Taxonomic revision and systematic notes on some Halecium species (Cnidaria, Hydrozoa)

Although the genus Halecium is easy to recognize, identifications at the species level are often difficult, this even for quite common and supposedly well‐known species of the north‐eastern Atlantic. This paper revises and re‐describes some Halecium species which resemble each other closely and which are not easy to distinguish. Additional information on a few rare species is also provided. The study is based on material collected from the North Atlantic, Mediterranean, South Africa, and New Zealand. Halecium scutum Clark, 1877 is recognized as a valid species and distinct from both H. beanii and H. halecinum; colony form and microscopic characters allow a distinction. The pinnate colony form of H. halecinum is a characteristic trait, but not all colonies show this growth form. Halecium beanii can occur in monosiphonic and polysiphonic colonies. Monosiphonic colonies of H. beanii have probably been misidentified by some authors as H. lankesterii. The differences of H. lankesteri to H. beanii and H. petrosum are discussed. The South African population of the reportedly cosmopolitan H. beanii has distinct gonothecae and could belong to a separate species. The Mediterranean Halecium mediterraneum is hardly distinguishable from the New Zealandic H. delicatulum, but it is kept separate mainly for biogeographic reasons. The rare Mediterranean Halecium banyulense is re‐described based on a second find from Naples. The male gonothecae of Halecium corrugatissimum are described for the first time.


Introduction
The genus Halecium comprises currently about 120 valid nominal species (Vervoort and Watson 2003). They often form large, conspicuous colonies and can be found in all oceans. The genus is also well represented in shallow waters and therefore regularly mentioned in faunal synopses and ecological surveys. Due to the characteristic short hydrothecae, it is quite easy to recognize members of the genus. However, many nominal Halecium species are very difficult to distinguish and the status of a considerable number of them is still unclear. Species identification relies mostly on microscopic characters and due to the relative simplicity and morphological variability it is sometimes quite a cumbersome task. Taxonomically, the shape of the female gonothecae is certainly the most important character. Several species are currently only reliably distinguishable if female gonothecae are present. For these species, it is therefore impossible to identify immature or male specimens, at least if they are not part of a series that also includes reproductive female colonies. Although it has been known for quite some time that single, immature colonies are often not identifiable, many authors nevertheless continued to do so. These insecure identifications contributed largely to the difficult taxonomic situation of the genus Halecium. Many of its nominal species are indeterminate and dubious.
Some characters rarely used in Halecium systematics could improve the problem. Cornelius (1998) found some important differences in the living hydranths. Unfortunately, these traits are mostly impossible to study in preserved material and material from deeper waters is usually dredged and therefore often moribund or dead once it reaches the surface. Other largely unexploited characters are the nematocysts and zooxanthellae, which to some degree can also be observed in preserved material. The colony form and the branching pattern are other important characters that have unfortunately been neglected too often.
Another important aspect of the hydrozoans is their large number of cosmopolitan or circumglobal species (cf. Boero and Bouillon 1993). Some Halecium species are also reportedly cosmopolitan. That such broad distributions can also be the result of overconservative systematics has been shown for other lower invertebrates (e.g. Klautau et al. 1999), and sibling species in the sea could be much more common than previously thought (Knowlton 1993). Many cosmopolitan hydrozoan species could thus turn out to be complexes of several biological species.
The present publication therefore aims at revising taxonomic errors, at providing a discussion on how to distinguish closely similar species, and to show that some circumglobal species nevertheless show some regional differences. Only species for which new material was available are treated here.

Material and methods
The material used for this study is based on material of the Natural History Museum of Geneva (MHNG), which also includes recent collections made by the author in New Zealand, South Africa, and the Mediterranean. Some new material from Iceland collected by the BIOICE programme and not yet reported in my earlier publications (Schuchert 2000(Schuchert , 2001 was obtained from the Zoological Museum in Copenhagen. The majority of this material is now in possession of the Icelandic Museum of Natural History in Reykjavík. Material from the Faroes was obtained through the BIOFAR collection. The BIOFAR programme was a predecessor programme of BIOICE and intended to survey the benthic fauna of the Faroe Islands (Nørrevang et al. 1994). The hydroids of the BIOFAR programme are now held by the Kaldbak Marine Laboratory, The Faroes. Some slide material of the BIOFAR and BIOICE collections is also kept by the Natural History Museum of Geneva (MHNG). For more details on methods see Schuchert (1996Schuchert ( , 2001. Technical terms are explained in Millard (1975) and Cornelius (1995).

Description
Colonies erect, up to 10 cm (reportedly up to 30 cm), imperfectly pinnate to shrubby, some colonies perfectly planar. Stem sometimes covered by bark-like overgrowth of anastomosed stolons. Main stem and branches polysiphonic, thinning out to monosiphonic ends. Internodes roughly equal in length; nodes oblique, sometimes indistinct. Hydrotheca sitting at the end of segment like a prolongation of it, not or only slightly inclined, slightly overtopping distal node of segment. Hydrotheca very shallow, walls straight, diameter at base 0.18-0.25 mm, depth 20-40 mm, with desmocytes, without pseudodiaphragms, rim of hydrotheca often damaged. Secondary hydrothecae on short hydrophore, the latter usually shorter than 1.5 times the diameter. Male gonotheca club-to oblong sac-shaped, 1.4 mm long. Female gonotheca club-shaped, up to 2 mm long and 0.8 mm broad, straight, somewhat flattened, distal end rounded, opening at side at approximately middle of gonotheca on a distinct shoulder; opening partially covered by a characteristic hood-like process, two protruding polyps, 10-20 eggs per gonotheca. Nematocysts: two types of capsules, type I (9-10)6(3.5-4) mm; type II (5.5-7)6(1.5-2.5) mm.

Distribution
Circumpolar in the Arctic Sea, in the North Atlantic south to Cape Farewell in Greenland, Iceland and Faroe Islands. Type locality: Clark (1877) gives several localities in Alaska.

Remarks
The trophosome and male gonothecae of Halecium scutum, H. halecinum and H. beanii can resemble each other closely and it is difficult to identify immature or male specimens. In contrast, their female gonothecae are quite distinct and allow a relatively secure identification (compare Figures 2D, 4D and 5G). The female gonotheca of H. scutum has a characteristic lip or hood above the hydrothecae on its side wall ( Figure 2E). Unfortunately, the shape of the female gonotheca of H. scutum can occasionally be so variable as to make it resemble that of H. halecinum (Broch 1918, Figure 14). It is therefore not surprising that Broch (1910, as H. halecinum forma gigantea) considered H. scutum as an arctic form of H. halecinum. Later, Broch (1918) revised his opinion and kept both species distinct. Calder (1970) likewise regarded H. scutum as valid and distinct from H. halecinum.  synonymized H. scutum and H. beanii, a view later hesitatingly adopted by, for example, Cornelius (1975) and myself (Schuchert 2001).
Examination of material from the Faroes and Greenland, however, convinced me that H. scutum is clearly distinct from both H. beanii and H. halecinum. Although my material from the Faroes did not contain female gonothecae, there were many large infertile and male colonies that agreed well with female colonies from Greenland and with the description of Broch (1918).
Mature colonies of H. scutum are larger and considerably more robust than those of H. halecinum and especially H. beanii (compare Figures 1A-C, 3A-C and 5A-C). The best character, to distinguish H. scutum from H. halecinum and H. beanii, apart from the female gonothecae, is the relative shortness of the secondary hydrophores (Figure 2A-C). These secondary hydrophores are maximally as long as 1.5 times their diameter, but normally much shorter. In H. halecinum and H. beanii, they are longer, the ratio of length to diameter being in the range of 2 and more (compare Figures 2A-C, 4A-C and 5D, F). Halecium scutum also lacks the characteristic pinnate growth often found in H. halecinum and its hydrophores are in line with the main axis of the internode. In H. halecinum the hydrophores are inclined ( Figure 4A, B). In H. beanii, the diameter of the hydrotheca is distinctly smaller (0.12-0.16 versus 0.18-0.25 mm), the hydrotheca is usually held oblique relative to the internode axis, and the size of the type I nematocyst capsule is larger.
In the material from the Faroe Islands, one large colony of H. scutum was overgrown with several other hydroids, among them were female and male shoots identified as H. beanii. The differences in the microscopic structure were quite evident. This syntopic occurrence is further evidence for them being separate species. For further details see below under H. beanii.
Another peculiarity of most large colonies of H. scutum is a bark-like covering of the stem by a layer of tightly anastomosed stolons. A similar, but less anastomosed formation may occur in some larger colonies of H. halecinum.
Another European species that can be confused with the species discussed above is Halecium sessile Norman, 1867. According to Cornelius (1995), this species has very short secondary hydrothecae without distinct hydrophores. This trait may allow a reliable distinction.
Halecium boreale Lorenz, 1886 is likely a synonym of H. scutum. Lorenz (1886) described H. boreale as devoid of female gonothecae and the eggs are reported to develop in the gastric cavity of the feeding zooids (sic!). The males have normal gonothecae. Lorenz depicts a male colony that is indistinguishable from H. scutum. The feeding zooid of this colony contains spherical bodies. It seems to me that Lorenz misinterpreted food items, perhaps even eggs of another animal, as the eggs of his Halecium boreale. Although it is thinkable that there exist Halecium species without gonothecae, I am not aware of any other such species. Moreover, the presence of normal male gonothecae and the position of the eggs within the gastric cavity make Lorenz's assertion improbable.

Description
Colonies erect, stiff, up to 10 cm high, colony form often regularly pinnate, sometimes irregular or with several pinnate side-branches. Side-branches alternate, straight, evenly spaced, at constant angle of 40-60u to main axis, secondary branching possible. Stem and main branches polysiphonic, thinning out to terminal monosiphonic branches. Segments of similar length within a region of the colony, segment length much variable among colonies, nodes approximately transverse. Hydrotheca sitting on a shallow hydrophore which is oblique to main axis of internode, opening plane inclined. Hydrothecae alternate, walls straight, with desmocytes, hydrotheca diameter at base 0.12-0.14 mm, depth 30-50 mm. Secondary hydrothecae frequent, on a long hydrophore, the latter at least twice as long as diameter, secondary hydrophore near base usually with a kink. Male gonotheca clubshaped, length about 0.9 mm, aperture terminal. Female gonotheca paddle-shaped, tapering below, distally having a terminal aperture on one side and with right-angled ''shoulder'' on other side, aperture formed by a fused pair of hydrothecae that overtops gonotheca like a chimney, one or two polyps protruding from aperture, one to four eggs. Nematocysts: two types, type I (8-9.5)6(3.5-4) mm, type II (5-6)6(1.5-2) mm.

Distribution
East and West Atlantic, Mediterranean, South Africa, Pacific, Arctic Seas. Type locality: Whitstable, Kent, England.

Remarks
As discussed under H. scutum, this species can be difficult to distinguish from H. scutum and H. beanii. In waters around the British Isles and continental Europe, Halecium halecinum often forms quite regular, pinnate colonies ( Figure 3A, C). Cornelius (1995) considers this an important diagnostic trait of the species. In the material from the Faroes, the colonies were often quite irregular ( Figure 3B). The pinnate colony form, however, is also found in Faroese waters ( Figure 3C). Also some of the colonies examined from the English Channel were not so obviously pinnate. Hincks (1868, Plate 42, Figure a) depicted a colony originating from the British Isles which also deviates somewhat from the strictly pinnate form shown in the same plate. The colony form of H. halecinum is therefore not always regularly pinnate. But if present, the pinnate colony form is certainly a diagnostic trait of H. halecinum. Infertile or male colonies with atypical colony form can be impossible to identify if they are not part of a larger series that also includes fertile females. Broch (1918) thought that the shape of the secondary hydrothecae in H. halecinum differs from those of H. beanii in having an asymmetric adcauline wall. Such a difference was not found in the material examined here and neither was it mentioned by Cornelius (1995) (compare Figures 4A-C and 5D, F). Millard (1966Millard ( , 1975 also found that only the female gonotheca allows a reliable identification. Medel and Vervoort (2000) found that side-branches arise frontally or at the rear in H. halecinum, while in H. beanii they arise laterally.

Description of European material
Colonies 1-10 cm, irregularly branched, varied in appearance. Fertile colonies polysiphonic, thinning out to monosiphonic. Internode length of one colony relatively homogeneous, variable between colonies. Nodes alternately oblique. Hydrotheca sitting on a shallow hydrophore, the latter held oblique to main axis of internode, opening plane of hydrotheca inclined. Hydrothecae alternate, not adnate to internode, rim not overtopping level of distal node. Hydrotheca with straight walls, walls not much diverging, diameter at level of diaphragm 0.12-0.16 mm, depth 25-45 mm. Secondary hydrothecae frequent, their hydrophores long, length two to four times longer than diameter, often with basal kink. Female gonotheca 1-1.2 mm long, curved, approximately kidney-shaped, in middle of concave side two fused hydrothecae, in side-view held oblique, without hood-like cover, normally with two hydranths (may be reduced), convex side of gonotheca without longitudinal crests or crease-lines, four to six eggs per gonotheca, develop in situ into planulae. Male gonothecae club-shaped with terminal aperture, length up to 1 mm, breadth up to 0.25 mm. Nematocysts: two types of capsules, type I is a heteroneme (13-15)6(6-6.5) mm; type II (5.5-6.5)6(1.5-2) mm.

Characteristics of South African material
Like European material, but hydrothecae adnate; female gonotheca with polygonal crosssection and with five to six longitudinal crease-lines or shallow ridges, egg number 8-20. Male gonotheca also polygonal in cross-section, but more irregular, distal end often wedgeshaped, aperture terminal. Nematocysts: two types of capsules, larger one is a macrobasic eurytele (11-12)6(4-5) mm; smaller one as above.

Characteristics of New Zealand material
Like European material, but hydrothecae adnate and smaller, diameter 0.10 mm, depth 25-30 mm. Female gonothecae also identical to European population, with four to five embryos or eggs.

Distribution
Considered to be nearly cosmopolitan, found in mostly shallow to moderately deep waters (0-150 m). Type locality: near Scarborough, Yorkshire, England (Cornelius 1975).

Remarks
The examined fertile female colonies from the northern Atlantic and Mediterranean were all rather small and delicate (see Figure 5A-C). One sample (MHNG INVE 33593) was particularly confusing as it was growing on a large colony of H. scutum. There were female as well as male stems present, although in different regions. The host had therefore been colonized by several, independent colonies (genets). The gonothecae bearing shoots included typical, polysiphonic ones, but also many monosiphonic ones. These monosiphonic colonies were thus hardly distinguishable from H. lankesterii sensu Cornelius (1995), a species closely resembling H. beanii. The principal difference between H. lankesterii and H. beanii according to Cornelius (1995) are the monosiphonic stems in the former versus polysiphonic stems in the latter. Halecium lankesterii is, however, doubtless a valid species and comparison with material from Brittany showed that it is distinct from H. beanii. I suspect that at least some material identified as H. lankesterii by Hamond (1957), Vervoort (1959) and Cornelius (1995) was in fact monosiphonic H. beanii (see under H. lankesterii).
The microscopic structure of Halecium beanii is almost identical to that of H. halecinum (cf. Millard 1966Millard , 1975. Both species can only reliably be distinguished by comparing the female gonothecae. Halecium halecinum additionally forms often quite regularly pinnate colonies, which helps to distinguish it from H. beanii and H. scutum (compare Figures 1A-C, 3A-C and 5A-C).  synonymized Halecium scutum Clark, 1877 and H. beanii, a view later hesitatingly adopted by, for example, Cornelius (1975) and myself (Schuchert 2001). However, examination of the material described in this report convinced me that H. scutum is a distinct species. The female gonothecae of the two species allow a distinction. There are also significant differences in the colony size and details of the hydrotheca (see under H. scutum). The occurrence of H. beanii on a colony of H. scutum described above strongly argues in favour of them being distinct. Vervoort (1972) and Hirohito (1995) also thought that they are separate species.
The material of H. beanii from New Zealand described above was in no significant way different from European specimens, confirming the reports of Ralph (1958). However, part of Ralph's material allocated to H. beanii has recently been attributed to a new species, Halecium ralphae, by Watson and Vervoort (2001). The female gonothecae of this species is strongly curved, forming nearly a complete circle. Watson and Vervoort (2001) postulated also that H. beanii has female gonothecae with one hydrotheca only. This is evidently not the case, as also all here examined female gonothecae had a pair of fused hydrothecae.
The material from South Africa deviated in several aspects from European specimens. Especially the female gonothecae with their polygonal cross-section and the longitudinal ridges or crease-lines rendered it distinct. Also the increased egg number per gonotheca is a significant difference. Following the general practice in current hydrozoan systematics, all these differences would justify the postulation that the South African population belongs to a different species. Because the limited number of examined colonies does not allow the variability of the South African population to be assessed, no new name is proposed here.

Description
Colonies erect, stems unbranched or sparingly branched, up to 8 mm high and with up to 40 hydranths. Stems arise from creeping hydrorhiza. Stems always monosiphonic, irregularly segmented, regularly with intercalated segments without hydrotheca, these internodes short and with two marked bulges at their ends ( Figure 7C). Internodes generally short and stout, sides often undulated. Hydrotheca sitting at the end of segment like a prolongation of it, rim perpendicular to segment axis. One or two side-branches, if present, arising immediately below hydrotheca, angles quite acute ( Figure 7D). Hydrotheca with slightly diverging wall, these quite straight, not curved; diameter at level of diaphragm 0.15-0.17 mm, depth 30-45 mm. Secondary hydrothecae present, with hydrophore. Hydranths with 20-24 tentacles, body and tentacles containing zooxanthellae. Gonothecae arising from within primary hydrothecae. Female gonotheca pear-shaped, about 0.8-0.9 mm long, slightly compressed, opening on side, formed by a short, vertical, completely adnate tube with a median septum, opening without distinct hydrotheca, but rim with some desmocytes; opening plane perpendicular to long axis of gonotheca; two protruding hydranths present; about three eggs per gonotheca. Male gonotheca not observed; according to Babić (1913) oblong ellipsoid, 0.68-0.85 mm long, smooth, without hydranths.

Distribution
Mediterranean, tropical and subtropical eastern Atlantic, English Channel, North Sea, South Africa, Moçambique. Type localities: near Plymouth, England (details in Cornelius 1975).

Remarks
The type material of Bourne (1890) did not include gonothecae, a character usually indispensible for a correct identification of Halecium species. Two other traits, however, allow recognition of the species. This is for one the presence of zooxanthellae, for the other the segmentation pattern with its short, ahydrothecate segments with bulging ends (or segments constricted in the middle). Bourne's material had both characteristics. Babić (1913) and Bedot (1914) described fertile material. Later authors, like Vervoort (1959), Millard (1975) and Cornelius (1995), also included material lacking zooxanthellae and with smooth perisarc. The female gonothecae depicted by Hamond (1957), Vervoort (1959) and Cornelius (1975Cornelius ( , 1995 differ from the one depicted in Figure 7E and they appear identical to that of H. beanii ( Figure 5G). The egg number is also higher than usually found in H. lankesterii. These differences prompted Peñ a Cantero and García Carrascosa (2002) to postulate that the mentioned material of Cornelius and Vervoort actually belonged to a different, so far unnamed species. I concur with Peñ a Cantero and García Carrascosa (2002). The occurrence of fertile, monosiphonic stems of Halecium beanii (see under this species) led me to suspect that the material of Hamond (1957), Vervoort (1959) and Cornelius (1975Cornelius ( , 1995 was actually composed of monosiphonic colonies of H. beanii. Hamond (1957) acknowledged that H. beanii and H. lankesterii resemble each other very closely, but he insisted that they are distinct. He kept them apart on account of the sharply curved side-branches. His colony in Figure 9 resembles more H. lankesterii and it thus appears that his material contained indeed some H. lankesterii besides monosiphonic H. beanii. Halecium lankesterii also occurs in the Mediterranean and can be confounded with H. petrosum if no female gonothecae are present. The presence of zooxanthellae, the wider hydrothecae and the characteristic ahydrothecate segments of H. lankesterii help to distinguish them (compare Figures 7 and 8).
Halecium nanum Alder, 1859 resembles closely H. lankesterii as described above. Both have zooxanthellae in their tissue. The female gonotheca of H. nanum has two hydrothecae that reach to the distal end (Calder 1991), while they end roughly in the middle in H.
lankesterii. Their different ecology additionally suggests that they are likely separate species. Halecium nanum is usually epiphytic, mostly occurring on pelagic Sargassum sp. in warmer waters of the Atlantic. Cornelius (1995) thought that it might be wider spread and occur on a wider range of substrata.

Biology
Depth range 1-200 m, grows on a variety of solid substrata. Fertile colonies observed from June to October (Peñ a Cantero and García Carrascosa 2002). Ecological aspects were investigated by Coma et al. (1992).

Description
Colonies up to 10 cm, arborescent, irregularly branched, predominantly in one plane, stem and some branches polysiphonic. Internodes of unequal length within and between colonies, perisarc corrugated or smooth. Nodes alternately oblique. Hydrotheca on short hydrophore at distal end of internode like a prolongation it, hydrophore not delimited by node, rim of hydrotheca reaching just to level of distal node of segment, sometimes overtopping it. Hydrotheca short, wall distinctly recurved. Renovations frequent, secondary hydrotheca on hydrophore that is several times as long as depth of hydrotheca, walls usually corrugated. Primary hydrophore often with adcauline semi-circular perisarc thickening (pseudodiaphragm), sometimes at adcauline base of hydrophore a projecting perisarc fold ( Figure 9B-D). Hydranths with 20-24 tentacles. Gonothecae without hydranths, females ovoid, compressed by about a factor of two, with short pedicel, distal end with oval opening that may be on a shallow neck-like process. Male gonothecae similar but smaller.

Distribution
Arctic-boreal species, in the North Atlantic reaching south at least to Brittany and North Sea, perhaps even Spain and Azores (Cornelius 1995). Also Northern Pacific and Japan. Type localities: Northumberland coast, Shetland, Moray Firth, UK (Cornelius and Garfath 1980).

Remarks
Halecium labrosum has very variable internode lengths, so much so that this variability itself becomes a diagnostic character. The lengths are variable within and between colonies. In his key to the British thecate hydroids, Cornelius (1995) characterizes H. labrosum as having a wrinkled or corrugated perisarc. While some colonies indeed have such a corrugated perisarc, especially the secondary hydrophores, there were also regularly colonies with smooth perisarc among the examined material from the North Atlantic (compare Figures 9B, C and 10B). Broch (1918) considered Halecium labrosum to be an Arctic or northern Atlantic species. Likewise, Cornelius (1975) considered the English Channel as its probable southern limit. As already mentioned by Broch (1918), it is likely that at least some Mediterranean records of H. labrosum are due to a confusion with other Halecium species, notably H. mediterraneum. Halecium mediterraneum (see below) is by no means easily separable from H. labrosum and in fact might represent a southern form of the latter. The differences between the two forms are only gradual. It differs from H. labrosum in forming smaller, mostly monosiphonic shoots (but some are weakly polysiphonic!), the long hydrophore which makes the hydrotheca always overtop the distal node of the segment, the smaller diameter of the hydrotheca, and the smaller diameter of the segments (internodes).
The growth forms are also different, with H. mediterraneum tending to form bushy, tangled masses, while H. labrosum is always arborescent (compare Figures 10A, E and 11A).

Description
Colonies growing on rock, algae and hydroids, occasionally mass occurrence covering areas of square metres forming a lawn or tangled mass of stems. Stolons tubular, creeping, ramified. Individual shoots 1-3 cm in height, irregularly branched, usually monosiphonic, occasionally weakly polysiphonic, bushy, limp when out of the water, distal parts often geniculate, with regular succession of alternately inclined nodes, nodes distinct. Each internode with a distal hydrophore, oblique, alternately pointing left or right, long and distinctly surpassing level of distal node, hydrophore not delimited by a node. Length of internode 0.4-0.6 mm (mode 0.5 mm), diameter 0.10-0.14 mm, internode length quite homogeneous. Perisarc of internodes smooth, near nodes somewhat bulging, relatively thin. Primary hydrotheca gradually widening and rim somewhat everted or rolled, diameter at base 70-125 mm (mode 80 mm), depth 20-40 mm (mode 30 mm), ring of desmocytes present. Sometimes below diaphragm a semicircular thickening on adcauline side (pseudodiaphragm). Secondary or higher hydrothecae often present, length of their hydrophores as in primary one or longer, often corrugated. Ramification of stems originate from primary hydrophores or from below them; branching usually dichotomous, occasionally trichotomous. Hydranth with 18-20 tentacles. Gonothecae arise in upper axils of branching points and hydrophores; gonothecae dimorphic, the two sexes on separate shoots. Gonothecae of both sexes without protruding hydranths. Female gonothecae smooth, ovoid to rectangular, length 0.7-0.8 mm, breadth 0.5 mm, always compressed but degree variable, along sides a crease-line, distal end obtuse, with or without curved notchlike opening of variable breadth and depth ( Figure 11L, M), without distinct lateral ears. Mature female gonothecae have a thin perisarc capsule on inside ( Figure 11L). This secondary capsule envelops the gonangium, towards the opening its wall gets thicker. Gonangium oblong, surface epidermis with numerous nematocysts, without hydranth, 6-13 eggs embedded in tissue of ovoid shape, egg diameter about 0.1 mm. Development to planula takes place within gonotheca (larviparity). Male gonothecae smaller, ovoid, length up to 0.7 mm, compressed, sides with crease, distal end rounded, without notch or inner secondary capsule, opening slit-like. One oblong sperm mass. Living colonies have a characteristic yellow-brown colour; the pigment is extracted in formalin fixative.

Distribution
Mediterranean, perhaps also adjacent Atlantic Ocean from Morocco to Galicia. Type locality: Naples, Italy, Mediterranean.

Remarks
Upon closer examination of mature female gonothecae, I noted that there is a delicate tubelike inner capsule enveloping the gonangium. This secondary capsule attaches to the outer capsule along its distal opening ( Figure 11L). Such a secondary capsule was also found in H. delicatulum and in H. labrosum, although in the latter it is very thin and difficult to see. Ralph (1958) also found this inner capsule in H. delicatulum. Halecium mediterraneum is almost indistinguishable from Halecium delicatulum. Halecium delicatulum was first described by Coughtrey (1876) based on colonies found in Dunedin Harbour, New Zealand. Ralph (1958) re-described and revised it and she synonymized several similar nominal species. She also suspected that Halecium mediterraneum might be conspecific with it. Rees and Vervoort (1987) agreed and formally synonymized both names. Subsequent studies dealing with Mediterranean collections (e.g. Ramil and Vervoort 1992;Peñ a Cantero and García Carrascosa 2002) continued this usage. Vervoort and Watson (2003), however, did not synonymize it.
Halecium mediterraneum was initially described by Weismann (1883)  Comparison of H. mediterraneum from the Mediterranean and H. delicatulum from New Zealand showed that at least the microscopic structure of their trophosomes is very similar. There were differences in the shape of the female gonotheca (compare Figures 11 and 12). Ralph (1958) documented the variability of female gonothecae and her figures also show shapes as found here in H. mediterraneum. Therefore, the differences apparent in Figures 11  and 12 are perhaps not really representative. There remain only some slight differences in egg numbers per gonotheca, colony colour and colony form. Because of the slight differences, and mainly for biogeographic reasons, both nominal species are here kept separate, although it is acknowledged that they are very similar. Genetic methods might hopefully clarify whether they really belong to the same biological species.
Among the European Halecium species, H. mediterraneum is uncomfortably intermediate between Halecium labrosum and H. tenellum and it is not always easy to draw a dividing line. This may have led García Corrales et al. (1978) to synonymize H. tenellum and H. delicatulum. This proposal has been rejected already by other authors (e.g. Ramil and Vervort 1992) and also the present author agrees that they are distinct. Halecium tenellum is more gracile and forms smaller colonies, it is sparingly branched only, the internodes appear more elongate, and it never forms polysiphonic colonies. Fertile Halecium labrosum, in contradistinction, are always polysiphonic, their gonothecae are about twice as large and their shape differs slightly (compare Figures 10E, F and 11E-G). Halecium labrosum often also has a characteristically undulated perisarc (Cornelius 1995), but this is not a diagnostic feature as the perisarc can be entirely smooth. These differences are valid for European populations only. Halecium mediterraneum may occasionally have been misidentified as H. labrosum. Colonies growing epizoically on dead stems of Eudendrium sp. can feign a strong polysiphonic stem and are prone to be confounded.
Another species within this species cluster is Halecium textum Kramp, 1911 (see below). The main trait to distinguish H. textum from H. mediterraneum is its undulated or corrugated perisarc. There are also differences in its branching pattern (frequent trifid branching) and the more pointed gonothecae. The distributions are disjunct. Halecium textum is an arctic or northern boreal species. Although I am convinced that they are good species, the diagnosis of the limits of all four species remains difficult due to the absence of clearly apomorphic characters. The infertile material from the Arctic Sea described by Leloup (1960) is very unlikely to be H. delicatulum. Better guesses may be H. minutum Broch, 1903 or H. textum Kramp, 1911 (see below).

Description
Colonies reaching 2 cm in height, stems irregularly branched, polytomies frequent, branches not in one plane. Larger colonies form a tangled mass. Branches curving sharply upwards, with irregularly occurring nodes. Perisarc strongly undulated or corrugated over stretches, but smooth stretches can also occur. Hydrotheca on hydrophore of about equal length as hydrothecal depth. Hydrotheca diameter at base 0.1 mm, depth 30-50 mm, upper part distinctly everted or even rolled-in, diaphragm fine, desmocytes present. Below diaphragm frequently a pseudo-diaphragm, variable in form and position, either formed as a horizontal or oblique annular thickening or as a mere process at the origin of the apophysis. Gonothecae borne on branches, dioecious. Female gonotheca about 1 mm long, oblong, in side view only half as thick, narrow truncated end with oval aperture. Male gonotheca slightly smaller, spindle-shaped, not much flattened, shape somewhat variable, gonad fills gonotheca nearly completely (after Schuchert 2001).

Distribution
Boreal-arctic species; known from Greenland, north-eastern Canada, western and eastern Iceland, Faroes. Some records of H. undulatum from the British Isles, the North Sea, Sweden, and Norway could also be H. textum. Type locality: Maroussia, eastern Greenland, 160-180 m (Kramp 1911).

Remarks
This species has recently been re-described by Schuchert (2001) and several records previously attributed to H. tenellum or H. undulatum could be referred to it. The newly examined material from eastern Iceland, from where it has not been recorded so far, is especially notable for the strongly corrugated perisarc ( Figure 14A, B).

Description
Colonies erect, up to 8 cm high and 6 cm broad, much branched, branches strictly in one plane, branching pattern quite regular, stem and branches polysiphonic, thinning out to monosiphonic terminal regions. Monosipohonic parts segmented through weak nodes, these alternately oblique. Segments elongate, about 1 mm long, with a subterminal apophysis for the next segment and a hydrophore overtopping node of segment, at junction of apophysis and hydrophore a tongue-like process projecting into lumen of internode. Hydrotheca shallow, basal diameter 0.15-0.17 mm, depth 20-40 mm, walls everted, sometimes rim rolled-in, desmocytes present, opening plane of hydrotheca tilted, secondary hydrothecae frequent, first one with hydrophore (length up to twice the diameter), tertiary or higher hydrothecae with short hydrophore ( Figure 13C). Hydranth with about 22 tentacles. Gonothecae on stem and branches, only female ones could be identified, diameter 1.1 mm, shape resembling a Pecten shell (scallop) through nine or 10 radiating depressions and elevations, these elevations and depressions rounded and without crease-lines, rim of gonotheca slightly undulated.

Further data
Motz-Kossowska (1911) describes the colonies as hermaphroditic, with male and female gonothecae occurring on the same stem. Female gonothecae contain four to six eggs. Male gonothecae are smaller, either smooth or also with ribs. Diameter of hydrotheca base 0.12-0.13 mm, depth 30-50 mm. Side-branches originate immediately below hydrothecae, curved steeply upwards. Gonothecae on stems and sometimes stolons, dimorphic, the two sexes on separate stems. Female gonothecae ellipsoid, straight, length 0.7-0.85 mm, diameter 0.35 mm, strongly sculptured by up to 10 transverse ridges, chimney-like opening in middle on side, short, formed by two fused tubes, with median separation line, two hydranths protrude from this aperture, four to five eggs that develop in situ. Male gonothecae, without lateral opening, straight, spindleshaped, smaller than female, 0.6 mm long and 0.25 mm diameter, less corrugated than female gonotheca, variable within the same colony. Vegetative propagules unknown.

Distribution
New Zealand, North and South Island. Type locality: St Clair, Dunedin, New Zealand.

Remarks
The male gonothecae of this species have not been described so far. As for many other Halecium species, their morphology differs from the female ones. Patriti (1970) identified closely similar colonies from Morocco as H. corrugatissimum. The hydrothecae of her material were clearly everted and therefore her material was more likely Halecium pusillum (M. Sars, 1857), a well-known Mediterranean species. Halecium pusillum is very similar to H. corrugatissimum, but usually has everted hydrothecae (see Broch 1912 for re-description of the type material; Babić 1913 for figures of gonothecae). Motz-Kossowska (1911) attributed similar, but infertile, colonies to H. pusillum, despite the hydrothecae having straight walls. Stechow (1919) therefore considered this material to belong to a separate species which he named Halecium annulatum. Because this name was preoccupied by H. annulatum Torrey, 1902, Leloup (1938 changed it to Halecium stechowi. Peñ a Cantero and García Carrascosa (2002) also attributed the records of Motz-Kossowska (1911), hence H. stechowi, to H. pusillum. The presence of the characteristic vegetative propagules in the material of Motz-Kossowska (1911) makes this quite reasonable, but allowing also straight hydrothecae for H. pusillum narrows uncomfortably the gap to H. corrugatissimum.
Another nominal species in this complex is Halecium speciosum Nutting, 1901, a species originally described from Alaska (see also Fraser 1937). Considering their climatic and geographic separation, it appears unlikely to me that H. corrugatissimum, H. pusillum and H. speciosum belong to the same biological species.
These three nominal species are likely another example of the limitations we must be aware of when we attempt to discriminate biological species by morphological characters.