70 



P.R.PUGH 



The broad, but relatively short, mouth-plate consists of two 

 rounded processes whose basal margins are marked by basal exten- 

 sions of the meso-lateral ridges. These ridges peter out, without 

 apparently uniting, on the lower surface of the nectophore a short 

 distance above ostial level. The infra-lateral ridges branch from the 

 meso-laterals approximately on a level with the mid-length of the 

 nectosac. The meso-laterals then run obliquely up the sides of the 

 nectophore to join the other ridges, approximately on a level with the 

 top of the nectosac. The infra-laterals form the ventral margins of the 

 ventro-lateral wings up to a level just above the top of the nectosac. 

 They then bend through 90° and run up to join the apico- and meso- 

 laterals (Figure 17C). The thickened ventro-lateral wings are well 

 developed and enclose a deep groove, which at its deepest occupies 

 half the height of the nectophore (Figure 17C). They are roundly 

 truncate apically at about four-fifths the length of the nectophore. 



The nectosac is a long tube, with only a slight apical emargination, 

 that occupies most of the main body of the nectophore, and has no 

 obvious dorso- ventral undulations. It is distinctly undercut adaxially 

 and is presumed to have a muscle-free zone in that region, although 

 this could not be established with certainty. The ostial opening is very 

 large and is only slightly directed towards the upper surface. The 

 pedicular canal (Figure 1 7B) typically only gives rise to the dorsal and 

 ventral radial canals. The course of all the canals is straight. 



Typically, the youngest nectophores show the absence of a central 

 thrust block (Figure 18 A), but with a clearly defined ridge pattern. A 

 slightly larger nectophore shows the gradual development of the 

 thrust block and the ventro-lateral wings (Figure 18B). 



Remarks concerning the identification of Bargmannia gigas. 

 B. gigas is known only from the nectophores of what is presumed to 

 be a single specimen, collected in the equatorial Atlantic at a depth 

 of 1510-2000m. The nectophores easily can be distinguished from 

 other Bargmannia sp. by their incredible size and the distinctive 

 pattern of ridges. 



Etymology. The specific name gigas refers to the giant size of 

 the nectophores. 



DISCUSSION 



As was noted in the Introduction, the content of the genus Barg- 

 mannia is debated. Although Totton (1965) included it in the family 

 Pyrostephidae, some of the characters that he listed in his diagnosis 

 of that family apply exclusively to the scope accorded to the genus 

 Pyrostephos, which is monotypic for P. vanhoeffeni. In particular, 

 these are the looping of the lateral radial canal on the nectosac of the 

 nectophore, and the three to four marked bends of the dorsal canal. 

 In Bargmannia all the canals are held to be straight, or only slightly 

 sinuous. Other characters, such as the number of tentilla on the 

 tentacle, and the structure of the bracts and gastrozooids probably 

 are more specific than familial. However, in both genera the 

 nectosome is long but again this is not a good familial character. 



At first glance, the nectophores of Pyrostephos vanhoeffeni (see 

 Totton, 1965, Figure 41) and Bargmannia species look strikingly 

 different. However, there are several similarities. Specimens of P. 

 vanhoeffeni have been collected recently by SCUBA divers (G.R. 

 Harbison, personal donation) and by net (Pages, Pugh & Gili, 1994). 

 It is apparent from these that the mature nectophores can vary greatly 

 in size; ranging from 8x5 mm (length x width) in the SCUBA 

 collected material (Figure 19B)to 15 x 1 8 mm, respectively, in the net 

 collected specimens (Figure 19C). Such large size ranges of the 

 mature nectophores of phy sonect species have not often been observed, 



although such is so in Nanomia bijuga (delle Chiaje, 1841) (Pugh, 

 pers. obs.). It is also known to be so in at least two Bargmannia spp. In 

 B. amoena (Figure 19 A), the size variation of mature nectophores is 

 even greater than that of P. vanhoeffeni, ranging from c. 6 x 3 mm 

 (length x width) to 25.5 x 1 2.5 mm respectively. Although the general 

 shape of mature Bargmannia nectophores does not change with size, 

 it appears that that of P. vanhoeffeni does. In the smaller specimens of 

 the latter (Figure 19B) there is a large triangular thrust block, 

 reminiscent of that on mature Bargmannia nectophores. However, in 

 the larger, preserved specimens (Figure 1 9C) the thrust block is folded 

 upwards producing a deep transverse furrow on the dorsal surface, just 

 basal to it. Neither P. vanhoeffeni nor Bargmannia spp. have large 

 apico-lateral processes. 



The nectophores of both genera have the same basic ridge pattern; 

 comprising apico-, infra, and vertical (meso-) laterals, but no lateral 

 ridges. In addition, in both, the apico-laterals divide into two branches 

 close to the ostium. The inner branch ('frontal ridge') of the larger 

 nectophores of Pyrostephos vanhoeffeni (see Totton, 1965, Figure 

 40) is relatively short, in comparison with Bargmannia spp., and 

 directed only toward the mid-line. However, the present material, 

 particularly that of the smaller specimens, shows that these ridges 

 can curve round basally and continue for a short distance towards the 

 ostium before petering out. Nonetheless, the species of these two 

 genera are not the only physonects to show this basic pattern of 

 ridges. It is also found on the nectophores of two others namely, 

 Frillagalma vityazi Daniel, 1966 (see Pugh, 1998) and Erenna 

 richardi Bedot, 1904 (PR.Pugh, personal observation). In addition, 

 an even simpler arrangement, in which the vertical lateral ridges are 

 absent, is found in two Marrus species, namely M. antarcticus 

 Totton, 1954 and M. orthocanna Kramp (1942). For these, the 

 branching of the apico-lateral ridges is weak and difficult to discern. 

 A third species, namely M. orthocannoides, that Totton (1954) 

 include in the latter genus probably does not belong there as its 

 nectophores do not have an adaxial muscle-free zone on the nectosac. 



Species referred to both Bargmannia and Pyrostephos have an 

 adaxial zone on the nectosac of the nectophore that is muscle-free and 

 deeply embayed. In addition, the lateral radial canals arise separately 

 from the dorsal canal.These appear to be important characteristics. Of 

 the other species previously mentioned Marrus antarcticus and M. 

 orthocanna show all of these characters. However, in Frillagalma 

 vityazi, there is no deeply embayed, muscle-free adaxial zone; al- 

 though the lateral radial canals do arise separately from the dorsal one, 

 albeit very close to the point of insertion of the pedicular canal (Pugh, 

 1998). Further, this species has many marked differences from the 

 others under consideration and need not be considered further in this 

 discussion. Erenna richardi does have a muscle-free zone, but it lies at 

 the apex of the nectosac, which is not deeply embayed adaxially. Thus, 

 from the basic arrangement of the ridges and nectosac, the nectophores 

 of Bargmannia, Pyrostephos and Marrus species are very similar. 

 Another common feature is that they all have relatively short pedicular 

 and relatively long, ascending pallial canals. But how do their 

 siphosomal elements compare? 



Most siphonophores are believed to be hermaphrodite (mono- 

 ecious), bearing both male and female gonophores. However, 

 specimens of Physalia, the Portuguese Man O'War, and probably all 

 other cystonect siphonophores, are single sexed (dioecious). It 

 should be noted that Mackie, Pugh & Purcell (1987, p. 1 00) used the 

 terms monoecious and dioecius erroneously. In physonect 

 siphonophores, species of the benthic family Rhodaliidae appear to 

 be dioecious (Pugh, 1983), as are Marrus antarcticus, Pyrostephos 

 vanhoeffeni (Totton, 1965), and from the present study Bargmannia 

 spp. According to Andersen (1981) M. orthocanna is monoecious, 

 but the male gonophores he illustrated were only minute, bud-like 



