New and non-indigenous species of Bryozoa from Iberian waters Author Reverter-Gil, Oscar 001DFC48-0F01-43AF-90EC-452BE6954DDF Museo de Historia Natural da Universidade de Santiago de Compostela, Parque Vista Alegre s / n, 15705 Santiago de Compostela, Spain. oscar.reverter@usc.es Author Souto, Javier E14FC8A9-40B5-4BF6-8551-3F2A1D991463 Institut für Paläontologie, Fakultät für Geowissenschaften, Geographie und Astronomie, Geozentrum, Universität Wien, Josef-Holaubek-Platz 2, 1090 Wien, Austria. javier.souto-derungs@univie.ac.at text European Journal of Taxonomy 2023 2023-07-26 885 33 64 http://dx.doi.org/10.5852/ejt.2023.885.2187 journal article 10.5852/ejt.2023.885.2187 a959a92c-3474-4243-8228-9a9ffb751549 2118-9773 8205406 09B12B7A-5A03-447C-826C-10EA479BCCE5 Arbopercula angulata ( Levinsen, 1909 ) Figs 1–3 , 12A ; Tables 1–2 Electra angulata Levinsen, 1909: 149 , pl. 22 fig. 4a. Electra angulata – Harmer 1926: 207 , pl. 13 fig. 11. Electra tenella – Marcus 1937: 38 , pl. 7 fig. 15a–c. — Silén 1941: 18 , fig. 14. — Mawatari 1974: 42 , fig. 7, pl. 3 figs 3–4. — Rosso 1994: 241–246 , pls 1–2. Electra cf. tenella – López de la Cuadra & García-Gómez 1994: 109 . Arbopercula tenella – Subías-Baratau et al . 2022: 3 , fig. 4b. Arbocuspis angulata – Tilbrook & Gordon 2015: 260 (lapsus calami, see below). Arbopercula angulata – Reverter-Gil & Souto 2021a: 15 . Non Membranipora tenella – Hincks 1880: 376 , pl. 16 fig. 7 [= Arbopercula tenella ]. Material examined Holotype (by monotypy) SOUTH CHINA SEA • several fragments on wood; Gulf of Thailand , near Koh Samet ; coordinates not provided; depth 0 m (floating); 9 Feb. 1900 ; Mortensen leg.; NHMD-77254 ( Fig. 1 ). Other material MALAY ARCHIPELAGO • 1 colony on a wooden slide; Philippines ; 11° 37′ N , 123° 31′ E ; stn 208; depth 18 fms ; date unknown; Challenger exped. 1873–1876; NHMUK 1899.7.1.5088 • 1 colony on wood; Salah Bay , N. Suinbawa ; coordinates not provided; stn 312; depth 274 m ; 1899; Siboga exped.; NHMUK 1928.3.6.14 , NHMUK 1928.3.6.16 • 2 colonies on wood; Java Sea , Indonesia , Strait of Bali ; 8° 20′ S , 114° 25′ E ; depth 30 m ; 19 Jul. 1984 ; Snellius exped. II; G.C. Cadée leg.; boomkor [beam trawl]; CAD84/11; NHMUK 2003.2.28.104 . MEDITERRANEAN SPAIN • 1 large colony on a plastic debris; Balearic Islands , Menorca , Es Talaier ; 39.92667° N , 3.90278° E ; depth 0 m (washed upon the beach); 7 Jul. 2012 ; O. Reverter-Gil leg.; MHNUSC-Bry 643 ( Figs 2–3 ) . NORTH AMERICA • 1 colony on drift plastic; Florida , Fort Pierce , N. Beach ; coordinates not provided; depth unknown; date unknown; NHMUK 1986.8.14.7 . STRAIT OF GIBRALTAR • 1 young colony on wood, coated; Andalucía , La Línea ; 36.18167° N , 05.33333° W ; depth 0 m (floating); 7 Dec. 1986 ; C.M. López-Fé leg.; pers. coll. C.M. López-Fé • several fragments on wood; Andalucía , La Línea ; 36.18167° N , 05.33333° W ; depth: 0 m (floating); Aug. 2005 ; C.M. López-Fé leg.; MHNUSC-Bry 708 . Holotype of Arbopercula tenella NORTH AMERICA • 1 colony on algae; Florida ; coordinates unknown; depth unknown date; Hincks leg.; NHMUK 1899.5.1.648 ( Fig. 4 ) . Other material of Arbopercula tenella NORTH AMERICA • several colonies on glass; Florida , Biscayne Bay , Miami Beach ; coordinates unkown; depth unkown; 1945; L.W. Hutchins leg.; NHMUK 1947.2.4.1 . Table 1. Measurements (in mm) of Arbopercula angulata , holotype (NHMD-77254). Abbreviations: N = number of measurements; SD = standard deviation.

Mean	SD	Minimum	Maximum	N
Autozooid length	0.528	0.0455	0.450	0.676	46
Autozooid width	0.274	0.0271	0.216	0.349	46
Opesia length	0.408	0.0324	0.343	0.491	46
Opesia width	0.222	0.0257	0.177	0.296	46

Description Colony encrusting, multiserial, unilaminar or occasionally multilaminar, forming extensive crusts that cover the substrate on which they grow. Autozooids elongate oval or rectangular, arranged in series, separated by shallow grooves. Distal wall generally ascending towards the frontal surface and angularly bent from side to side or arch-like. Gymnocyst reduced to the proximal region. Two (rarely three) conical, hollow processes, generally open at the end, developed on the proximal gymnocyst, half-way between the central line and the lateral margins. These processes may sometimes be rudimentary or even absent in large parts of the colony. The first zooid in each of both series after bifurcation bears a single median process. Cryptocyst granular, poorly developed, absent at the distal end of the opesia, somewhat more evident at its proximal end, leaving an extensive, oval opesia. Communication via two rather large, multiporous rosette-plates situated in the basal corners of the distal wall. The distal half of each lateral wall has a single multiporous rosette-plate. Oral spines absent. The development of marginal spines shows great differences. Some zooids bear up to 8 pairs of marginal spines, not very thick, somewhat flattened, recurved on the opesia, reaching the middle of the area or even surpassing it, but in general the spines are smaller and fewer, and many zooids are completely spineless. There are no ovicells or avicularia. Irregular intercalary kenozooids, small, scattered, filling gaps between autozooids. Ancestrula unknown. Fig. 1. Arbopercula angulata ( Levinsen, 1909 ) , holotype (NHMD-77254), Koh Samet (Gulf of Thailand). A . Autozooids with well-developed spines and gymnocystal processes. B–C . Autozooids with fewer, poorly developed spines and gymnocystal processes. D . Autozooids without spines and gymnocystal processes. Fig. 2. Arbopercula angulata ( Levinsen, 1909 ) (MHNUSC-Bry 643), Menorca (Balearic Islands). A–B . View of part of large colony on plastic. C–E . Autozooids with fewer and less-developed spines. F . Second layer of spineless autozooids covering the first layer of spiny zooids. G . Autozooids of the growing margin with developing gymnocystal processes. Table 2. Measurements (in mm) of Arbopercula angulata (MHNUSC-Bry 643). Abbreviations: N = number of measurements; SD = standard deviation.

Mean	SD	Minimum	Maximum	N
Autozooid length	0.578	0.0585	0.500	0.712	23
Autozooid width	0.279	0.0424	0.216	0.384	23
Opesia length	0.485	0.0393	0.405	0.578	23
Opesia width	0.234	0.0370	0.178	0.324	23

Remarks Electra angulata was originally described by Levinsen (1909) from colonies collected on a ligneous core floating near Koh Samit, Siam (Ko Samet, Thailand ). The description is very clear and complete, but not so the only drawing ( Levinsen 1909 : pl. 22 fig. 4a), which provides no information on the variability of the species. Levinsen (1909: 149) , however, clearly stated that “The best provided ones [autozooids], which in the colonies examined are in a great minority, have on the margin 12 not very thick spines, which reach the middle of the area or even surpass it. A larger or smaller number of them is however often wanting, and many zooecia are altogether without spines. On the proximal gymnocyst we find in most zooecia 2 (more rarely a single median and still more seldom 3) short, thick, conical spines, generally open at the end, which are situated half-way between the central line and the lateral margins. These spines may sometimes be rudimentary, and in many zooecia (with or without marginal spines) they are absent”. On page 156 of the same paper, it is also stated that “Here we may find in the same colony some zooecia, which are entirely without spines, and others provided with a larger number of these structures.” It is clear then that according to Levinsen (1909) the zooids of E. angulata may have spines and gymnocystal processes at the same time, or only processes, or only spines, or even none of them, and both processes and spines may be very variably developed. We have been able to verify all this variation both in the type material of E. angulata (NHMD-77254 and Fig. 1 ), in our own material from the Mediterranean and the Strait of Gibraltar (MHNUSC-Bry 643, 708) ( Figs 2–3 ), and in different museum samples (see Material examined above). Nonetheless, later authors have apparantly incorrectly considered that only the presence of spines (as well as gymnocystal processes) is a typical character of the species. Fig. 3. Arbopercula angulata ( Levinsen, 1909 ) (MHNUSC-Bry 643), Menorca (Balearic Islands). A . Autozooids with well-developed spines and gymnocystal processes. B . Multiporous rosette-plates. C . Spineless autozooids. D . Second layer of spineless autozooids covering the first layer of spiny zooids. The species also has a marked tendency to form multilaminar colonies due to the overgrowth of some layers over others ( Figs 2F , 3D ). This does not seem to have been pointed out by other authors, but is clearly visible in our own material, a large colony on plastic debris (MHNUSC-Bry 643) ( Fig. 2A–B ). The upper layers are often spineless to a great extent ( Figs 2F , 3D ). Membranipora tenella Hincks, 1880 has also been subject to misinterpretation and even considered a synonym of E. angulata or at least misidentified (e.g., Marcus 1937 ; Silén 1941 ; Mawatari 1974 ; Rosso 1994 ; Subías-Baratau et al . 2022 ). However, as already stated about twenty years ago by Tilbrook et al . (2001) , they are clearly separate species. These authors ( Tilbrook et al . 2001: 40 ) first pointed out as a difference that M. tenella sensu stricto bears no marginal spines (see Fig. 4 ), but as we have already demonstrated (see above), zooids of E. angulata usually lack them. Tilbrook et al . (2001) did, however, accurately point out another important difference between the two species: the development of the gymnocystal processes in M. tenella , which are far more robust and knob-like, occupying a greater area of the gymnocyst ( Fig. 4A ). Importantly, the revision of type material of M. tenella (NHMUK 1899.5.1.684, Fig. 4 ) shows that the proximal cryptocyst is characteristically much more developed in this species than in E. angulata (see Figs 3A, C , 4A ). Fig. 4. Arbopercula tenella ( Hincks, 1880 ) , holotype (NHMUK 1899.5.1.648), Florida (USA). A . View of colony. Note the autozooids with well-developed proximal cryptocyst and gymnocystal processes. B . Autozooid with multiporous rosette-plates in basal corners of distal wall. (Photos by M.E. Spencer Jones). The resemblance of E. angulata with Conopeum papillorum Tilbrook, Hayward & Gordon, 2001 was already discussed in the original paper ( Tilbrook et al . 2001: 40 ) and we have nothing more to add. Finally, material of E. angulata was also considered as a new species (named Electra inexpectata ) in the unpublished PhD by López de la Cuadra (1991) , but this species was not formally published. It was reported later from the Strait of Gibraltar area by López de la Cuadra & García-Gómez (1994) as Electra cf. tenella . The systematic position of E. angulata has also been subject of discussion ever since its original description. Levinsen (1909) placed the species in Electra with some reservations, closely allied to Electra monostachys ( Busk, 1854 ) . This position was also accepted by Harmer (1926) . But the original author himself ( Levinsen 1909: 160 ) also related his new species to Aspidelectra melolontha ( Landsborough, 1852 ) , the type species and, at that time, the only species of his new genus Aspidelectra Levinsen, 1909 . Indeed, as Levinsen (1909) stated, both species share an angularly bent distal wall with a multiporous rosette-plate in each of the two basal corners, and 1–2 gymnocystal processes. These processes somehow replace the oral spines, absent in both species. Note that, for this reason, Aspidelectra defensa ( Kirkpatrick, 1888 ) and Aspidelectra densuense Cook, 1968 cannot remain in this genus because they do have true oral spines and lack gymnocystal proximal processes. This requires further discussion, which is beyond the scope of the present work. Anyway, in our opinion E. angulata cannot be placed in this genus either because Aspidelectra is characterized by a frontal shield of fused, flattened spines. In contrast, M. tenella was firstly placed in Electra Lamouroux, 1816 by Marcus (1937) , and the name Electra tenella was widely used by later authors, even although the species cited by Marcus (1937) and others was actually E. angulata . Nikulina (2007 , 2010 ) and Nikulina & Schäfer (2008) distributed a number of species previously attributed to Electra to new genera, but these did not include E. angulata and M. tenella . Tilbrook & Gordon (2015) pointed out some similarities of both species with Arbopercula bengalensis ( Stoliczka, 1869 ) , the type species of the genus Arbopercula Nikulina, 2010 , but they inadvertently erred in attributing both species to Arbocuspis Nikulina, 2010 in their paper. This was clearly a lapsus owing to the similarity of both generic names (D.P. Gordon pers. com.). Electra angulata and M. tenella do not conform to Arbocuspis as the cryptocyst in this genus is inconspicuous and spines are branching, bending across the opesia from its proximal end, forming a sort of shield. We agree with the original intention of the authors since the relationship between the three species is evident, but the inclusion of E. angulata and M. tenella in Arbopercula requires modifying the diagnosis of the genus, which also contains a serious error of understanding and an important omission: Firstly, Arbopercula was originally characterized (and named) based on a pair of bifurcating, chitinous spines on the operculum, which are absent in E. angulata and M. tenella . As Tilbrook & Gordon (2015) pointed out, however, these spines may be very small and easily overlooked. In our opinion, this character can be useful to differentiate the type species, but not to characterize the genus, so it should be eliminated from the generic diagnosis. Otherwise, no other species could be integrated into the genus because this character is exclusive of A. bengalensis . Secondly, the diagnosis of Arbopercula , but also Arbocuspis , is incorrect. Both diagnoses state that the zooids exhibit a pair of distal spines, stout, conical, non-articulated, but these are really gymnocystal blunt processes (not true spines) and are located at the proximalmost end of the succeeding zooid, such as those present in A. bengalensis , E. angulata and M. tenella (but also in A. melolontha ), probably replacing the oral spines, absent in these species. Thirdly, the absence of marginal spines in M. tenella and in many zooids of E. angulata must be included in the diagnosis. Finally, unfortunately no information was provided about the interzooidal communication pores of both genera. We have no information about A. bengalensis , but E. angulata and M. tenella have on either side a rather large, multiporous rosette-plate situated in one of the basal corners of the distal wall. In our opionion this must be also incorporated into the diagnosis of the genus Arbopercula . Tilbrook & Gordon (2015) tentatively also added Membranipora devinensis Robertson, 1921 to Arbopercula , but the presence of proximal pores and a presumed ovicell prevents the inclusion of this species, at least until it is correctly described. Material of A. angulata has been frequently incorrectly reported as E. tenella by many authors around the world, growing on different algae, on drifting or beach-stranded plastic or wood, as small colonies on the hulls of pleasure craft plying tropical and subtropical waters, or even as epibionts attached to the scales of sea snakes, shells of living nautilus or carapaces of sea turtles or horseshoe crabs (e.g., Key et al . 1995 , 1996 ; Pfaller et al . 2008 ; Gordon 2009 ; Tan et al . 2011 ; Subías-Baratau et al . 2022 ). Its first record in European waters was published by Rosso (1994 as E. tenella ) from Sicily. In Iberian waters ( Fig. 12A ), colonies have sometimes been observed on ligneous cores floating near the Strait of Gibraltar ( López de la Cuadra & García-Gómez 1994 , as Electra cf. tenella ; C.M. López-Fé pers. com.; MHNUSC-Bry 708). Colonies have very recently also been reported near the coast of Catalonia growing on plastic debris ( Subías-Baratau et al . 2022 as A. tenella ). Moreover, we collected a large colony also growing on plastic debris washed upon a beach at Menorca (Balearics) in July 2012 ( Reverter-Gil & Souto 2021a ) (MHNUSC-Bry 643; Figs 2–3 ). Because colonies of this species were not observed in Iberian waters growing on fixed substrates, the large number of recent records shows that A. angulata must be considered as a non-indigenous species (NIS) in Iberian waters, attached to different floating substrates.