On the Austral-Antarctic stenothoids Proboloides, Metopoides, Torometopa and Scaphodactylus (Crustacea Amphipoda) Part 2: the genus Proboloides, with description of two new genera and the transfer of two nominal species to Metopoides

Abstract This is the second part of a revision of the most plesiomorphic genera in the amphipod family Stenothoidae sensu lato (see Krapp-Schickel and Koenemann 2006 for an overview and Krapp-Schickel 2008 for the first part). 41 species not belonging to Metopoides were plotted in a matrix using the same 61 characters as in the first part. The resulting group of Proboloides species (most probably not existing in the Austral-Antarctic region) is discussed, a key for the members given and two new genera erected. Some species described as Proboloides are redescribed and 2 species transferred to Metopoides. A key for all actual members of. The remaining species, i.e. those actually being in the genera Torometopa and Scaphodactylus, will be dealt with in the final part of this series, together with a key to all of them.


Introduction
list 23 species belonging to the genera Metopoides and Proboloides, which were always diffi cult to diff erenciate. Since this publication the number of species has increased, while our knowledge on character states did not grow the same way. Th e results of a phylogenetic analysis of the entire family Stenothoidae sensu lato by Krapp-Schickel and Koenemann (2006) showed not only that the genera treated therein had many plesiomorphic character states, but also that too many characters were still unknown or poorly described. Th us before further studies of the phylogenetic relationships, several species required redescription or at least checking of new characters not described so far.

Material and methods
As many species as possible of this group were studied and redescribed, in order to replace the (initially numerous) question marks in the start-up matrix. Species were borrowed from diff erent Museums.
It is remarkable that all mouthparts are always unusually long and narrow and it could be imagined that they all together function as a sucking device (it is said that the name steno-thoids stems from the narrow mouthparts, stenos meaning narrow in Greek). Mxp has reduced plates and Md has more or less reduced molars and palps, while pars incisivus and lacinia mobilis are very well developed with acute and robust "teeth"; also both maxillae have robust setae, which could help to divide the food parts already cut by the mandible.
In many stenothoids P3 4 are longer but weaker than P5-7, and all are kept parallel to the coxae and never twisted. Interesting are the quite often acutely lengthened meri (in P3, 4 anteriorly, in P5-7 posteriorly) which warrant an additional capacity against fallling off the substrate.

Genus Proboloides Della Valle
Della Valle, 1893: 907 Type species. Metopa gregaria Sars, 1882: 93, t. 4, fi g. 6 Proboloides mainly occurs in the Atlantic, but nominal Proboloides species have been reported also from the Pacifi c, Indian and Antarctic oceans. Its species are often found living in deep waters and show a clear sexual dimorphism, usually their gnathopods are quite diff erent in size and shape, often with a strongly incised Gn2 palm, with palmar corner well defi ned in females, but not defi ned in males, and robust peraeopods.
Diagnostic characters. A1 peduncle art 1 usually short, length < 3× width, subequal to cephalon; A1 usually shorter than 2/3 body length, A1 accessory fl agellum lacking. Md palp with a very short or lacking art3, poorly setose; Mx1 palp 2 arts; Mx2 inner plate ordinary; Mxp inner plates well separated, outer plates usually reduced (less than 0.2 of merus length). Ratio Cx2:Cx1 > 3. Cx2 length equal or more than 1.5 × the width. Gn1, 2 diff erent in size and shape; Gn1 small, almost simple, rarely subchelate; carpus length equal to propodus; length of propodus Gn1 about half or less than half length of propodus Gn2; Gn2 palm has serrations or teeth, usually no incisions; Gn2 propodus is in males often, in females always smaller than Cx2; carpus shorter than wide, merus elongate. P5 basis linear, without posterodistal lobe; merus anterior margin shorter than 1.25 length of propodus anterior margin. P6, 7 basis expanded and lobate, merus tip reaching half to full length of carpus. Ep3 with acute posterodistal corner. U1 peduncle is longer than longer ramus. T length is shorter to equal the double width, triangular, laminar.
Proboloides. Md palp may have a shortened third article, the inner plates of Mxp may be fused. Antennae robust with 0-1 articulate acc. fl ag.; gnathopods with sexual dimorphism, Gn1 much smaller than Gn2; coxal plates enlarged; basis P6, 7 with strongly lengthened and widened merus.
Torometopa. Mouthparts ordinary. Antennae with 0-2-articulate acc. fl ag.; gnathopods without or with sexual dimorphism; coxal plates small or large; basis P5 rectolinear with posterodistal lobe lengthened and widened to varying degrees; P6, 7 merus weakly to strongly lengthened and widened. In short, characters of Metopoides and Proboloides together, but P5 basis with posterodistal lobe, which might have evolved independently. Th us this genus was the least convincing one.
To fi ll the gaps with question marks in the fi rst matrix (see also Krapp-Schickel 2009), I studied the following species in detail:

1) Proboloides porcellanus KH Barnard, 1932: 111-112, fi g. 61
Th is species has a small posterodistal lobe on P5 basis (like Torometopa and diff erent from all other Proboloides), but P6, 7 bases are not rounded (like in all Torometopa), but narrowing distad. Gn1, 2 show a striking similarity with those of Mesometopa sinuata Shoemaker, 1964 (female) from the American west coast, but P5-7 in that species are totally diff erent, and the other Mesometopa species are Pacifi c arctic-boreal (southernmost locality is S. California).
Etymology. Th e type species was collected from the pharynx of a large ascidian in the Falkland Islands, in Spanish Islas Malvinas.
Remarks. Th is genus has a posterodistal lobe on P5 basis like the members of Torometopa; it is diff erent from all other known stenothoid genera by the rectangularly widened, distally narrowing basis of P6,7. In Metopella and Mesoproboloides P5 is rectolinear without a posterodistal lobe, P6,7 are diff erently widened; in Hardametopa P5-7 all have a slender basis. Th e genera Metopelloides, Stenothoides, Vonimetopa and Zaikometopa diff er in having a 1-articulate Md palp, while the palp is absent in Parametopella.
Redescription of type material: Body smooth. Eyes rounded, large.
Coxae. Cx2 with rounded anterior margin, straight behind, angle rounded without tooth; Cx3 trapezoid-shaped, distally widening, Cx4 not excavated, anterior and posterior margin rounded, about as long as wide. Gnathopods. Gn1, 2 propodi similar in shape, diff erent in size. Gn1 dactylus ordinary; propodus with parallel margins, palm not defi ned, about twice as long as wide; carpus longer than propodus, subtriangular, longer than wide, proximally wider than distally; merus incipiently chelate; all articles densely beset with setae. Gn2 length of propodus more than 2/3 of Cx2; propodus subelliptical, twice the size of propodus Gn1; hind margin subequal to length of palm which has shallow incisions, palmar corner well defi ned by small tooth-shaped prolongation but no U-shaped incision. Dactylus same length as palm. Gn2 carpus shorter than wide, cup-shaped, merus not lobate.
Peraeopods. P4 merus anterodistal margin somewhat lengthened. P5 dactylus half length of slim propodus; merus posterodistal margin not reaching half of carpus length, basis rectolinear with short posterodistal lobe. P6 basis hind margin straight, with posterodistal lobe similar to P5, merus posterodistal corner acutely lengthened and somewhat widened, not reaching to half of carpus length. P7 basis proximally widened with lobe, distad narrowing with small posterodistal lobe, hind margin crenulate and excavated; merus lengthened and widened, reaching about half carpus length.
Uropods. U1 peduncle with many short robust setae, nearly twice as long as subequal rami; U2 peduncle also beset with many small robust setae, longer than longer ramus, rami somewhat unequal; U3 totally unarmed, peduncle longer than ramus, art 1 of ramus longer than art 2.
Telson. Not reaching end of peduncle U3; less than twice as long as wide; distally tongue-shaped rounded, naked.
Ecology. From pharynx of a large ascidian.
Remarks. M. porcellanus has extremely shortened A1, 2, no Mxp OP, a stout Gn1 and aberrant P6, 7: is this an adaptation to the life inside the pharynx of ascidians, where they certainly cannot swim but only crawl? All we know is that space there is at a premium.

2) Proboloides typicus (Walker)
Also this species is very sparsely described and fi gured. I found material at the Australian Museum Sydney, and compared it with one specimen deposited at the Verona Museum by Bellan-Santini. Both fi t the written description by Schellenberg well, and this species clearly belongs in Metopoides:
Mouthparts. Md incisor and raker spine row well developed; no clear molar cusp; palp with 3 arts, art 3 about 1/3-1/2 length of art 2, with 3 distal long setae (Walker: Md palp lacking third art, therefore creating a new genus Proboliella; but Schellenberg already noticed 1926: 323 fi g. 41, that there is a well-developed third article). Mx 1 IP with 1 distal seta, OP with 6 strong robust setae, palp with 2 arts; Mx 2 inner plate ordinary, shorter than outer; Mxp IP not fused, 2/3 length of ischium; OP narrow, well developed, reaching more than half of merus length; dactylus long, subequal to propodus.
Coxae. Cx2 with rounded anterior margin straight behind, angle rounded with small tooth; Cx3 narrow with parallel margins, Cx4 not excavated, inferior and posterior margin rounded, about as long as wide.
Gnathopods. Gn1, 2 propodi diff erent in size and shape. Gn1 dactylus ordinary; propodus with parallel margins, palm well defi ned (corner about 120°), somewhat longer than half length of propodus, about twice as long as wide; carpus shorter than propodus, triangular, longer than wide, merus incipiently chelate. Gn2 length of propodus more than 2/3 of basis in male, less in female; propodus subelliptical, twice the size of propodus Gn1; hind margin half length of palm which is in male and female with incisions, palmar corner well defi ned by acute tooth-shaped prolongation and U-shaped incision. Dactylus clearly shorter than palm, probably working together with robust setae of palmar corner. Gn2 carpus shorter than wide, cup-shaped, merus not lobate.
Ecology. Steeply sloping rock bottom with encrusted bryozoans and hydroids. Remarks: As this species clearly has an accessory fl agellum (although tiny), unspecialized gnathopod propodi and neither much lengthened nor widened merus on P5-7, it has to be placed in the genus Metopoides, and even is a very "typical" representative of that genus.

3) Proboloides stephenseni Ruff o, 1949
Shortly after the war the possibility to check foreign literature was restricted. As only few characters were illustrated, the character states included in the matrix most probably were not always appropriate.
Coxae. Cx2 with rounded anterior margin, straight posterior one, angle rounded with small tooth; Cx3 with parallel margins, Cx4 not excavated, anterior and posterior margin rounded, about as long as wide.
Gnathopods. Gn1, 2 propodi similar in shape, diff erent in size. Gn1 dactylus ordinary; propodus triangular, palm well defi ned, about twice as long as wide, about as long as hind margin; carpus shorter than propodus, trapezoid, longer than wide, with parallel margins; merus with very short distal free margin. Gn2 length of propodus = Cx2; propodus more than twice the size of propodus Gn1; hind margin much shorter than length of palm which has shallow incisions and crenulations, palmar corner well defi ned by small tooth-shaped prolongation but no U-shaped incision. Dactylus shorter than palm. Gn2 carpus shorter than wide, cup-shaped, merus not lobate.
Telson not reaching end of peduncle U3; less than twice as long as wide; distally pointed, marginally two robust setae.
Sexual diff erences. Females unknown. Distribution. Antarctica, 70°23'S, 82°47'W (?P. stephenseni Ruff o, 1949). South Africa (P. rotunda Stebbing, 1917 After Barnard and Karaman (1991: 696) the genus Proboloides has its distribution in the Atlantic Ocean, S-Africa and the Antarctica. Th erefore it was important to check also non-Antarctic species.  (2) Remarks. Although the shape of the gnathopods (especially the simple Gn1) creates doubt if it could not belong in Metopa or Stenula, checking of the mouthparts showed at least that this species has a Md palp with 3 arts and a palp of Mx1 with ? 2 arts (although the articulation is not clear, see Bousfi eld 1973: 98). It is diff erent from the other Atlantic members, but for the time being it should remain in the genus Proboloides.

4) ? Proboloides holmesi
At the Verona Museum I found a tiny specimen called "Metopa sp."(1,5 mm) which is extremely similar to Proboloides holmesi, except the rounded palmar corner (see Fig. 6, 7 and compare to Bousfi eld 1973 fi g. 16 (2)); also here there seems to be a fi ne indistinct line in Mx1 palp. I am adding the illustration also to stress the fact how small the diff erences between the genera are, and to show that also the genus Metopa has to be included in this basic group of stenothoid genera.

5) Stenothoe aequicornis Stephensen, 1931
During a stay at the Copenhagen Museum I checked Stephensen's type material of this species, as Barnard and Karaman (1991: 698) remarked "gnathopod 1 wrong, mouthparts unknown". And they were right: there is no doubt about a clearly developed 3-ar-ticulated Md palp and therefore this species cannot be a member of Stenothoe, where there is no Md palp at all. Less clear is the structure of the palp of Mx1: as often in other specimens, the articulation is not easily seen (Fig. 8, 9). But this character should not be the only one deciding if a species belongs to Metopa or to Proboloides, and the shape of gnathopods brings this material into the vicinity of the latter.
Coxae. Cx2 with rounded anterior margin, straight or even somewhat concave behind, front angle rounded without tooth; Cx3 with trapezoid-shaped margins, Cx4 not excavated, anterior and posterior margin rounded, wider than long.
Gnathopods. Gn1, 2 propodi diff erent in size and shape. Gn1 dactylus ordinary; propodus elongate, about 3× as long as wide, palm well defi ned, much shorter than hind margin; carpus longer than propodus, triangular, nearly 3× longer than wide, with parallel margins; merus with very long distal free margin. Gn2 length of propodus > Cx2; propodus about 3× the size of propodus Gn1; hind margin much shorter than length of palm which has shallow incisions and crenulations, palmar corner scarcely defi ned by group of robust setae, no U-shaped incision. Dactylus subequal to palm. Gn2 carpus very short, cup-shaped, merus acutely lobate.
Peraeopods. P5 dactylus > half length of slim propodus; merus posterodistal margin not reaching end of carpus length, basis rectolinear, width proximally and distally subequal, posterodistally rounded, but not lobed. P6 basis hind margin straight, merus posterodistal corner acutely lengthened and widened, reaching to end of carpus length; P7 similar to P6, but basis hind margin regularly rounded.
Telson. Not reaching end of peduncle U3; less than twice as long as wide; distally pointed, marginally 3 robust setae.
Sexual diff erences. Females unknown. Distribution. Between Faroes and Iceland, 375 m depth. At the Verona Museum I looked for the only species of the genus Torometopa cited in Barnard and Karaman (1991), where there is a question mark behind the genus name: Figure 9. Proboloides aequicornis (Stephensen, 1931): as above, photographs of the material taken with Olympus BX51 with cell imaging software. Ledoyer, 1986: 966, fi g. 381, 382 A Torometopa? armata Barnard & Karaman, 1991: 700 It is the unique (female) type specimen from Îles Glorieuses N of Madagascar, from 3718 m depth.

6) Proboloides armata
Unfortunately I could not examine the slide and confi rm the drawings of the subrectangularly widened Gn1 propodus and the 1-articulated Md palp, both very unusual characters in our treated group, as this type must be deposited elsewhere.
But at the Victoria Museum Melbourne I found a species from the Bass Strait from 770 m clearly belonging also to this basic species-complex of stenothoids, having a posterodistal lobe on P5 basis. To my big surprise it turned out that this species too had a 1-articulate Md palp.
As this character-combination does not fi t any of the extant stenothoid genera, a new one was erected: Figure 10. Proboloides aequicornis (Stephensen, 1931): as above, photographs of the material taken with Olympus BX51 with cell imaging software.
Mouthparts. Mdb palp one long article, on tip a fi ne articulation-line visible, marginally no setation, distally 1 long and 1 shorter seta. Mxp outer plate reduced.
Peraeopods. In all dactylus clearly longer than propodus. P3, 4 merus long, falcate curved, nearly twice the length of propodus. P5 basis distally somewhat widened but strongly lengthened to lobe maximal to minimal breadth 1.4-1.6; merus also nearly twice the length of propodus, posterodistal tip neither lengthened nor widened. P6 basis hind margin with straight margins, merus anterior and posterior margin subequal, distally not lengthened, reaching no carpus. P7 basis wider than in P6, but posterior margin also rather straight.

Cladistic analysis
A matrix of 38 species and 61 characters was built (Fig. 13): all presently known species in the genera Proboloides, Torometopa and Scaphodactylus were included. A hypothetical Gammarus species was chosen as out-group (see also Krapp-Schickel 2009, Fig. 5 without the species of Metopoides).
Th e programs MacClade 4.06 (Maddison and Maddison 2003) and PAUP 40B.10 (Swoff ord 2002) were applied. Using 38 taxa and 61 characters a heuristic analysis with a hypothetical Gammarus as an outgroup -species was performed and the majority rule consensus tree of 28 trees illustrated in Fig. 14. Th e diff erence between the trees concerned only the arrangement within the groupings.

Results
At the beginning of the present analysis 16 species were cited for Proboloides: Eight species were reported from the Atlantic or Arctic Ocean: P. calcaratus, clypeatus, grandimanus, gregarius, holmesi, schokalskii, schuleikini, zubovi. In most species at least the females have a pronounced palmar corner in Gn2, while Gn1 is weak and slender. In the deep-water species P. calcaratus and P. gregarius the males have Gn2 propodus + dactylus very much lengthened and the eyes large, Cx3 has no parallel margins, but becomes wider distally and is not much narrower than Cx 4, while P. holmesi has a narrow Cx3 with parallel margins, very diff erent from Cx4. P. clypeatus must remain a species dubia, as it is too poorly described. Th e species P. gregarius and schuleikini (originally only subspecies of P. gregarius) show diff erences only in the even more elongated Gn1 in the latter, and I think P. schuleikini is a big female of P. gregarius. I have also my doubts about the description of P. grandimana (Bonnier), where all details match P. gregarius except the big and triangular Cx1 which should be even larger than Cx2, an extremely unusual character in stenothoids; it seems quite probable that this is an error and that Cx2 is repeated. -Branch et al. (1991) illustrate a Proboloides sp. with U3 with 2 rami, which undoubtedly is also an error of the drawing.
Th us the only certain Atlantic-Arctic members are P. calcaratus, gregarius, schokalskii and zubovi.
Th ere are 5 nominal Proboloides species from S-Africa and the Antarctic-Subantarctic region: P. porcellanus, rotundus, stephenseni, typicamimus, typicus. Th e species P. stephenseni and P. rotundus are morphologically similar and may be synonymized; P. typicus is redescribed and both could be transferred to Metopoides, as they have more plesiomorphic character states than members of Proboloides. -P. porcellanus is redescribed and is the type of a new monotypic genus Malvinometopa.
Th e remaining species P. typicamimus would then be the only member of the genus Proboloides living in the Antarctic, but it seems quite probable that also this species does not belong to this genus. But it is incompletely described though, based on a single specimen and knowledge about its character states is still very inadequate.
Th ere are two nominal Pacifi c species of Proboloides remaining, P. tundus and pacifi cus, which may well be synonymous: the shape of Gn2 matches (the only illustrated detail of the fi rst), and the written description of the shape of P5-7 merus in P. tundus "narrow, scarcely produced" matches the description by Shoemaker, 1964 for P. pacificus "very slightly expanded". Furthermore both species were found off California at greater depth (302 fathoms = 552 m and 718 fathoms = 1313 m, among hydroids on the back of a spider crab). -Shoemaker describes and illustrates P. pacifi cus with slender P5-7 merus, but at the end of his remarks he adds: "as shown here, the merus [of the last peraeopods] is widely expanded", which must be a lapsus linguae.
Proboloides anophthalmus is the only species living in the Indian Ocean (Madagascar). It is very similar to the Atlantic species P.? holmesi, however there are some diff erences in the shape of T, the presence of pearls on the Cx3 margin and the shape of A1, 2. Th is deep-sea species has no eyes.
Two species can be added here: "Metopa nordmanni" sensu Shoemaker, 1955: 128 fi g. 10 a-j (non Metopa nordmanni Stephensen) has to be described as new member of Proboloides, but as I could not see the material it must be cited as Proboloides sp. (Shoemaker, 1955) for the time being.
gregarius (Sars, 1882: 93 Th e habitus sketches added to the resulting tree in Fig.14 (from above: Proboloides, a large Scaphodactylus, Torometopa with A1> A2, Torometopa with A1 < A2, small Scaphodactylus and Metopoides) may give an idea about the diff erences in the body shapes in this basic group of stenothoids: e.g. members of Proboloides have a disto-posteriorly lengthened and widened merus on the last three peraeopods, which is much less the case in all other groups. Ed Bousfi eld (in litteris) opines that this must have an important hydrodynamic function, which could well be imagined. Also the relative length and width of Cx 4 (in Scaphodactylus gigantocheirus strikingly small) or the relation of the antennae (in the group near Torometopa antarctica and crenatipalmata with the second one always being longer and stronger) could tell us something about the swimming (or even digging?) ability, if we would know more about their life style. However, it seems probable that Proboloides members are mainly free-living and do not live in association with, or at least not inside of, other animals; they ought therefore to be good swimmers and have a rather strong sexual dimorphism. Th e remaining species in Fig. 14., not belonging to the genus Proboloides, will be treated in the following and fi nal part.