Stenothoids living with or on other animals (Crustacea, Amphipoda)

This paper describes new or little known species of Stenothoidae, collected from sea anemones, bivalves or hermit crabs. A key to world Stenula species is provided.


Introduction
Associations between amphipods and other animals are probably not all that rare, but few have been recorded hitherto, mainly because collection methods earlier were too crude.With the advance of Scuba-diving, new associations are being discovered at a rapid pace.The present paper adds a few more examples from the family Stenothoidae.
Material examined.1 spec.from Bermagui (400 km S of Port Jackson, Australia), 8/6/1989, Wim Vader collected on a hermit crab in shallow water.Stored at the Australian Museum (AM xxx) Sydney.
Remarks.In Haswell (1879) two species of Montagua were described from Port Jackson, Sydney: M. miersii, directly followed by M. longicornis.It appears that the first was the female, the latter the male of the same species, belonging to Stenothoe.J.L. Barnard (1974) described four species of Stenothoe from Australia.One of these he called S. ?miersi, doubting about the synonymy, as no type material is available.Barnard's description matches the different populations around most of the Australian coastline, also the Lizard Island specimens (Krapp-Schickel 2009: 873-875), and again the illustrations given here of a female.
Ecology.It may well be that this specimen lived among the encrusting hydroids and bryozoans growing on top of the hermit-crab-shell and thus had no direct association with the crab; it was the only Stenothoe specimen found among many hermit crabs.
In the European register of marine species Bellan-Santini and Costello 2001 cited 15 Stenula species: the ten above specified by Barnard 1962, plus S. alexanderi Tzvetkova & Golikov, 1990, S. nordmanni (Stephensen, 1931), S. peltata (where they mistakenly cited Della Valle 1893 as author instead of Smith 1874) and -following Lincoln 1979 -they placed Stenothoides latipes Chevreux & Fage, 1925 in junior synonymy with S. rubrovittata (Sars, 1892), which according to them therefore should be the actual type at the moment.We do not think this is correct, as Stenothoides latipes remains in any case the type.
Thus at the beginning of this study 14 species belong to Stenula.Judging from the illustrations of the mandible in Tandberg 2011: fig.25, M. invalida Sars, 1892 from N. Norway has to be added as 15th species.These species are mainly living in the far north region (N-Atlantic, N-Pacific, Arctic), only two of them were described by J.L. Barnard from California.
Tandberg & Vader could demonstrate in Tandberg (2011), that the character of Gn 2 palmar corner present/absent does not bring any clear results in a cladistic analysis.E.g.Metopa clypeata (the type species) or M. palmata, both with strongly rectangular palmar corner, strangely enough are not grouped together with M. alderi = M. spectabilis or M. norvegica, probably because of the strong allometry, which shows their members with very different palmar corners depending on age.It might therefore be more helpful to look at the shape of Gn 1, which shows normally much less allometry and which can be basic (the type of Stenula plus several other members of this genus and a lot of Metopa, with the carpus shorter than or equal to the propodus) or elongated (type of Metopa and some other Stenula, with Gn 1 simple and carpus, often also propodus, much lengthened and narrow).
Tandberg 2011 cites in her thesis at the beginning a letter from G.O. Sars to Sparre-Schneider, writing "I have advanced to the supposedly most difficult of all amphipod-families: Stenothoidae".There is no doubt that there is a great difference between having a fully developed mandibular palp (Metopa) or none (Stenothoe), but the genus Stenula, as presently conceived, gathers all transitions, and is with high probability heterogeneous.
There are also various transitions within the maxillae, having two (Stenothoe) or one (Metopa, Stenula) articles in Mx1 palp, where often one cannot clearly decide if and where an articulation is present; while the Mx2 plates may sit in tandem-position (many Metopa like M. affinis, aequicornis, groenlandica, glacialis, clypeata) or riding position (in some Stenothoe and Stenula), with all steps in-between.
In three species we have no information about the mandible palp at all: S. rubrovittata, S. modosa, S. peltata.The following have a short stump, about as long as the width of the mandible-incisor: S. angusta, S. bassarginensis, S. ratmanovi.All other species have a uniarticulate mandible palp which is clearly longer than the mandible-incisor: S. alexanderi, S. arctica, S. beringiensis, S. incola, S. serripes, S. ussuriensis, and also Metopa derjugini Gurjanova, 1948, which is therefore here also transferred to Stenula (see above).Just 1980: 52 looked at the mandible of Metopa nordmanni using the type specimen, and found again a uniarticulate palp longer than the mandible-incisor (also illustrated by Tzvetkova and Golikov 1990), while Shoemaker 1955: 127 found material from Point Barrow with strikingly similar legs but different antennae (A1>A2), a two-jointed Mx1 palp and a 3-articulated mandible palp.Although he cites Metopa nordmanni Stephensen, 1931 in the synonymy-list, his species belongs to Proboloides and thus is a different animal with nearly identical body but different mouthparts, an observation which can be made rather frequently within Stenothoidae.
Stenula species could also be divided by the ratios of articles in Gn 1, having propodus subequal to carpus, or clearly much longer resp.clearly shorter.The first group is formed by the majority: S. beringiensis, S. derjugini, S. incola, S. latipes, S. modosa, S. peltata, S. ratmanovi, S. serripes; propodus is longer than carpus in S. angusta; propodus is shorter than carpus in S. arctica, S. bassarginensis, S. nordmanni, S. ussuriensis and also S. alexanderi (this species is very aberrant also in the shape of Gn 1 dactylus).
It is the great help of a cladistic analysis that one can test the states of many characters together, and if a group of characters is changing together, it is more probable that a naturally related clade is found.But in the above listed species there are A1 subequal A2 or much different, Gn 1 propodus simple, rounded or with strong palmar corner, Gn 1 carpus short or extremely lengthened, Gn 2 propodus regularly rounded or deeply excavated, P 6, 7 strongly rounded or with widened but parallel margins, telson spinose or naked.And even using more than 60 characters as in the very exhaustive phylogenetic analysis of Tandberg & Vader (Tandberg 2011), there remains the big danger that the character states are not homologous.As an example, several analyses bring Stenula incola J.L. Barnard, 1969 from the intertidal of California always closely together with Stenula serripes Gurjanova, 1955: both show a one-articulate mandibular palp of medium length, Gn 2 with a well defined palmar corner, P 7 basis about as wide as long and P 7 merus very much lengthened and widened, and they are thus "very similar" after the coded characters.But their biogeography, Gn 1 and P 6 are obviously quite different, and they are most probably not closely related at all.
At the moment there is nothing else to do than to continue "making order" within this complicated family of Stenothoidae in describing as completely as possible its single members.
First we tried to find material of Metopa rubrovittata Sars from the northern North Atlantic (type loc.Christiansund and Finnmark) for comparing it with material of Stenula latipes (Chevreux & Fage) from the English Channel (type loc.Saint Vaast la Hougue, see Chevreux 1908Chevreux : 42, 1925: 130): 130).(Sars, 1883) by Lincoln 1979.But until now really nobody had looked at the mouthparts of M. rubrovittata, an often cited species, which nevertheless is rarely found in Museum collections.

Metopa rubrovittata Sars, 1883
In fact, the studied material shows a classical mandible palp of Metopa species with 3 articles, though it has to be admitted that it was quite a difficult task to see always the articulations.But nevertheless, already the length of the mandible palp is very different in the material from the Channel (cf.Fig. 8 with Fig. 10B, C), thus it can be confirmed that Chevreux was right: both Metopa rubrovittata Sars, 1883 as well as Stenula latipes (Chevreux & Fage, 1925) do exist, and they show extremely similar body morphology, colour pattern and even ecological niche, only the mouthparts are somewhat different.Thus, Metopa rubrovittata Sars, 1892 is herewith revalidated and Stenothoides latipes (Chevreux & Fage, 1925) re-mains the type species of Stenula.It is not clear what the geographical distribution of the two species is, as all the numerous citations cannot be judged without examining their mandible.
For control Jean-Claude Sorbe sent us material from the Bay of Biscay, and the single specimen he had collected affirms this decision.As the original paper is not easily accessible and as there is some confusion about the authors, I repeat herewith the type-description by Smith:

Description.
Female.Eyes round and nearly white in alcohol.Antennulae (=A1) considerably shorter than epimera of the 4th segment (Cx 4); first article of the peduncle stout, subequal to head, the second shorter, the third very short and similar to the arts of the flagellum; flagellum scarcely longer than the peduncle, with 8 arts.Antennae (=A 2) slightly longer than antennulae; peduncle art 4, 5 about equal in length; flagellum subequal to flagellum of antennulae.Cx 2 (fig.5) nearly ovate, twice as high as broad; Cx 3 somewhat rectangular, not wider than the second but considerably deeper; Cx 4 (fig.6) very large, slightly deeper than Cx 3 and 1/3-1/4 longer than deep, being about as long as the first five segments of the thorax, the inferior margin regularly curved and the posterior convex in outline.Gn 1 (fig.7) small and slender; merus   triangular and distally broader than the carpus, which is not quite twice as long as broad and has the lateral margins parallel; propodus narrower but slightly longer than the carpus and narrowed distally; dactylus about half as long as the propodus.Gn 2 (Fig. 5) stouter; merus short triangular, carpus much broader than long and only slightly produced beneath the propodus; propodus about as long as the breadth of Cx 2, nearly twice as long as broad; palmar margin (Fig. 8) convex in outline, slightly oblique, with an acute lobe and a spine at the posterior angle, within which the top of the dactylus closes.P 4, 5 slender and nearly naked, P 5 basis slender, four times as long as broad, not wider than the merus.P 6, 7 slightly shorter than P 5, basis posteriorly dilated and squamiform in both pairs, but broader in P 7. U 3 ramus slightly longer than the peduncle.
Length of largest specimen, from front of head to tip of telson, about 6 mm.
The mandibles are without palp or molar tubercles, and in all other characters the species agrees with the genus Stenothoe as restricted by Boeck, but it seems to be very distinct from either of the European species.
Discussion.The hint after the original description, that this species should belong to Stenothoe as it has no mandible palp, was not convincing: no Stenothoe is described from the region off Massachusetts or Connecticut, nor from the entire Atlantic, with gnathopods similar to the ones illustrated.
The incomplete illustrations of S. ratmanovi (Gurjanova, 1948) are very similar to what little we know about ?Stenothoe peltata, and the two species may well be synonymous, in spite of the large geographic distance between the type localities.In that case the older name Stenula peltata (Smith, 1874) would become the valid name of the taxon.
We hoped to get more information by studying the single type specimen (see Fig. 11,12) and illustrate here all what we could see; but there were no mouthparts except the maxilliped, and we still don't know anything about the shape of the mandibular palp.
A sample in the collections of the Smithsonian Inst.(Washington) raised new hope to shed light in this situation: there could exist a Stenula sp. from the coelenteron of Haliactis arctica.Will this be S. peltata?Etymology.the epitheton should remind on the shape of the propodus Gn 2, which looks somewhat like a small fist, in Latin "pugilla"; it is used as noun in apposition.
Remarks.The note ‚from the coelenteron of Haliactis' on the label of this sample may as well just have meant that the sea anemones had contracted on collection.
Discussion.Within the above discussed criteria of dividing Stenula species into groups, the new species belongs to the majority having Gn 1 propodus subequal to carpus (together with S. solsbergi, see below, here transferred to Stenula) and to the few members having a very short stump of mandible palp.The shape of Gn 2 palm male and female defined by a thumb-like hump is unique and quite helpful in identifying this species.This seems to be the very first time that the mouthparts were checked, and a reduced, uniarticulate mandibular palp could be illustrated, moving also this species to Stenula.
The amphipod genus Stenula is probably not a monophyletic entity (cf.Tandberg and Vader 2011, this paper), and it is at present not possible to decide which taxa belong to it.In this key we have therefore cast our nets widely, and we include all species in the Metopa-Stenula complex with a uniarticulate mandible palp.This palp is very short in what we might call ‚typical Stenula', a bit longer, but still shorter than the incisor of the mandible, in a number of other species, also traditionally placed in Stenula, and still a bit longer, but clearly uniarticulate, in a few Metopa species: M. hearni, M. palmata and M. sinuata.Just's (1980) "Stenula sp." is in our opinion identical with M. sinuata, as that author himself already suspected.
The task has been made more difficult by several factors: many species have only been partly illustrated, and at least for the species S. angusta, S. invalida, S. modosa and ?S. peltata, as well as possibly some of Gurjanova's species, males are still unknown.

A short survey of associations between stenothoids and larger marine invertebrates (Table 1).
Table 1 lists the associations between stenothoid amphipods and other marine invertebrates known to us, with the exception of those reported from sponges, hydroids or bryozoans.These latter are excluded because in most cases it is unclear what the exact niche of the amphipods is: usually the labels say only ‚among hydroids and bryozoans' or ‚found together with sponges'.Among the others the associates of various large coelenterates and also those found in ascidians generally do not seem to be obligate symbionts.Although now and then found in large numbers (e.g.Metopa bruzelii and Proboloides calcarata on gorgonians), the same species are also regularly found apparently free-living.
The situation is different for the associates of mollusks (all Metopa species) and those on sea anemones (mostly Stenothoe and Stenula species).Practically all these species appear to be obligate associates of only a single or in some cases a few hosts, and they have never been found free-living (for a possible exception see Blain and Gagnon 2014, who claim to have found numbers of Stenothoe brevicornis on the alga Desmarestia viridis).The amphipod associates of sea anemones always live on the column of the host or among the tentacles (Elmhirst 1925, Krapp-Schickel and Vader 2009, Krapp-Schickel et al. 2015;Vader 1983, Vader andKrapp-Schickel 1996), with the possible exception of Stenula pugilla, found according to the label ‚in the coelenteron of Haliactis arctica' (this paper).At least Stenothoe brevicornis, somewhat surprisingly, feeds to a large extent on host tissue (Moore et al. 1994), contrary to earlier assumptions (Vader 1983).Large numbers of amphipods are usually found on a single host, and ovigerous females are commonly present.Interestingly, sexual dimorphism is in most cases much less developed in the associates of sea anemones than in related free living stenothoids (see also Vader 1983).
In contradistinction to the case with sea anemones, all the stenothoid associates of bivalve mollusks are Metopa species.Once more the associations seem to be obligate ones, the amphipods are rarely found free-living (and never leave their hosts in laboratory observations) and they are confined to a single host or, in the case of Metopa glacialis, to a series of closely related host species.A partial exception is Metopa alderi, usually an associate of large hydroids and hydromedusae, that recently was found in Musculus spp in N. Spitsbergen (Tandberg et al. 2010b).The data on mollusk-associated stenothoids have recently been reviewed by Tandberg et al. (2010a): the amphipods live inside the host and feed on that part of the ingested material that the host does not consume itself.In addition, the stenothoid symbionts of bivalves seem to exhibit territoriality as well as extended parental care: invariably only a single pair of adults is present within a single host, often together with several cohorts of juveniles.
In the case of the single, quite aberrant Stenothoe species that lives on a spider crab, S. symbiotica Shoemaker, 1956, its biology is as yet completely unknown, but also this association appears to be an obligate and probably species-specific one; the species has never been collected elsewhere and it has clearly prehensile peraeopods.Also the amphipod associates of hermit crabs and their tenanted mollusk shells are of unknown biology.Metopelloides paguri Marin &Sinelnikov, 2012 andM. micropalpa (Shoemaker, 1930) have slightly but clearly prehensile posterior peraeopods, and may therefore well be direct associates of their host hermit crabs (Vader 1983b).But the somewhat mysterious pair of Metopa rubrovittata Sars and Stenula latipes (Chevreux & Fage) do not have prehensile peraeopods (even though the posterior legs carry maybe more spines than is usual in Metopa species?) and many authors have associated these species primarily with the Hydractinia-cover of the tenanted gastropod shells rather than with the hermit crabs themselves, although without any proof.These two species occupy the same niche, and slightly different, but possibly overlapping distributions, and have the exactly identical, quite special coloration pattern, but according to present classifications they have to be placed in different genera.Also the species associated with hermit crabs seem to be largely obligate symbionts, although possibly occurring on a larger range of hosts.

Table 1 .
Associations of Stenothoidae with larger marine invertebrates.