Revision of Leucothoe (Amphipoda, Crustacea) from the Southern Ocean: a cosmopolitanism concept is vanishing

Among the 125 currently recognized species of the panoceanic genus Leucothoe, L. antarctica was described in 1888 from the Antarctic seas, but was soon synonymized with the so-called cosmopolitan Leucothoe spinicarpa Abildgaard, which was cited from the Southern Ocean about 70 times since this fi rst record. After erecting a new Antarctic species again only in 1983, “morphological variants” were observed and discussed. In this paper, we revalidate the fi rst defi ned Antarctic species (Leucothoe antarctica), redescribe the second one (L. orkneyi), describe 5 new Southern Ocean species (L. campbelli sp. nov., L. longimembris sp. nov., L. macquariae sp. nov., L. merletta sp. nov. and L. weddellensis sp. nov.) and provide a key to all Antarctic and sub-Antarctic species.


Introduction
In the context of growing impact of environmental change on polar marine ecosystems (Turner et al. 2009;Verde & di Prisco 2013), it appears crucial to establish comprehensive baseline information on the Antarctic marine biodiversity as a sound benchmark against which future change can reliably be assessed. The Census of Antarctic Marine Life programme (www.caml.aq) was partly set up to answer this need and organized several exploratory campaigns in various parts of the Southern Ocean (Gutt et al. 2010;Schiaparelli et al. 2013). This prompted the establishment of a comprehensive register of Antarctic marine species by a large panel of specialists (De Broyer & Danis 2011).
species Leucothoe spinicarpa (Abildgaard, 1789) (see Lowry & Bullock 1976;De Broyer et al. 2007). In the course of the revision of the Antarctic and sub-Antarctic amphipod fauna in the framework of the "Synopsis of the Amphipoda of the Southern Ocean", we revise here the Leucothoe Leach, 1814 material from some recent Southern Ocean collections.

Material and Methods
Complete specimens were studied in temporary glycerine slides, preparations made in Faure's fl uid and both studied under various light microscopes. Pencil drawings were scanned in and 'inked' by applying the software Adobe Illustrator CS3 and Wacom tablets A4 and A5, using the method described in Coleman (2003Coleman ( , 2009. Examined specimens are stored in the various depositories mentioned in the material sections. Geographic coordinates and depth data between square brackets are approximate and based on the SCAR ENEA Composite Gazetteer of Antarctica (http://www.scarmarbin.be/gazetteer.php?p=search) and the GEBCO map respectively (see detailed methods in De Broyer et al. 2007).
In this paper the following morphological terms are applied in the same way as in most of the earlier publications of the fi rst author (see also Krapp-Schickel 2011: 2): tooth = non-articulated pointed ectodermal structure spine = stout, articulated structure (synonymous to "robust seta" of Watling 1989) seta = slender, fl exible articulated structure.
The length ratio between Gn dactylus and propodus is expressed by the shortest distance between the extreme ends of each.
Cx 4 length subequal to width or longer than wide, smooth, bare, anterior margin straight, distal margin rounded, posterior margin shorter than anterior one, not excavate, facial setae absent. P 3, 4 basis very narrow, a bit wider than merus; dactylus reaching nearly half length of propodus, posterior margins with some short thin spines.

Pleon
Ep 2 with spines on distal margin, Ep 1, 3 distally bare, but some small setae on posterodistal corner of Ep 3. Ep 1 distally rounded, Ep 2 posteroventral corner not acutely upturned, but produced with rounded corner, Ep 3 posteroventral corner blunt and rectangular.
Telson ratio length:width = 3, tip tridentate because of two indentations near distal end, with a short seta inserted in each one.
No sexual dimorphism known.

Remarks
Pfeffer (1888) compared his new species only with the Atlantic species Leucothoe articulosa (Leach, 1814), insuffi ciently described and actually kept it -with question marks -as a synonym of L. spinicarpa (Abildbaard, 1789); there is, however, no doubt that it belongs to the "spinicarpa-group".
The original description and illustration of the single type are not very helpful (see Fig. 1), but at least the semicircular posterior margin of P 7 and the scarcely upturned rounded posterodistal corner in Ep 2 are visible in the drawing.

Etymology
The latinized adjective indicates the original locality of the new species.

Head
Anterior margin rounded, anterodistal margin rectangular with rounded corner. Mid-cephalic keel not prominent. Rostrum small.
Eyes large, round.
Gn 1 basis scarcely infl ated, posterior margin with short setae; ischium with one seta; carpus basis more than half as long as propodus, distal part linear and narrow, about 10 x longer than wide; propodus straight, about 5 x as long as wide, palm with 7 short setae; dactylus smooth, reaching about 1/3 propodus length.
Cx 4 length subequal to width or longer than wide, smooth, bare, anterior margin straight, distal margin rounded, posterior margin shorter than anterior one, not excavate, facial setae absent. P 3, 4 basis very narrow, scarcely wider than merus; dactylus reaching about half length of propodus, posterior margins without spination.
No sexual dimorphism known.

Distribution
Campbell Island.

Remarks
This species has gnathopods similar to the ones of the small specimens of L. weddellensis sp. nov., but the peraeopods are more robust and the posterodistal corners of Ep 2, 3 different; while the here described species lives in shallow depths, L. weddellensis sp. nov. was always found much deeper.
The species has the same ecology as L. macquariae sp. nov., but differs from the latter in Gn 1, 2 propodus, P 7 basis and Ep 3.
Holman & Watling (loc. cit.) described their "variant 2" as following: "Coxa 3 with smoothly rounded ventral margin, coxa 4 not excavate posteriorly. Gnathopod 1 dactyl 1/4 length and slightly more than 1.5 times as wide as propodus, latter almost 7 times as wide. Third epimeron with quadrate posterodistal corner [...]. The fi rst gnathopod has a considerably shorter dactyl than variant 1 [see L. weddellensis sp. nov. below], but appears similar to Sars' illustrations fo L. spinicarpa in most other respects. In stations where both variants 1 and 2 occurred, there is also a clear size dichotomy. Large specimens correspond consistently to variant 1, while smaller ones correspond to variant 2 [...]. In addition, several variant 2 females were found to be ovigerous." This text stresses how small the morphological differences are, but at the same time confi rms the size differences of the ovigerous females.
The ecology, the quite shallow depth range as well as the shape of Gn 2 dactylus and Ep 2, 3 are good indications for this species.

Etymology
Many articles (in Latin membrum, abl. plural membris) like antennae, Gn 1, but especially the bases of the peraeopods 5-7 are unusually elongated (= longis) and this character is different from all other species hitherto known from the Antarctic or sub-Antarctic regions. The chosen specifi c name is used as non-fl exible apposition.

Type material
Holotype

Head
Anterior margin rounded, anterodistal margin rectangular with rounded corner. Mid-cephalic keel not prominent. Rostrum small.
Eyes absent.
Gn 1 basis slender, not infl ated, both margins with short setae; ischium smooth; carpus basis about half as long as propodus, distal part linear and extremely narrow, about 10 x longer than wide; propodus straight, about 6 x as long as wide, palm with short setae; dactylus smooth, reaching about 1/3 propodus length.
Cx 4 length shorter than width, smooth, bare, anterior margin nearly straight, distal margin rounded, posterior margin shorter than anterior one, not excavate, facial setae absent. P 3, 4 basis very narrow, wider than delicate merus; dactylus reaching nearly half length of propodus, all margins naked.
Uropods: U 2 the shortest, with very unequal rami (ratio about 1:2), the shorter one shorter than peduncle. U 1 and U 3 with subequal rami.
Telson ratio l:w = 3, tip fi ve-dentate because of four indentations near distal end.

Remarks
The new blind species is similar to the (also blind) deep-sea species L. ayrtonia Bellan-Santini, 1997 (5 mm) and L. atosi Bellan-Santini, 2007 (12 mm), found in the Mid-Atlantic Ridge at 37°N, 31°W at a depth of 850 m, and at the Barbados Trench, 10°N, 58°W at a depth of 1947 m respectively. Differences from L. ayrtonia are in the large rostrum, Cx 4 with posterodistal acute corner, Ep 3 quadrate with small tooth, from L. atosi are in much shorter antennae, blunt posterodistal corner at Ep 2, less unequal U 2 and broader telson. Also L. ushakovi Gurjanova, 1951 (type from Greenland at a depth of 3892 m, specimens from 8-34 mm length; see Krapp-Schickel & Vader 2012) has P 5-7 bases not medially, but proximally widest and is blind; differences with L. longimembris are are the setose bases of Gn 1, 2 and very unequal rami in U 2 in the here described material (vs. naked bases and less unequal rami in L. ushakovi), a longer Gn 1 dactylus and a fi ve-dentate slender telson here (vs. short dactylus and threedentate broader telson in L. ushakovi).

Head
Anterior margin rounded, anterodistal margin rectangular with rounded corner. Mid-cephalic keel not prominent. Rostrum small.
Eyes large, round.
Gn 1 basis scarcely infl ated, anterior margin with some short setae; ischium smooth; carpus basis more than half as long as propodus, distal part linear, narrow, about 6 x longer than wide; propodus straight, about 3-3.5 x as long as wide, palm with 8 short setae; dactylus smooth, reaching nearly 1/2 propodus length.
Gn 2 basis not infl ated, on anterior margins 3 short setae; carpus nearly reaching one third of propodus length, curved, distally rounded, densely setose; propodus anterodistally without prolongation, with a bundle of setae, posterior margin in male with many low humps, in female smooth, palm convex, proximally near dactylus-end no corner; facial setae medially and submedially present; dactylus curved, both margins smooth, bare, reaching somewhat more than half of propodus length.
Cx 4 length subequal to width, smooth, bare, anterior margin straight, distal margin oblique, posterior margin clearly shorter than anterior one, not excavate, facial setae absent. P 3, 4 basis very narrow, a bit wider than merus; dactylus reaching nearly half length of propodus, posterior margins naked.
Uropods: U 2 the shortest, with unequal rami (ratio about 1.4), the shorter one shorter than peduncle. U 3 peduncle more robust than in the other species, margine spinose; rami unequal, ratio 1.2 to 1.25.

Remarks
This species has some similarity to L. antarctica, but P 7 basis is less wide and has always a lengthened posterodistal corner (nearly semi-circularly rounded, corner not or scarcely lengthened in L. antarctica); Gn 1 propodus is much more robust here and U 3 rami are unequal (vs. U 3 slender with equal rami in L. antarctica); Ep 3 has a somewhat upturned rounded posterodistal corner (vs. rectangular corner in L. antarctica).

Diagnosis
Eyes round, yellowish. Mandibular palp art. 3 nearly as long as art. 2. Cx 2 smooth, in shape of a parallelogram, anterodistal corner acutely lengthened. Cx 4 also with anteridistal corner acutely lengthened, but posterior margin much shorter than anterior one. Gn 1 propodus l:w = 4, dactylus reaching nearly 1/2 of propodus length. P 5-7 basis oval, with regularly rounded hind margin, ratio l:w about 1.5; posterior margin in P 7 smooth. Ep 2 distoposterior corner slightly upturned, Ep 3 posterodistally with rectangular corner. T l:w not known.

Etymology
In Italian "merletto" is the bobbin lace; as the very acutely elongated Cx 2 and Cx 4 can remind on lacecollars, this epitheton (Latinized fl exible adjective) was coined.

Type material
Holotype

Head
Anterior margin rounded, anterodistal margin rectangular with rounded corner. Mid-cephalic keel not prominent. Rostrum small.
Eyes round, yellow.
Gn 1 basis somewhat infl ated, smooth; ischium smooth; carpus basis more than half as long as propodus, distal part linear, narrow, about 7-9 x longer than wide (width very different); propodus straight, about 5 x as long as wide, palm smooth; dactylus smooth, reaching more than 1/3 propodus length.
Cx 2 about as wide as long, in the shape of a parallelogram, width similar to Cx 3, distally smooth; anterior margin scarcely excavate, anterodistal corner acutely lengthened, inferior and posterior margin straight, facial setae absent.
Gn 2 basis not infl ated, on anterior margin some setae; carpus nearly reaching half propodus length, curved, distally rounded, with small incision, setose; propodus anterodistally without prolongation, with some setae, posterior margin with many low humps, palm convex, proximally near dactylus-end with weak blunt corner; facial setae medially and submedially present, but not abundant; dactylus curved, both margins smooth, bare, reaching somewhat more than half of propodus length.
P 5-7 similar, bases l:w ratio about 1.5, margins smooth. Merus with posterodistal corner scarcely lengthened and not widened.
Uropods: U 2 with unequal rami, the shorter one shorter than peduncle. U 3 unknown.
Telson ratio l:w unknown.

Remarks
Until now we know only one representative of this new species, but the peculiar shape of Cx 2 and Cx 4 makes it easy to recognize it within the Antarctic and sub-Antarctic species. Within the hitherto known Leucothoe species L. lilljeborgi (Boeck, 1861) as well as L. ayrtonia Bellan-Santini, 1997 have a similar Cx 4 with a pointed anterodistal corner but rounded Cx 2, while L. angusticoxa (Ledoyer, 1972), described as "Leucothopsis n. gen.", has Cx 2 pointed and Cx 4 rounded.

Head
Anterior margin rounded, anterodistal margin rectangular with rounded corner. Mid-cephalic keel with acute projection, extending past epistome. Rostrum small.
Gn 1 basis somewhat widened, lacking setae; ischium smooth; carpus basis about half as long as propodus, distal part linear and extremely narrow (width not much differing, about 9 long); propodus straight, 7-8 x as long as wide, palm minutely dentate with 12 short and up to 8 longer setae; dactylus smooth, reaching about 1/5 propodus length.
Cx 2 subquadrangular, about as long as wide, much wider than Cx 3, distally smooth; anterior margin straight, anterodistal corner rectangular, inferior and posterior margin straight, facial setae absent.
Gn 2 basis not infl ated, on both margins up to 14 short setae and on anterior margin one longer one; carpus reaching about 1/3 propodus length, curved, distally truncate, with small serration, densely setose; propodus anterodistally with short acute prolongation and bundle of setae, posterior margin with many low humps, palm convex, proximally near dactylus-end widening with a characteristic blunt corner; one row of facial setae below the middle line; dactylus curved, both margins smooth, bare, reaching more than half of propodus length.
Cx 4 wider than long, smooth, bare, anterior margin scarcely convex, distal margin rounded, posterior margin shorter than anterior one, not excavate, facial setae absent. P 3, 4 basis slender, a bit wider than merus; merus anterodistally lengthened; dactylus reaching more than half length of propodus.
Telson ratio l:w > 3, tip tridentate because of two indentations near distal end, with a short seta inserted in each one.

Remarks
The material examined here matches the description and fi gures of Holman & Watling quite well, but our specimens are always considerably smaller, found at shallower depth and their peraeopod bases are posteriorly serrate. Moreover the original description tells about specimens of 4.5-6.5 mm from 61°27'S, 41°55'W, while others from a similar area were between 6.5 and 9 mm, both localities from about 600 m depth. On p. 230-231 (loc. cit.) the authors wrote: "[...]of P 7, the basis of which appeared broadly rounded. Some slight variability in the degree of expansion of this article was seen [...], so its diagnostic value is questionable." Thus it could very well be that this material consists of a group of closely related and morphologically extremely similar species.

Etymology
The latinized adjective refers to the fact that much of the material comes from the Weddell Sea.

Preliminary remarks
Before starting to cite the localities, we have to explain the division in two parts: Quite like Holman & Watling (loc. cit.) we hesitated between attributing the entire material to one species, or dividing the material into two new species. Like the cited authors we fi rst separated the material into two different size classes with large or small ovigerous females, but then could not fi nd any clearcut morphological defi nition which would not show here or there intermediate structures. Unfortunately, we have no clue about the life span of these females and how often they become ovigerous, thus all the observed tiny differences could also be due to allometry.
Therefore we kept the locality citations separated and fi rst gave information about the smaller specimens, followed by the larger ones. This might help to continue gathering observations within this so puzzling genus. Fig. 19. Leucothoe weddellensis sp. nov., 14 mm. A 1, 2 = antenna 1, 2; UL = upper lip; Mx 1, 2 = maxilla 1, 2; Md = mandible; Gn 1, 2 = gnathopod 1, 2.      Cx 4 length subequal to width, smooth, bare, anterior margin rounded, distal margin straight to excavate, posterior margin shorter than anterior one, straight to slightly excavate, facial setae absent. P 3, 4 basis very narrow, a bit wider than merus; dactylus reaching less than half length of propodus, posterior margins with some short thin spines.
Uropods: U 1 with equal rami, shorter than peduncle; U 2 the shortest, with unequal rami, the shorter one clearly shorter than peduncle; U 3 with slightly unequal rami much shorter than peduncle.

Remarks
In Holman & Watling (1983: 227, Fig. 10) males are illustrated which are 19-24 mm long and match perfectly to the here illustrated material of 20 mm. Holman & Watling (l.c.: 229) defi ned their variant 1 as follows: "...ventral margin of coxa 3 bilobed in larger specimens, coxa 4 not excavate posteriorly. Gnathopod 1 very similar to that illustrated by Sars (1895) for L. spinicarpa, dactyl nearly half length and twice width of propodus, latter 5 times as long as wide. Gnathopod 2, distal margin of article 5 bifurcate in large males. Third epimeron posterodistal corner with small tooth but without distinct sinus above; however, smaller specimens tending not to have this tooth or distinctive coxa 3 shape...." Differences to the smaller material are, besides the nearly double length of mature females, the length of Md palp art. 3 (we doubt that there can be such a strong allometry), the shape of Cx 3, P 5-7 basis and U 3 rami. However, the main reason of our hesitation is that we found samples with both small and larger ovigerous females, and with larger specimens with and smaller ones without the characteristic shape of Cx 3 living together.
Within all hitherto known Leucothoe species only L. serraticarpa Della Valle, 1893 has a distal lobation on a coxa in large specimens (like here on Cx 3), but in the former it is in Cx 4 and has a different shape.

Discussion
Widespread or "cosmopolitan" distribution for sublittoral amphipods of the Southern Ocean was reported several times (see De Broyer et al. 2007, for complete distribution records of benthic, bentho-pelagic and pelagic species occurring in the Southern Ocean). In clarifying the taxonomic status of species previously considered cosmopolitan, this study -on the basis of signifi cant material -contributes to establish step by step, a more refi ned view of the biogeography of the Southern Ocean, where the benthic amphipods present a high rate of endemism of 83% (De Broyer et al. 2007).
Recent molecular studies demonstrated that many species considered widespread or cosmopolitan on the basis of morphological criteria were in fact composed of several cryptic species. A particularly demonstrative case is the combined molecular and morphological analysis conducted by Havermans et al. (2013) on Eurythenes gryllus (Lichtenstein, 1822), a cosmopolitan deep-sea benthopelagic lysianassoid, which appears to represent in fact a complex of 9 putative species-level clades, with more restricted distribution.
For the Antarctic amphipod fauna, these and the present results emphasize the need to re-examine morphologically in more details and to test with molecular methods the validity of some widespread distributed species, in particular the so-called panoceanic species.