sizes of the largest ones, which were probably about 30 cm in height. FIGURE 7. Spicules of Rossella racovitzae Topsent, 1901 . A, mesodiscohexaster of SMF 11877, B, microdiscohexaster of SMF 11878, C, microdiscohexaster of SMF 11729, D, complete ray of calycocome of SMF 11871, E, complete ray of calycocome with microdiscohexaster between its secondary rays from SMF 11895, F, fragmentary calycocome of SMF 11891. Some representatives of rather unusual spicule shapes: G, calycocome of SMF 11736, rays in front and back are missing, H, complete ray of calycocome of SMF 11736, I, microdiscohexaster of SMF 11736, J, mesodiscohexaster of SMF 11889, K, complete ray of calycocome of SMF 11889. Spicules of R. vanhoeffeni armata Schulze & Kirkpatrick, 1910 (ZMH S2932) for comparison: L, microdiscohexaster, scale bar 10 µm, M, complete ray of calycocome, scale bar 10 µm. A high variation is also found within the spiculation. Even the calycocome, which is usually the sclere with the highest diagnostic value amongst Rossella spp. spicules, occurs in several varieties, some of which can even assume a shape very similar to that of other species (see remarks). The calycocomes ( Figs. 7 D–H) have a total diameter of up to 340 µm (and probably more, as most larger calycocomes were broken and therefore only lengths of complete rays of the largest ones can be reported, Tab. 6), with strong middle pieces of 15 to 35 µm length. The secondary rays are very stout, mostly straight, but occasionally curved, up to 160 µm in length and 3–7 in number. Mesodiscohexasters ( Figs. 7 A, I–J) were found to show an extraordinary size range: 60–175 µm (see remarks). Oxyhexasters and oxyhexactins are extremely large in some specimens (up to 250 µm), but can also be very small (80 µm). Their shape is of the typical kind. Microdiscohexasters ( Figs. 7 B–C, I) have secondary rays of two clearly different lengths and a distinct capitulum at the end of the primary rays, from which the secondary rays originate. They are 40 µm in diameter. TABLE 7. Spicule sizes of Rossella vanhoeffeni Schulze & Kirkpatrick, 1910 . Values in μm are given as follows: minimum – mean – maximum (number of spicules measured). For comparison, values from Schulze & Kirkpatrick (1910b) and Barthel & Tendal (1994) are given. parameter ZMH S2932 Schulze & Kirkpatrick, Schulze & Kirkpatrick, Barthel & Tendal ( 1910 b), . vanhoeffeni ( 1910 b), R. vanhoeffeni ( 1994 ) vanhoeffeni armata rough Pentactin tangential ray (L) 75–104–140 (30) proximal ray (L) 80–89–100 (10) rough Hexactin (D) 175–311.8–470 (30) Stauractin ray (L) 150 (1) Oxyhexactin (D) 100–102.5–110 (4) Oxyhexaster (D) 80–111.8–140 (31) 140 120–130 95–140 Heterodiactin (L) Discohexactin (D) Microdiscohexaster (D) 35–43.4–52.5 (30) 47 46 40–85 Mesodiscohexaster (D) 90–110.6–120 (8) 160 97 90–160 Calycocome

(D)	180–190–200 (2) 140–190	190	140–380
complete ray (L)	60–88.5–100 (30)
primary ray (L) middle piece (L)	6.25–11.4–16.25 (30) 14 2.5–4.1–7.5 (30)	14

secondary ray (L)	47.5–72.4–90 (30)	65
number of sec. rays	3–4.8–7 (30)	2–7	2–7
Curvature	S: 3
S-C: 16
C: 11

Remarks. Two rather unusual specimens have to be discussed in more detail here: SMF 11736 ( Figs. 7 G–I) and SMF 11889 ( Figs. 7 J–K). Both show very distinct similarities to the current concept of R. vanhoeffeni (following Schulze & Kirkpatrick 1910b; Barthel & Tendal 1994), so that the identity of the specimens seemed questionable, until after long consideration we decided on assignment to R. racovitzae . One specimen (SMF 11736) has calycocomes with strongly curved secondary ray, as is the main diagnostic feature of R. vanhoeffeni (Schulze & Kirkpatrick 1910b, see also Fig. 7 m for comparison of R. vanhoeffeni armata original material, of which no official type has been designated yet). The most significant difference of calycoccomes in our specimen to the typical vanhoeffeni-calycocomes are the strong middle pices, which in R. vanhoeffeni are almost absent. Another specimen (SMF 11889) is characterized by very large mesodiscohexasters of up to 175 µm, which according to Barthel and Tendal (1994), following Schulze and Kirkpatrick (1910b), are typical for R. vanhoeffeni . However large mesodiscohexasters up to 168 µm were reported in R. racovitzae by Kirkpatrick (1907). In comparative vanhoeffeni- material (ZMH S2932, Rossella vanhoeffenis armata , Tab. 7 ), mesodiscohexasters of only 90–120 µm could be found. Additional to the large mesodischohexasters, our specimen SMF 11889 has calycocomes with straight secondary rays, as they are typical of R. racovitzae ( Fig. 7 K). Thus, the size of mesodiscohexasters is not a suitable character for species differentiation. We conclude that the only clear character of R. vanhoeffeni appears to be the combination of curved secondary rays and reduced middle pieces in calycocomes. Thus, at the moment, the justification for the species R. vanhoeffeni seems rather weak, it might prove to be but a variety of R. racovitzae . Topsent (1901b) depicted three specimens of the R. racovitzae , two smaller ones with numerous long diactins and a larger specimen with very few protruding spicules. None of these sponges have a velum of pentactins, which is rated as the species most outstanding characteristic by Barthel and Tendal (1994). The first specimen recorded that had a velum was described as R. hexactinophila by Kirkpatrick (1907), but was re-described as a subspecies of R. racovitzae by Schulze and Kirkpatrick (1910a, b) after they realized that a specimen clearly representing R. racovitzae (described as subspecies minuta ) had a velum as well, thus for the first time proving the wide variance in characters of the species. It seems obvious, that the veil of pentactins and even the protruding diactins are characters of juvenile sponges only, which more or less disappear while the sponge is growing. Therefore, an exact identification of this species from its outer appearance is not possible. It has to be pointed out though, that the variability might partly be due to the possibility of this species comprising several cryptic species, as is underlined by the variability in characters between specimens (see above), as well as by the long synonymy list. Still, a splitting into further species on the basis of morphologic characters (as it was tried before) is very likely to lead only to artificial species which will not prove to be consistent when comparing enough material. Our data extends the size range for most spicule categories. In no previous report, oxyhexasters and oxyhexactins larger than 180 µm were documented, here we found some of up to 250 µm in several specimens. Also, the mesodiscohexasters showed an extremely large growth of up to 175 µm (discussed above). The calycocomes documented here are rather small (largest reported by Barthel and Tendal (1994) was 490 µm). Still, in some specimens which were not suitable for general spicule measurements, some fragmentary calycocomes were found which seem to have reached much larger sizes.