Stephanomia amphytridis Lesueur & Petit, 1807

Pugh, P. R. & Baxter, E. J., 2014, A review of the physonect siphonophore genera Halistemma (Family Agalmatidae) and Stephanomia (Family Stephanomiidae), Zootaxa 3897 (1), pp. 1-111 : 83-102

Stephanomia amphytridis Lesueur & Petit, 1807

Stephanomia amphytridis Lesueur & Petit, 1807 , pl. 29, fig. 5; Lamarck, 1816, p. 462; 1840, p. 25; Bedot, 1896, p. 382; Totton, 1954, p. 46; Bonnemains & Carré, 1991, fig. 1 (5); Dunn et al., 2005a, p. 2005; Dunn & Wagner, 2008, p. 752; Lindsay, 2006, p. 105; Pugh, 2006, p. 42; Burton & Lundsten, 2008, p. 30 (fig.); Cartwright et al., 2008, p.4 Lindsay & Miyake, 2009, p. 419; Pugh & Gasca, 2009, p. 399.

Stephanomia amphitritis Tilesius, 1814, p. 181 ; Cuvier, 1817, p. 65; 1830, p. 288; 1843, p. 206; Eschscholtz, 1825, p. 744; 1829, p. 155; Lesson, 1830, p. 62; Griffith & Pidgeon, 1834, p. 492; Oken, 1825, p. 744; 1835, p. 197; Leuckart, 1854, p 321; L. Agassiz, 1862, p. 368; Claus, 1878, p. 272.

? Stephanomia levigata Quoy & Gaimard, 1825, p. 585 , Pl. 86, fig. 2.

Stephanomia amphitrides Blainville, 1830, p. 108 ; 1834, p. 119.

Stephanomia amphitridis Lesson, 1843, p. 476 , Pl. 10, fig. 1; 1841, p. 197; Leuckart, 1861, p. 197; Schneider, 1898, p. 118; Totton, 1936, p. 236; Leloup, 1955, p. 13.

Stephanomia Amphitritidis Leuckart, 1854, p. 341 ; 1875, p. 455.

Stephanomia amphitrites ; Haeckel, 1869, p. 45; 1888a, p. 40; 1888b, p. 184, p. 221; Claus, 1872, p. 190.

Cupulita amphitrites, Bedot, 1896, p. 408 .

Halistemma Amphitrites Claus, 1876, p. 238

"? Halistemma amphytridis " Totton, 1965, p. 60 ; Daniel, 1974, p. 47, text—fig. 3, H–J;

Halistemma amphytridis Daniel, 1984, p. 75 , fig. 16, a–b; Bonnemains & Carré, 1991, fig. 12; Carré & Carré, 1995, p. 565; Haddock & Case, 1999, p. 575; Hunt & Lindsay, 1999, p. 80, figure 5A; Poupin et al., 1999, p. 20.

Halistemma amphitridis Daniel, 1973, p. 867 ; Ortman et al, 2010, p. 2150; Bucklin et al, 2010, p. 2238.

Halistemma amphytridens Stepanjants, 1967, p. 127 ; 1977a, p. 55.

Halistemma aff. amphytridis Lindsay & Hunt, 2005, p. 506 .

? Stephanomia sp. Lens & van Riemsdijk, 1908, p. 84, Pl. XV, figs. 113–114.

non Stephanomia Amphitritis Chamisso & Eysenhardt, 1821 , Pl. XXXII, fig. 5 = Agalma okeni .

non Stephanomia amphitridis Huxley, 1859, p. 72 , Pl. 6; Bigelow, 1911, p. 287, pl. 18, figs. 1–8; Kawamura, 1954, p. 110 = Halistemma foliacea .

non Stephanomia nereidum Haeckel, 1888a, p. 40, 1888b , p. 221 = nomen nudum.

non Halistemma amphytridis Mapstone, 2004, p. 231 , figs. 1–3; Araujo, 2006, p. 49 = Halistemma foliacea .

The family Stephanomiidae fam. nov., monotypic for the species Stephanomia amphytridis , belongs to the dioecious clade of physonect siphonophores (Dunn et al., 2005). That clade is characterised by the following characters:

Each colony is of a single sex; although there appear to be exceptions to this in the genera Marrus and Bargmannia ;

The nectophores have only an ascending mantle canal;

The nectophores, with the exception of the family the family Rhodaliidae , possess a muscle-free zone at the proximal end of the nectosac; usually more prominent on the lower side;

The lateral radial canals on the nectosac are usually without major loops; with the major exception being Stephanomia amphytridis ;

The nectophores are budded off on the ventral side of the nectosome.

Other characters, such as the ridge pattern on the nectophore, the structure of the tentillum, and the presence or absence of palpons, are often of familial or generic significance.

Diagnosis: Stephanomia amphytridis is dioecious; its nectophores are budded off on the ventral side of the stem, and the mature ones have an extensive muscle-free zone on the proximal, lower side of the nectosac. Only an ascending mantle canal is present. The sinuous course of the lateral radial canals is very characteristic; as is the ridge pattern that, uniquely, includes a divided vertical lateral ridge complex.

The presence of palpons, and a tentillum, with a loosely coiled cnidoband, without involucrum, and a single terminal filament are characters that help to separate this genus from certain other dioecious genera.

Material examined:

JSL II Dive 1457-DS8 3 rd September 1987 24°00.7'N 82°16.6'W Depth 829 m. *

JSL II Dive 1685 - CG5 10 th October 1988 26°27.8'N 77°58.9'W Depth 846 m. *

JSL I Dive 2138-DS2 2 nd August 1987 39°56.7'N 70°14.9'W Depth 777 m. *

JSL I Dive 2644 - CG2 9 th November 1989 26°02.0'N 77°32.0'W Depth 869 m. *

JSL II Dive 2667-DS8 17 th September 1995 c. 27°N c. 85°W Depth unknown

JSL II Dive 2677-DS4 22 nd September 1995 c. 27°N c. 84°W Depth unknown

JSL I Dive 3272-DS1 No Station data available

Tiburon Dive 746-D4 6th October 2004 35°28.8'N 123°51.9'W Depth 1161 m.† Tiburon Dive 1074- SS9 23rd January 2007 35°50.1’N 122°40.1’W Depth 1158 m.*† Tiburon Dive 1110-SS12 1st August 2007 35°23.0'N 122°40.0'W Depth 1401 m. †

* Nectalia stage . † in situ frame grabs.

Thanks to the kindness of Susan Svonthun, from the Video Laboratory at MBARI, a brief video of the specimen of Stephanomia amphytridis collected at 1401 m during Tiburon Dive 1110 will be placed on the website www.youtube.com/user/MBARIvideo.

Description: Two photographs of a large specimen, part of which was collected during Tiburon Dive 746, are shown in Figure 94 . The specimen was probably c. 5 metres long and was very robust, with a rigid siphosome and over 25 large nectophores in the nectosome.

Pneumatophore: The pneumatophore was featureless and bore no signs of pigmentation. In its preserved state it measured c. 7.6 mm in length and 3.8 mm in diameter. It was borne on a very muscular stalk that clearly could relax and extend itself considerably so that the pneumatophore would appear well above the nectosome.

Nectosome: The very large nectophores were budded off on the ventral side of the nectosome and, usually, there were numerous nectophoral buds at its anterior end.

Nectophores: The basic ridge pattern on the nectophores was very similar to that of Halistemma species , and consisted of pairs of upper and lower laterals, laterals and what will be referred to as the "vertical lateral ridge complex".

At the bud stage ( Figure 95A ) these ridges were very distinct and all were completely separate from each other. The upper laterals had a distinct notch in them. The nectosac filled almost the entire nectophore, with the axial wings being undeveloped and the thrust block very small. The mantle and pedicular canals were very prominent. The lateral ostial processes were extremely well-developed, forming hemispherical structures comprised of numerous darkly-staining ectodermal cells that surrounded a central core of 40–50 nematocysts ( Figure 95B ). These nematocysts measured c. 260 x 50 µm and possibly were the same as those found at the tips of the bracts (see below).

A very young nectophore, measuring 6.5 mm in all directions, is shown in Figure 96 . The narrow axial wings were closely apposed to each other, and the thrust block remained as a small process on a mesogloeal swelling. The upper and lower lateral ridges merged at the proximal end of the nectophore. The latter ran down the lower side, at first close together, but then diverging, before petering out at about half the height of the nectophore ( Figure 96B ). The upper laterals also remained together for most of their course but, slightly above the ostial level, they curved outwards through 90° and continued laterally until they joined the lateral ridges. The lateral ridges branched from the upper laterals above the apex of the nectosac and were first directed obliquely downwards before curving through 90° and running parallel to the nectosac. After the upper laterals had joined them they continued to end just above the much reduced lateral ostial processes.

What we will refer to as the vertical lateral ridge complex appeared to become more complex as the nectophores developed. In fact, for the very small nectophore shown in Figure 96 , that on the right side of the nectophore differed from that of the left. On the left side ( Figure 96C ) there were two entirely separate ridges, neither of which connected with any other ridge. The more proximal one lay in the lower half of the nectophore, while the more distal one was longer and extended from a short distance away from the upper lateral obliquely down to end almost below the ostium. On the right side ( Figure 96D ) the two ridges had become joined together and each bore an additional side branch.

The canal system was very characteristic. There was an ascending mantle canal extending down from just below the apex of the thrust block along a mesogloeal protuberance, until it gave rise to the pedicular canal. This ran directly to the lower proximal side of the nectosac where it gave rise directly to all four radial canals. The upper and lower canals had a straight course to the ostial ring canal, while the laterals first ran upwards and slightly proximally to the lateral sides of the nectosac. There they formed a broad loop upwards and then downwards, to the lower side of the nectosac, before looping back up to its mid-height and then curving and running directly to the ostial ring canal. At this stage of development there was no sign of a muscle-free zone on the nectosac.

The mouth plate was a thickened structure surrounding the basal half of the ostium. The lateral ostial processes were circular areas with a stalk-like area connecting them to the ostium. They no longer bore nematocysts, just darkly staining ectodermal cells that possibly were sites of bioluminescence. Large, totally transparent cells were present in a single row below the point of insertion of the lateral radial canals onto the ostial ring canal. There was also, on each side, a long patch of ectodermal cells, which was easily detached, on the lateral sides of the nectophore in its proximal half.

A medium-sized nectophore is shown in Figure 97 . The ridge pattern was now clearly defined with the upper laterals uniting with the lower laterals at the proximal end of the axial wings. The upper laterals remained widely apart in the proximal half of the nectophore, and then began to approach each other rapidly, before turning toward the ostium. From then on they were considerably narrowed and much less obvious. The lower laterals extended along the lower lateral side of the nectophore to end at the base of the mouth plate. The vertical lateral and lateral ridges arose separately, but quite close together, from the upper lateral ridges. The laterals were first directed obliquely downwards and then curved to run distad to the thickened sides of the mouth plate. The vertical lateral ridge complex was now more clearly defined and consisted of an oblique ridge extending down from the upper lateral ridge to join a longitudinal ridge on the lower lateral side of the nectophore. During its course it gave off two separate branches.

The course of the radial canals was almost the same as that described above. The long ascending mantle canal ran over the surface of a pronounced mesogloeal thickening before giving rise to the pedicular canal. The latter did not give rise directly to all four radial canals, as the upper canal arose a short distance along the right lateral canal. By this stage the nectosac clearly had a muscle-free zone on its lower side in its proximal half, although this was not apparent in lateral view. The mouth plate was well developed and thickened laterally, with a pronounced median indentation. The giant ganglion could clearly could be seen traversing the mouth plate and extending to the lower side of the ostium.

The fully mature nectophores ( Figure 98 ) measured, in their preserved state, up to 55 mm in length and 45 mm in width, thereby exceeding the largest previously known nectophores belonging to Bargmannia gigas Pugh (see Pugh, 1999b). They possessed well-defined, proximal truncated, axial wings, and a large central thrust block with a shallow, but broad central emargination. On the lower side of the nectophore the thrust block was hollowed out centrally and laterally thickened and, at its base, were two roundly triangular mesogloeal swellings, between which ran the mantle canal. This canal was quite short and soon gave rise to the pedicular canal that connected with the radial canals on the nectosac in the middle of the muscle-free zone on the lower side of the nectosac. There were slight, but not consistent, variations with regard to the origin of each radial canal. The courses of the upper and lower canals remained straight, and the lateral radial canals had the same looping course as described above. The mouth plate was still well developed and slightly emarginated distally in the mid-line,. It was thickened with mesogloea, particularly at the lateral margins.

The arrangement of the upper and lower lateral ridges remained much the same as described above, but the lateral and vertical lateral complex became further developed. The main feature was that the two ridges became totally, or so it appears, separated from the upper lateral ridges, but ran parallel with the latter for a short distance. There were indications of a further continuation, proximally, of the ridges but they never were seen to unite with the upper laterals. The vertical lateral complex appeared to break up into a number of separate ridges that had a very variable arrangement even, as noted above, on the different sides of the same nectophore. In general, a proximal vertical lateral ridge extended down, to a variable extent, from the lateral ridge, often curving distad toward its lower end. There was a separate ridge distal to this that again was very variable, but often took on a Yshaped appearance ( Figure 98 (left), on right-hand side of nectophore). On some occasions there was a complete loop, while on others the lower distal extension of the ridge shown in Figure 97C was associated with it. There was usually a further vertical, usually short, ridge distal to this, and again of varying design. It is probable that the ridge pattern could be better seen in fresh material, as shrinkage gives rise to a very complex pattern of ridges and excrescences in the preserved material.

Siphosome: In life the siphosome, like the nectosome, appeared to be almost rigid but the in situ pictures of the specimen from Tiburon Dive 746 (cf. Figures 94 (lower) and 99) showed that the stem could undergo a modicum of relaxation, as the gastrozooids were more spread out in the latter than in the former. The gastrozooids always appeared as discrete entities as did the clusters of palpons, of which up to four appeared to be present in each cormidium. However, it was not clear where the gonodendron was is attached in each cormidium. The foliaceous bracts thickly covered the siphosome, particularly dorsally and laterally.

Bracts: There were two basic types of bract, both of which were large and quite solid, occurring in enantiomorphic pairs. The Type A bracts were far more abundant than the other type, which probably means that they formed paired lateral rows.

Type A —The mature bracts of this type were extremely asymmetrical, extending proximally only on one side ( Figure 100 ). They measured up to 6.5 cm in both length and width. Typically on the foreshortened side of the bract there was a lateral tooth but, in the younger bracts particularly, another tooth could be present on the opposite side. The proximal tip of the bract was often swollen with mesogloea on the upper side, and distally on the same side there was a median mesogloeal thickening. On the bracts, particularly the younger ones, there were occasional small patches of ectodermal cells, either just one or a pair on either side of the mid-line.

The bracteal canal arose on the upper side of the bract but soon traversed onto the lower side and continued straight toward the distal tip. The attachment lamella would have been present for about half its length. The canal remained in contact with the lower surface of the bract until close to the distal tip. In the younger bracts the canal often was greatly expanded in this region and, in the youngest, the proximal regions also were inflated. The canal then ran obliquely through the mesogloea to end below a small cupulate process at the distal end of the mesogloeal thickening. In the youngest bracts it appeared that nematocysts were present in the cupule but, in general, it was just an empty hollow.

Type B —These bracts were more symmetrical and slightly smaller, measuring up to 5 cm in length and 4.5 mm in width ( Figure 101 ). They were quite variable in their form; some being almost rounded with a swollen proximal region, while others bore lateral or distal teeth in varying numbers. About 40% of the bracts bore no teeth, and a slightly lower number had just one lateral tooth. The remainder were made up of bracts with 2 lateral teeth, or 1 or two distal teeth, or two lateral and one distal, or two lateral and two distal. The degree to which these teeth were developed also varied considerably. The bracteal canal again arose proximally on the upper side and, particularly in the younger bracts, could have a variable thickness. The region of attachment extended to about a third of the length of the bract and in this region the canal ran along the near surface of a narrow mesogloeal ridge. As in the Type A bracts the canal continued in close contact with the lower surface of the bract until close to the distal tip, where it similarly ran obliquely through the mesogloea to end below a somewhat large cupulate process. Instead of this process containing nematocysts, in this case it formed the socket for a large spherical ball packed full of nematocysts ( Figure 102A ). However, apart from the very youngest bracts, this ball of nematocysts had become detached.

The hundreds of nematocysts in this distal ball were all of one type, which measured in their undischarged state c. 210 µm in length and 55 µm in diameter ( Figure 102B ). A bract was found ( Figure 102C ) where some of these nematocysts had discharged into it. These showed a short shaft that was obviously grooved, but no spines could be discerned. Once again we are uncertain as to the identity of these nematocysts.

Gastrozooid and tentacle: The preserved gastrozooids were relatively large measuring about 3 cm in length with the basigaster occupying the proximal third. However, such sizes are meaningless as they only represent the contracted size in preservation The tentacles were extremely long bearing numerous large tentilla.

Tentilla: Mature preserved tentilla ( Figure 103A ) consisted of a relatively short pedicle, a long, loosely coiled cnidoband and a long terminal filament. The cnidoband was made up of a broad band of nematocysts and a double elastic band attached firmly to its distal end. When fully stretched out the cnidoband extended to c. 7.5–8 cm in length. Although there was a slight swelling at the junction between the pedicle and the cnidoband no distinct involucrum was apparent.

However, on the younger tentilla ( Figure 103 B,C ) attached to the tentacle of the first adult gastrozooid of the Nectalia post-larval stage, the preserved material ( Figure 103B , arrowed) gave the impression that an incomplete horse-shoe shaped involucrum was present. However, this was probably a preservation artefact as a photograph of a similar tentillum from a living specimen ( Figure 103C , arrowed) showed just a small swelling. For these younger tentilla the cnidoband was more obviously spirally wound, with 7–9 whorls.

On the cnidoband there were two types of nematocyst, with one type occurring in a wide range of sizes. These ( Figure 104A ) measured from c. 140 to 195 µm in length and from c. 15 to 32 µm in diameter. They were by far the more numerous nematocyst and were arranged in numerous (30+) rows across the cnidoband. There was an indication that they gradually decreased in size from the sides toward the middle. Unfortunately no discharged nematocysts of this type were found and so it was not possible to determine their exact type.

The second type of nematocyst ( Figure 104B ) were of about the same length as the others (c. 175 µm), but were considerably broader (c. 50 µm in diameter). They were probably stenoteles although, again, no discharged ones were found. These nematocysts occupied the lateral margins of the cnidoband, but not in a regular fashion. There were none at the very proximal end of the cnidoband, but after a short distance distally they began to form a tightly packed single row on each side. Distally their number began to increase so that they formed two alternating rows. At about a third of the length of the cnidoband their numbers began to decrease, so that again a single row was present distally. The distance between the individual nematocysts then gradually increased so that at the distal end of the cnidoband these nematocysts were infrequently found.

There were also two types of nematocyst found on the terminal filament. The first type ( Figure 104C, D ) were arranged in three alternating rows along one side of the filament, while the second type ( Figure 104E ) were far more numerous and covered the remaining surface. The first type measured, in the undischarged state, c. 80 µm in length and 45 µm in diameter. Quite unusually, when compressed these nematocysts tended to discharge and evaginated a bottle-like structure with a long, narrow distal process, without any obvious spines and apparently open at its tip. It is not clear what type of nematocyst this represents, and no one on the Cnidaria Network has responded to our request for information. Apart from the apparent absence of spines, microbasic euryteles appear to be the most closely related, although ours have a very short tubule. Although Werner (1965) suggests that such nematocysts are absent from siphonophores, more recently Carré (1969) has found them to be present on the tentilla of Rosacea villafrancae, a calycophoran, and Carré (1980) on the tentilla of Agalma elegans , but neither look anything like the present ones. They might also be strange acrophores, which are usually found on the tentilla of physonects, but they are usually small and their distal tip is closed, while that of the present material appeared to be open.

The second, more abundant, type of nematocyst on the terminal filament ( Figure 104E ) measured c. 62 µm in length and 20 µm in diameter. No discharged nematocysts of this type were found and so, again, we cannot be sure as to which type of nematocyst they belong. As noted above, acrophores along with desmonemes are usually found on the terminal filaments of the tentilla of physonect siphonophores, but these did not resemble either of those types.

Palpons: The palpons ( Figure 105A ) were simple bag-like structures up to c. 25 mm in length. Usually there was a long, narrow proximal section, an inflated middle part, and a nozzle-shaped distal proboscis terminating in an opening. No nematocysts were found anywhere on its surface. At the proximal end of the narrowed basal region a long semi-annulated palpacle was attached. Nematocysts were scattered randomly along one, somewhat flattened side ( Figure 105B ). These nematocysts ( Figure 105C ) measured c. 30 µm in length and 14 µm in diameter and, from their appearance, were probably haplonemes, although none were found discharged. They resembled the second type of nematocyst ( Figure 104B ) found on the terminal filament of the tentillum, but they were only half the length and relatively broader.

Gonophores: Each colony is dioecious and, as for Apolemia species , it is difficult to find mature gonophores as, usually, it is only the anterior end of these large colonies that is collected, where all the gonophores are immature. Each cormidium possessed a single gonodendron, with potentially hundreds of gonophores branching from a single gonostyle. A female gonodendron and some, probably, immature female gonophores are shown in Figure 106 . The gonophores had a broad, proximally tapering, pedicle and a large bell in which the four lateral radial canals could clearly be seen. They measure up to 4.5 mm in length and c. 2.75 mm in diameter, but that probably misrepresented their mature size. Several clear oil droplets were found in the subumbrella cavity.

The male gonodendra ( Figure 107 ) probably included hundreds of male gonophores, but the only ones found were very young and hardly developed. The oldest of these gonophores were borne on a long narrow pedicle, about as long as the bell itself. The latter measured c. 2.1 mm in length and 1 mm in diameter, but they almost certainly grew to a much greater size when mature.

Nectalia - stage: Of all the specimens of Stephanomia amphytridis examined for this study five were at the Nectalia post-larval stage, and two of these are shown in Figure 108 .

Larval bract: The fully formed larval bracts ( Figure 109 ) were elongate with a distinct distal extension, measured up to 60 mm in length, and usually numbered two or three. These were surrounded anteriorly by up to four smaller bracts at various stages in development ( Figure 110 ). On the upper side of the bract there was a median longitudinal ridge that ran almost the entire length of the mature bract. Proximally, a pair of lateral ridges arose from it and demarcated the lateral margins of the upper side of the bract. Proximal to them the thickness of the bract decreased sharply so that the digitate mesogloeal extension lay on the lower side. In the younger bracts these lateral ridges varied considerably in the degree to which they extended distally, and in the youngest they merely curved out laterally. On the mature bracts they extended down to about the mid-point and then petered out. Their places were taken by a separate pair of lateral ridges that, at first, ran parallel with the other pair, but then continued to the distal tip of the bract and joined with the median longitudinal ridge.

The lower side of the youngest bracts ( Figure 110 , top left) was quite flat apart from in the region of attachment where a distinct central keel was formed, with the bracteal canal following its contours. With age the lower side developed a pair of lateral flaps that arose separately at the proximal end but joined together, distally in the mid-line. Shortly before uniting, for the mature bracts, each gave rise to a pronounced tooth; although they were not present on the more developed of the younger bracts ( Figure 109 , right). A similar pair of flaps was present proximally. When fully developed the flaps enclosed a deep cavity on the inner side of the bract. From the junction point of the lateral flaps, which was raised above the lower wall of the more distal part of the bract, there arose a median ridge that also ran to the distal tip, so that the distal extension of the bract had four regular facets.

On the upper surface of the bracts there were numerous small, rounded spots (not illustrated) that were distributed randomly throughout. There were also larger patches of ectodermal cells that usually, in the younger bracts, formed long slightly curved structures distal to the origin of the lateral ridges; rather like the pair of patches on the nectophores. On the older bracts small patches could occur, particularly on the distal extension, but they were easily detached and their presence difficult to detect.

The canal arose on the lower side of the bract at the same level as the junction between the median and lateral ridges on the upper side. In the region of attachment it was thickened and, as noted above, in the younger bracts ran over a distinct keel. However, in the adult bracts that keel had all but disappeared. After a short distance the canal turned abruptly upwards and penetrated into the mesogloea. In this region the walls of the canal may have had small outward processes. The canal then bent through another right angle and proceeded toward the distal end of the bract. In the distal extension the canal might have had a very wavy course. Just before it terminated, particularly in the younger bracts, it slightly increased its diameter. It ended below a small concavity in the distal tip of the bract. In the mature bracts this concavity was usually empty, but in the younger ones it was found to contain a number of large ectodermal cells along with, in the central region, up to ten large nematocysts, measuring c. 133 x 34 µm.

Larval tentillum: The larval tentillum of Stephanomia amphytridis ( Figures 111 , 112 ) was quite different from that of Halistemma species. It was borne on a long, narrow pedicel and consisted of a large featureless sac, a cnidoband and a short "terminal" filament. At an early stage of development the sac, about 2.5 mm in length and with a maximum width of 1.8 mm, was laterally flattened and somewhat heart-shaped, with a shallow median indentation distally ( Figure 112A ). The cnidoband, total length 2.5 mm, was positioned on one side with a portion, with a reduced compliment of nematocysts, extending distally and from which arose the "terminal filament". In life ( Figure 111 ) the cnidoband bore orange pigmentation, apart from the "distal" portion, which was colourless. When fully developed ( Figure 112B,C ) the cnidoband extended from the distal end of the more inflated sac. In actuality, the cnidoband was attached to the sac by a broad tube that ran up one side of the cnidoband to the portion with a reduced nematocyst compliment ( Figure 113 , arrowed). The tube, much convoluted, could be traced down one side of the sac and appeared to arise slightly proximal to the sac's mid height. The nematocysts on the cnidoband were positioned so that they abutted this tube. This must mean that for them to come into effect the "proximal" end of the cnidoband must break free (see Figure 112B ) and, presumably by means of hydrostatic pressure, the cnidoband would be flipped through 180° in order to bring the nematocyst batteries into contact with a prey item. The "terminal" filament would then lie between the sac and the cnidoband.

Two types of nematocyst were found in the cnidoband ( Figure 114A,B ). About 20 pairs of the larger ones ( Figure 114A ), measuring c. 132 x 45 µm, were situated on either side of the cnidoband at its "proximal" end, relative to the unactivated state of the cnidoband. Unfortunately, no discharged nematocysts of any type were found and so their true identity remains unknown. However, in most physonect cnidobands these few, large nematocysts are usually heteronemes ( Purcell, 1984) and as the present ones do not have the appearance of stenoteles it is presumed that they may be mastigophores.

The second type ( Figure 114B ), measured c. 72 x 15 µm, and were present in great numbers. They filled most of the remainder of the cnidoband on one side. The number of rows of them increased from both the distal and proximal ends. In the "distal" portion, from which the "terminal" filament arose, they were largely confined to the outer walls, with no central concentration, as in the remainder of the cnidoband. These nematocysts are presumed to be haplonemes that, according to Werner (1965), are most likely to be monotrichous anisorhizas.

Three types of nematocyst were found on the "terminal" filament ( Figure 114C ). The largest of these, measuring c. 34 x 17 µm, were ovoid capsules containing, along with the invaginated tubule, 4 or 5 spherical inclusions. The tubule itself occupied less than half the volume of the capsule. The middle sized nematocysts, measuring c. 30 x 15 µm, were also ovoid and the highly coiled tubule filled the capsule. The smallest, measuring c. 19 x 12 µm, were almost spherical and again the tubule filled most of the capsule. In many physonect species the terminal filament of the adult-type tentillum usually contains two types of nematocysts, acrophores and desmonemes. But without any discharged nematocysts to look at it is futile to speculate on the identity of these three types. It was difficult to discern the arrangement of the nematocysts in the "terminal" filament (see insert to Figure 113 ), but there was some indication that the middle sized ones formed outside pairs, with the larger and then smaller ones arranged between them.

Remarks: It is hoped that the present description firmly establishes the species Stephanomia amphytridis as a valid species although, of course, we can never be absolutely sure that it is the same as that originally figured by Lesueur & Petit (1807). It is regrettable that Totton did not publish a description of the species, as that would then have avoided some of the confusion that has more recently arisen. We have already established that the three main re-descriptions attributed to this species, i.e. Huxley (1859), Bigelow (1911) and Kawamura (1954), together with Mapstone (2004) can now all be referred to the species Halistemma foliacea . After Totton's (1965) doubtful attribution of the species to the genus Halistemma, Daniel (1974) described and illustrated two nectophore and a single bract from the Indian Ocean. The nectophores measured 28 and 33 mm in length and 25 mm in width, but clearly were in poor condition, as no ridge pattern could be discerned.

Daniel (1985, p. 76) reproduced her earlier figures under the name Halistemma amphytridis , but she remarked "The nectophore described below differs a great deal from the typical " Halistemma " type of nectophore and warrants a creation of new genus and new species, but in this account it is still retained as H. amphytridis (= Stephanomia amphytridis )". Unfortunately she did not say what these distinguishing features were, and one can only presume that it was based on the arrangement of the lateral radial canals. However, her figure did appear to indicate the presence of only an ascending mantle canal, but Daniel drew no attention to this significant feature.

We have already remarked in the Introduction (see Figure 10 ) that the species Sphaeragalma rotunda described by Margulis (1976) most certainly belongs to Stephanomia amphytridis as the pattern of the ridges and the lateral radial canals certainly shows, as well as the presence of only an ascending mantle canal, although the nectophores were clearly very young as they only measured 1.1–1.2 cm in length.

Distribution: The locality where the original specimen figured by Lesueur & Petit (1807) was collected remains uncertain; the report indicating only that it was collected somewhere between Le Havre and Mauritius. As noted above, the four main re-descriptions of specimens attributed to this species by Huxley, (1859), Bigelow (1911), Kawamura (1954) and Mapstone (2004) in actuality all refer to the species Halistemma foliacea .

The first records for specimens that are now attributed to this species appear to come from Totton (1936), who found several specimens in the Beebe Collections from the south-east of Bermuda, and he also listed several specimens of Nectalia loligo , which may well be the post larval stages of that species. Later, Leloup (1955) recorded a specimen from the Michael Sars Expedition at c. 35°N, 48°W. There are several more recent records from the North Atlantic, e.g. Haddock & Case (1999) recorded a specimen from the Bahamas; Dunn (2005) from 40°18.6'N 68°07.8'W at a depth of 800m; Bucklin et al. (2010) and Ortman et al. (2010) from 33°38.5'N 69°47.7'W and 14°05.8'N 54°46.8'W, with a further record from the Gulf of Maine (Pagès, personal communication). Several specimens have also been collected by the JSL submersibles in the vicinity of the Bahamas and the Dry Tortugas (see Table 1). In addition, specimens have been collected at several Discovery station in the North Atlantic Ocean, together with further specimens just south of the equator (see Table 2).

Tiburon 1150 4-Nov-2007 36°36.61'N 122°09.52'W 1103 m Nectalia

Tiburon 0260 27-Feb-2001 36°34.20'N 122°31.20'W 1192 m

Tiburon 1110 1-Aug-2007 36°22.47'N 122°40.07'W 1401 m

Tiburon 0962 6-Apr-2005 36°19.78'N 122°53.97'W 1372 m Nectalia

Doc Ricketts 0026 27-May-2009 36°06.98'N 122°45.20'W 1993 m

Doc Ricketts 0105 13-Dec-2009 36°04.07'N 122°17.92'W 1983 m

Doc Ricketts 0101 10-Dec-2009 36°04.14'N 122°17.89'W 1319 m

Doc Ricketts 0030 20-May-2009 35°56.66'N 123°57.35'W 1227 m Nectalia

Doc Ricketts 0030 20-May-2009 35°56.51'N 123°57.20'W 1434 m

Doc Ricketts 0555 21-Nov-2013 36°04.12'N 122°17.84'W 1168 m

Tiburon 0949 2-Feb-2006 35°45.66'N 122°42.16'W 1304 m

Tiburon 0949 2-Feb-2006 35°45.66'N 122°42.16'W 1130 m Nectalia

Tiburon 1074 23-Jan-2007 35°50.22'N 122°40.07'W 1178 m Nectalia

Doc Ricketts 0029 29-May-2009 35°29.79'N 125°00.53'W 1066 m Nectalia

JSL II 2666 17-Sep-1995 c. 27°0.0'N c. 85°0.0'W Depth unknown

JSL II 2677 22-Sep-1995 c. 27°0.0'N c. 85°0.0'W Depth unknown

Doc Ricketts 0370 31-Mar-2012 26°44.99'N 111°09.71'W 1388 m Nectalia

JSL I 2648 12-Nov-1989 26°30.9'N 78°17.8'W 899 m

JSL I 2888 19-Dec-1990 26°22.8'N 78°46.3'W 761 m.

Tiburon 1040 1-Oct-2006 24°17.41'N 124°03.12'W 1368 m

In the Indian Ocean Daniel (1973, 1974, 1985) found occasional well dispersed specimens, from off South Africa, the Arabian Sea, and off the west coast of Australia. The specimen of Sphaeragalma rotunda ( Margulis, 1976) was collected in the southern Bay of Bengal, east of Sri Lanka (07°39'N, 87°54').

In the Pacific Ocean, several further specimens have been observed from submersibles in the vicinity of Monterey Bay (see Table 1), and Burton & Lundsten (2008) picture a specimen close to the sea bed on Davidson Seamount, 120 km SW of Monterey, California, at a depth of 1304m. Hunt & Lindsay (1999), Lindsay (2006) and Lindsay & Miyake (2009) have recorded specimens from the North West Pacific in Japanese waters, but there appears to be no records from the South Pacific Ocean.

Figures

FIGURE 10. Upper (left), lower (centre) and lateral (right) views of Paragalma birsteini (top) and Sphaeragalma rotunda (bottom). Redrawn, with a few alterations, from Margulis (1976) figures 1–6. Scale bar 5 mm.

FIGURE 94. Stephanomia amphytridis. In situ photographs of specimen from Tiburon Dive 746. The length of the "detritus" sampler in view, including the lids, is 12 inches. (Photographs courtesy of Dr. S.H.D. Haddock, © MBARI.)

FIGURE 95. Stephanomia amphytridis. A. Lateral view of nectophoral bud from Tiburon Dive 746 specimen. lop. lateral ostial process; rl, rul, and rvl lateral, upper lateral, and vertical lateral ridges, respectively; n. notch. Scale bar 2 mm. B. Nematocyst from lateral ostial process. Scale bar 100 µm

FIGURE 96. Stephanomia amphytridis. A. Upper, B. lower and C. left and D. right lateral views of small nectophore From JSLI Dive 2138. Scale bar 2 mm.

FIGURE 97. Upper (A), lower (B), and lateral (C) views of medium-sized nectophore of Stephanomia amphytridis from T1074-SS9 specimen. Scale bar 5 mm.

FIGURE 98. Stephanomia amphytridis. Upper (left) and lower (right) views of mature nectophore. Scale bar 1 cm.

FIGURE 100. Stephanomia amphytridis. Type A bracts. A. Mature, and B. young bracts. Proximal at top, distal at bottom. Scale bars A.10 & B. 5 mm.

FIGURE 101. Stephanomia amphytridis. A. and B. Two forms of mature Type B bracts. Proximal at top, distal at bottom. Scale bar 1 cm.

FIGURE 102. Stephanomia amphytridis. A. Detail of nematocyst cluster at distal tip of young Type B bract. Scale bar 1 mm. B. Undischarged and C. discharged nematocysts from Type B bract. Scale bars 100 µm.

FIGURE 103. Stephanomia amphytridis. A. Mature tentillum and B, C young tentilla from the tentacle of the first adult gastrozooid of the Nectalia post-larval stage. (C. © Stefan Siebert). For B & C proximal at top, distal at bottom. Scale bars 5 mm.

FIGURE 104. Stephanomia amphytridis. Nematocysts from the cnidoband (A, B, scale bar 100 µm) and terminal filament (C–E, scale bars 50 µm) of the tentillum.

FIGURE 105. Stephanomia amphytridis. A. Palpon and palpacle. Scale bar 5 mm. B. Portion of palpacle showing attached nematocysts. Scale bar 100 µm. C. Nematocyst. Scale bar 10 µm.

FIGURE 106. Stephanomia amphytridis. Part of female gonodendron (Upper), and some individual female gonophores (Lower). Scale bars 5 and 1 mm, respectively.

FIGURE 107. Stephanomia amphytridis. Part of a male gonodendron from T1110-SS12 specimen, with (inset) individual male gonophore. Scale bars 5 and 1 mm, respectively.

FIGURE 108. Stephanomia amphytridis Nectalia-stages from JSL Dive 1685 (left) and Tiburon Dive 1074 (right). Scale bar 1 cm.

FIGURE 109. Upper (left) and lower (right) views of a larval bract of Stephanomia amphytridis from JSL I Dive 2138 specimen. Proximal at top, distal at bottom. Scale bar 5 mm.

FIGURE 110. Upper views of young larval bracts of Stephanomia amphytridis from JSL I Dive 2138 specimen. Proximal at top, distal at bottom. Scale bar 5 mm.

FIGURE 111. Nectalia stage of Stephanomia amphytridis from Tiburon Dive 1074. Detail of gastrozooid and larval tentilla. Scale bar 2.5 mm.

FIGURE 112. Larval tentilla of Nectalia post-larval stage Stephanomia amphytridis from Tiburon Dive 1074. A. younger tentilla; B. intermediate tentilla; C. Older tentilla. Scale bars 1 mm.

FIGURE 113. Cnidoband and terminal filament of older tentillum of Stephanomia amphytridis (scale bar 500 µm), with detail (top right) of nematocysts on terminal filament, and (bottom left) of large nematocysts at proximal end of cnidoband. Scale bars 100 µm.

FIGURE 114. Stephanomia amphytridis Nectalia-stage. Types of nematocyst found on larval tentilla. A. at proximal end of cnidoband (scale bar100 µm); B. throughout cnidoband (scale bar 100 µm); and C. the three types found on the terminal filament (scale bar 25 µm).