A new North‐East Atlantic species of Bacescomysis (Mysidacea, Crustacea) with original moth‐like antennules

A new Mysidacea, Bacescomysis papilio, with a unique, supernumerary complement of sensilla on the antennules, is described from the Porcupine Abyssal Plain (4848 m) in the North‐East Atlantic. Comparisons are made with the species of Bacescomysis and of the related genus Hansenomysis. Biological and biogeographical remarks on these two genera are provided, i.e. role of antennule, adaptations to deep water, and correlations between depth range of species and latitude.

tufts of sensilla on the inner side of each article; anterior part of cephalothorax, and more particularly mouth area, violet coloured; lateral margins of carapace deeply indented.

Description
Bacescomysis papilio is a medium-sized species (20.8 mm). The carapace is short (length 5.8 mm), leaving the last two thoracomeres uncovered (Figures 1a, 2a). It is soft, membraneous and transparent, so that the mandibles are clearly visible laterally ( Figure 1c). Two well-marked grooves run across the dorso-lateral area, before and behind the mandibles, the posterior one probably corresponding to the cervical sulcus since it indicates the limit between the head and the thorax. Frontal margin gently rounded, with slightly upturned rim. Eye notch present. Antero-lateral corners acutely produced. Posterolateral half margin deeply indented with four notches (Figures 1d, 2a). Postero-dorsal margin emarginate, fitting with the antero-dorsal part of the seventh thoracic somite, shaped like the pommel of a saddle ( Figure 2a).
Eyeplate well evident, sticking out from beneath the carapace, flanked by two anterior horn-like projections considered to represent the reduced eyes (Figures 1c,3a). Only the appendages having a taxonomical and/or a biological importance in the genus will be described. Antennules long, 11 mm, and stout (Figure 1a, b). Peduncles robust, with the Tattersall organ situated proximally on the dorsal proximal side of the first segment ( Figure 3a) as a pit-like depression partially overlapped by a rounded flap projecting forward. Outer flagellum considerably thicker than the inner one all along (Figure 1a, b). The proximal 33 segments are swollen and provided on inner edges with two transverse clusters of long and densely set sensilla (Figures 1b, 4a, b). The six distal segments are normally sized and devoid of sensilla. Each sensillum is made of two parts: a very short basal portion, rounded in cross-section, and an elongated distal, strap-like part (Figure 4be). A nerve is visible in the sensilla core when broken (Figure 4c, e). Antennal peduncles shorter and more slender than those of the antennules (Figure 1a-c). Antennal scale lanceolate, setose all along its inner margin and along the distal four-fifths of its outer margin ( Figure 2b). Two spines separate the naked and setose parts of the outer margin. First segment of the flagellum longer than the second; all the others considerably shorter. Mandibles with a small blunt, feebly sclerotized pars incisiva; mandibular palp long and rather slender (Figure 1c), with numerous long setae.
Endopod of the first pair of thoracic limbs enlarged and violet as the rest of the mouth area ( Figure 5a). Endopod of the third to the fifth thoracic limbs subchelate, whilst those of the last three pairs with a slender claw. Eighth thoracic limbs with, at their base, an ampulla-like genital protuberance in postero-internal position (Figure 5b).
Pleopods biramous. The endopods of the first pair are reduced to a single segment ornamented with setae ( Figure 2c). Those of the fifth pair are about twice as long as the exopods (Figure 6a, b) and bear a few setae distally, at the level of the very short terminal tongue-shaped segment; no other trace of segmentation is seen on their length. Exopods of the second pair bearing one long serrulated dagger-like seta on the inner distal part of the basal segment (Figure 5c-e). Uropods shorter than the telson (Figure 6b). Exopods lacking distal articulation; their outer margin naked, with two small spines distally ( Figure 2d); apex and inner margin ornamented with setae. Endopods almost as long as the exopods, but half their width, totally fringed with setae. Telson elongated, bearing spines of different size along its parallel margins and rounded apex (Figures 2e, 6c): three or four isolated spines, then three sets of two, one set of three and one of six spines (in each set, spine length increases distally); the two longest are the subterminal pair which flank eight small spines separated by a large median one. Ventral integument particularly thin and translucent allowing clear observation of the nerve cord and the thoracic and abdominal ganglia between the muscular masses ( Figure 5a).

Etymology
The name of the species refers to the aspect of the antennules which recall the antennae of a moth.

Comparison with other species
The genus Bacescomysis has been separated from Hansenomysis Stebbing, 1893 by Murano and Krygier (1985). It differs from the latter in four main characteristics: (1) exopods of the uropods unsegmented; (2) vivid violet coloration of the anterior portion of the cephalothorax; (3) bi-horned anterior margin of the eyeplate; and (4) exopods of the second pleopods of males with three or four long, thick and denticulate setae. Nevertheless, these authors questioned the validity of the last characteristic since Hansenomysis rostrata Birstein and Tchindonova, 1970 displays such modified setae. The validity of the third character must also be questioned; indeed, H. tropicalis Bacescu, 1967 andH. pseudofyllae Lagardère, 1983 have two well-defined prominences (the former), and horn-like projections (the latter), on the eyeplate.
A fifth characteristic to separate the two genera is the lateral ornamentation of the telson. In Bacescomysis, there are only one or two long spines near the tip of the telson; the lateral edges of the telson are provided along most of their length with more or less small-sized spines. In Hansenomysis, there are three to six long spines flanking short series of smaller spines; in each series, the length of the spines slightly increases distally.
Thus, except for the colour, which often disappears after a long period of preservation, the ornamentation of the telson and the unjointed or two-segmented condition of the exopod of the uropods are the best characters to distinguish the two genera. The presence of a joint in this appendage would be a primitive feature, according to Burkenroad (1981). Its presence in only one species of Euphausiacea, Bentheuphausia amblyops G. O. Sars, 1885, considered to be the most primitive representative of this order of pelagic crustaceans (Casanova 1984), argues for Burkenroad's hypothesis. On the other hand, its variability in the Mysidacea weakens this view. Indeed, it exists in three genera of the primitive suborder Lophogastrida (Gnathophausia W.-Suhm, 1873, Eucopia Dana, 1852, Paralophogaster Hansen, 1910, but not in the three others (Lophogaster M. Sars, 1857 and the two monospecific genera Chalaraspidum W.-Suhm, 1895 and Ceratolepis G. O. Sars, 1883). In the evolved suborder Mysida, it is present in the family Lepidomysidae but not in the Stygiomysidae (although it seems to be partially expressed in Stygiomysis aemete Wagner, 1992), both being monogeneric, and in the three other genera (Petalophthalmus W.-Suhm, 1874, Pseudopetalophthalmus Bravo and Murano, 1997and Ceratomysis Faxon, 1893 of the family Petalophthalmidae, to which Bacescomysis and Hansenomysis belong. In the last and largest family, Mysidae, considered to be the most evolved, it is only observed in the small subfamily Siriellinae, although this articulation line is incomplete according to W. M. Tattersall and O. S. Tattersall (1951), and in the monogeneric subfamily Rhopalophthalmidae where both rami of the uropods are segmented. Thus, the presence of such an articulation in Hansenomysis or its absence in the related genus Bacescomysis seem to have neither a phylogenetical nor an ecological significance. Bacescu (1971) said that the seven species of Hansenomysis he studied in the Peru-Chile Trench split into ''two different groups''. One of them was later considered to represent a new genus, after the discovery of Bacescomysis pacifica Murano and Krygier, 1985, as noted above. It comprises now three species described by Bacescu (1971), i.e. B. birsteini, B. peruvianus and B. tattersallae, two species described by Lagardère (1983) in the Bay of Biscay, namely B. abyssalis and B. atlantica, B. pacifica off the Oregon coast, and B. papilio n. sp.
Bacescomysis papilio n. sp. is immediately recognizable by its laterally indented carapace, and the high number of enlarged articles on the outer flagellum of the antennules, 33, that is to say about twice that of the species of Bacescomysis and Hansenomysis, 13-18. Moreover, these articles are provided with numerous sensilla that have never been reported in the other species of the two genera. This is well evident in a male of B. abyssalis (Figure 4f-h) which exhibits these swollen articles devoid of sensilla, while the narrower one following this series (the appendage is broken at this level) is provided with two inner areas where probable sensilla-like structures were inserted, as indicated by the nerve arising from the regular perforations in the tegument. Lastly, the presence of only one modified stout seta on the endopod of the second pleopod of the male contrasts with three or four displayed by the rest of the species of Bacescomysis, and their absence in those of Hansenomysis, when this character has been looked for. So, it would not be surprising if further information on B. papilio, such as molecular data from adequately preserved specimens that might be caught in the future, lead it to be raised to generic level.

Sensitive role of antennule
Male-specific tufts of hair-like sensilla, looking like those of Bacescomysis papilio, are known on the antennule of mysids. They are situated on the third peduncular segment in Boreomysis arctica (Krøyer, 1861) and B. megalops G. O. Sars, 1872, and on the appendix masculina situated ventrally at the distal end of the same segment in Neomysis integer (Leach, 1814) and Praunus flexuosus (Mü ller, 1776) (Johansson and Hallberg 1992). They would have an olfactory role as chemoreceptors of sexual pheromones, as demonstrated by Guse (1983). Why does the male of Bacescomysis papilio possess so numerous sensilla on almost every article of the antennules? The absence of a terminal pore indicates they cannot be contact chemoreceptors (taste) (see Johansson and Hallberg 1992 for review). Therefore, they are not involved in food detection. Moreover, the diet of these animals probably does not require particular detection skills. Indeed, Bacescu (1971) suggested these benthic mysids feed on substratum detritus because of the great number of setae on their peri-oral appendages. SEM observations of the molar process of the mandibles of Bacescomysis abyssalis, totally smooth and concave by comparison with those of phytophagous and carnivorous species, have confirmed this hypothesis (De Jong-Moreau et al. 2001). The other species of Bacescomysis and Hansenomysis that live in the same habitat are not provided with so numerous antennular sensilla. Their role remains unknown in Bacescomysis papilio, unless this species is considered as very rare (it is absent in the numerous samples of Lagardère from the Bay of Biscay), implying thus the need for high-performance sensorial equipment to enable individuals to find one another.
The role of the Tattersall organs is more enigmatic yet. They are found in all the species of Bacescomysis and Hansenomysis of both sexes. They were not mentioned by Tchindonova (1958, 1970) in H. armata Birstein and Tchindonova, 1958and H. rostrata Birstein and Tchindonova, 1970, but Tattersall (1961, who wrote to the latter, reported that Tchindonova confirmed they existed. Tattersall described these ''pit-like'' organs, and concluded they ''have some kind of sensory function'', recalling that she previously (Tattersall 1955) suggested their chemoreceptor role. Later, Bacescu (1971) agreed with her; he suggested they were sensory organs ''of a primitive type'', and proposed the name of Tattersall organs. SEM observations of these organs have been made to try to elucidate their role. As there was a risk of damage to the holotype, specimens of Bacescomysis abyssalis were used for this purpose. The results are not conclusive. It appears (Figure 3b, c) that in the bottom of the horseshoe-shaped, pit-like depression there is a small circular aperture through which a rounded mass is seen. To our knowledge, they do not resemble sensorial organs already known in the Crustacea. Unfortunately, the internal tissues of these specimens are too damaged to allow histological sections for TEM observation. A chemoreceptor role seems probable.

Adaptations to deep water
Since the eyeplate is situated just above the Tattersall organs, it was interesting to observe the horn-like projections, considered to be remnants of the eyes. No trace of ommatidia are visible (Figure 3d), even degenerated as those of the common deep planktonic Lophogastrid Eucopia unguiculata (W.-Suhm, 1875), which nevertheless lives less deeply, i.e. below 500-600 m in the world ocean (Fage 1942). According to W. M. Tattersall and O. S. Tattersall (1951), these projections ''probably serve a sensory function''; this hypothesis is not evidenced since no sensory organs (aesthetasc sensilla, setae…) are observed. Besides the loss of the eyes, another adaptation to abyssal life in B. papilio is the reduction of the locomotory musculature, evidenced by the clear observation of almost the entire course of the nerve cord, and the ganglia, in ventral view through the thin integument ( Figure 5a). This is difficult, indeed impossible, to be seen in individuals of Boreomysis artica Krøyer, 1861 and Praunus flexuosus Mü ller, 1776 of the same size, living respectively at mid-depths and in neritic waters, since their longitudinal muscular masses are contiguous. A comparable adaptation has already been reported for deep-water decapod crustaceans such as Sergestes (Sergia) japonicus Bate, 1881 (described as S. ''mollis'' by Smith 1884 since ''the whole integument is very soft and all parts of the animal exceedingly fragile''), and numerous deep-sea fish families (Denton and Marshall 1958); this would be an adaptation to the scarcity of food resources in the abyssal waters.

Biogeographical remarks on the related genera Bacescomysis and Hansenomysis
Five species of Hansenomysis were known when Bacescu (1967Bacescu ( , 1971) described four species of this genus and three others of Bacescomysis. These seven new species plus H. falklandica O. S. Tattersall, 1955 (only represented by 12 individuals) were caught between 1238 and 4450 m deep on the sea-floor off Peru. Later, Lagardère (1983) described four new species (two in each genus) in the Bay of Biscay, and remarked the astonishing variety and the vigorous evolution of Hansenomysis (the two genera were not still separated) in the abyssal realm. The discovery of Bacescomysis papilio, the fifth species belonging to these related genera in the same area, strengthens Lagardère's remarks.
These mysids are more or less linked with the bottom, since most of them have been caught among the epibenthic fauna. The specific variety of the deep bentho-planktonic fauna has already been shown for the copepods (Grice and Hulsemann 1970;Grice 1972), the ostracods (Angel and de Baker 1982), and the chaetognaths, which are also more diversified in the tropical regions than in the Antarctic (Casanova 1986). The latter trend in biodiversity is also noticeable for the species of Bacescomysis and Hansenomysis. Indeed, only six species are known at high latitudes: Hansenomysis fyllae (Hansen, 1887) between Greenland, New England and the British Isles; H. antarctica Holt and Tattersall, 1906 andH. angusticauda O. S. Tattersall, 1961 from the Ross Sea; H. falklandica off the southern South America; and H. armata Birstein and Tchindonova, 1958 and H. rostrata from the Kuril-Kamtchatka Trench. They also live at shallow depths in these areas: from 100 m downwards in the Ross Sea, 185 m in the Magellan Strait, as compared to 650 m off New England, 950-1000 m in the New Caledonia area (Casanova 1993), 1280 m off Peru, and more than 1900 m in the Bay of Biscay for the species caught in these areas. Moreover, it appears that the species of Bacescomysis live deeper than those of Hansenomysis. Indeed, all have been observed at depths >2000 m, and mostly .4000 m, while the latter are found at depths (2000 m, except for two records of H. pseudofyllae at 3000 and 4800 m (Lagardère 1983).
The species of the two genera also seem to be strictly localized: around the Falkland Islands and the Magellan Strait for H. falklandica in the ''Discovery II'' collections from the southern Atlantic and the Antarctic Oceans (Tattersall 1955); eight species (among which one individual of H. falklandica) ''in only a few localities … despite the amount of sampling'' in the Peru-Chile Trench according to Bacescu (1971). This is perhaps an explanation for the display of numerous chemoreceptors if Bacescomysis papilio is one of the rarest species of this interesting group of mysids.