Systematic revision and palaeobiology of Rosenfeldia triasica and Rogeryon oppeli gen. et comb. nov. (Eucrustacea, Polychelida)

Polychelidan lobsters (Decapoda: Polychelida) are crustaceans with extant species which are restricted to deep water environments. Fossil species, however, used to live in more varied palaeoenvironments, from shallow water to deep water, and were more diverse morphologically. We redescribe two species of polychelidan lobsters, the Late Triassic Rosenfeldia triasica Garassino, Teruzzi & Dalla Vecchia, 1996 and the Late Jurassic Eryon oppeli Woodward, 1866, recently assigned to the same genus, Rosenfeldia, based upon only a few characters. Our investigation of all available material of both species leads us to distinguish these two species and to erect Rogeryon gen. nov. to accommodate Eryon oppeli. The palaeobiology of both species is interpreted for the first time. Rosenfeldia triasica with its stout first pereiopods and mandibles with both incisor and molar processes (documented for the first time in Polychelida) was benthic and probably fed either on slow-moving sedentary preys or was a scavenger. Rogeryon oppeli gen. et comb. nov. was benthic, visually adapted to shallow water palaeoenvironments, and possibly had a diet similar to that of slipper lobsters and horseshoe crabs. The redescription of these two species highlights the palaeobiological diversity of fossil polychelidans.

Cross-polarized light images (see Bengtson 2000) are realized by illuminating the specimen with a polarized white-light source and fi ltering the light directly refl ected on the specimen (same axis of polarization as light source) with a second polarizer on the camera lens. This way, refl ections are cancelled and contrast is increased.
Green fl uorescence (see Haug et al. 2009;Haug & Haug 2011) is obtained by illuminating the specimen with a green light (reduced spectrum compared to white "natural" light). In some outcrops, the minerals which replaced organic structures during the fossilisation process are fl uorescent, and re-emit an orange fl uorescence when exposed to green light. To observe this fl uorescence, an orange/red fi lter is required to fi lter out the green light and mostly keep the orange fl uorescence. Images are then colour equilibrated (often also desaturated) to simplify the visualisation of fl uorescent parts.
UV fl uorescence, more commonly used than green fl uorescence, relies on a simple principle. Here, the incident light is a "dark light" (blue to UVA). This incident light is fi ltered by an orange fi lter (care was taken that this fi lter was not fl uorescent under UV) to fi lter out blue, violet and UV light and keep the yellow/orange fl uorescence. As for green-orange fl uorescence, images are then colour equilibrated to simplify the visualization of fl uorescent parts. Malacostraca Latreille, 1802Decapoda Latreille, 1802Pleocyemata Burkenroad, 1963 Reptantia Boas, 1880 Polychelida Scholtz & Richter, 1995 Eryonoidea de Haan, 1841 Uncertain family Genus Rosenfeldia Garassino, Teruzzi & Dalla Vecchia, 1996 Rosenfeldia Garassino, Teruzzi & Dalla Vecchia, 1996: 33-41, fi gs 6-10, 17-20.

Original diagnosis
Subround carapace, with longitudinal median carina originating from the posterior margin; two thin carinae run parallel to the median carina; two parallel rows of tubercles are located in the median portion of the cephalic region; chela of pereiopod I with the dactylus longer than the index and bent at the distal extremity; subrectangular telson; exopodite without diaeresis. AUDO D. et al., Rosenfeldia triasica and Rogeryon oppeli gen. et comb. nov.

Emended diagnosis
Subquadrate yet slightly rounded carapace (in dorsal view); inner angle of ocular incision blunt, not forming a spine; ocular incision fringed laterally by an expansion of carapace; cervical incision shallower than postcervical incision; posterolateral angle short, not projecting along pleonite 1; very rounded telson and strengthened by a media carina; third maxilliped ischium kidney-shaped; uropodal exopod not articulated (no diaeresis).

Remarks
Rosenfeldia was assigned to Eryonidae de Haan, 1841, because (1) it did not fi t in other families then described (Coleiidae Van Straelen, 1925, Tetrachelidae Beurlen, 1930, Polychelidae Wood-Mason, 1874, and (2) based upon of its longitudinal median carina, two thin lateral carinae, elongate P1 propodus, subrectangular shape of pleonal somites and uropodal exopod without a diaeresis (Garassino et al. 1996). This opinion was followed by Ahyong (2009), De Grave et al. (2009 and Feldmann et al. (2013), who did not discuss further the familial assignment. Our reinvestigation questions this familial assignment. Indeed, as noted by Garassino et al. (1996), the telson of Rosenfeldia is rounded in dorsal view, contrary to that of Eryonidae, which is triangular. Other important differences include: the scaphocerite is rounded with a median carina, a disposition well known in Coleiidae, but not in Eryonidae (fusiform, without median carina); the third maxilliped ischium is kidney-shaped, once again a character well known in Coleiidae, contrary to Eryonidae, in which it is subtriangular. Garassino et al. (1996) argued that the "lack of cervical and branchiocardiac grooves" was also typical of Eryonidae, however, the cervical and postcervical grooves at least are present, although poorly preserved. In fact, these grooves occur in all Polychelida (Fig. 2), but are often restricted to the median part of the carapace in Eryonidae. This is not the case in Rosenfeldia. Finally, the lack of diaeresis is a poor indicator since it is present in Coleiidae, Tetrachelidae and in Hellerocaris Van Straelen, 1925, which was considered as being close to Polychelidae (Audo et al. 2014c). A diaeresis is absent (= uropodal exopod in one part) in all other polychelidan lobsters. Taking into account these characters, the assignment of Rosenfeldia to Eryonidae is equivocal, and will be tested in a future work, the phylogenetic analysis of Polychelida. Garassino, Teruzzi & Dalla Vecchia, 1996 Figs 1, 2N, 3

Emendation to description
SHAPE OF CARAPACE. Dorsoventrally fl attened carapace, subquadratic in outline in dorsal view; concave frontal margin; anterolateral angle blunt, pointing slightly outward; spiny lateral margin with ocular, cervical, and postcervical incisions; U-shaped, almost semi-circular ocular incision; anterolateral margin slightly rounded posteriorly, oblique compared to longitudinal axis, poorly preserved; cervical and postcervical incisions opening in anterior half of lateral margin; cervical incision subtriangular; mediolateral margin shorter than anterolateral margin, rounded posteriorly; postcervical incision subtriangular, slightly deeper than cervical incision; posterolateral margin, slightly rounded, fringed by small spines; posterolateral angle round, short, not fringing pleonite 1; posterior margin slightly concave, wider than frontal margin. CARAPACE GROOVES AND CARINAE. Postrostral carina not visible; postcervical carina well-marked in posterior half of carapace; very thin branchial carina, with anterior branch possibly incorporating (merged with) postorbital carina and extending on ⅓ of carapace length, posterior branch separated from anterior branch by cervical and postcervical grooves, extending to posterolateral angle; cervical groove poorly preserved, oblique, extending from cervical incision toward median line; postcervical groove poorly preserved, extending from postcervical incision toward median line, passing near cervical groove; branchiocardiac groove not visible; gastro-orbital groove thin, poorly preserved, extending obliquely from cervical groove toward median line. PLEON AND TELSON. Pleon and telson slightly shorter than carapace; pleon narrower than carapace; pleonite 1 shorter and narrower than others; s2-s5 with subrectangular terga crossed by a pair of transverse grooves and an axial carina cutting the posterior transverse groove; s6 with subtrapezoidal tergum; s1 and s6 with narrow subtriangular tergopleura (also called "pleura" although it is structurally a part of terga); other tergopleura also subtriangular but wider than s1 and s6 tergopleura; telson rounded, strengthened by a pair of slightly oblique longitudinal carinae and with fi nely serrated margins. EYE AND CEPHALIC APPENDAGES. Eye not protruding from ocular incision; antennula poorly preserved; antenna poorly preserved, with an ovoid scaphocerite strengthened by a median longitudinal carina; mandible with an asymmetric molar process (Fig. 1D) and an incisor process forming sharp irregular triangular teeth. THORACIC APPENDAGES. Third maxilliped with kidney-shaped ischium carrying a few stocky podomeres; P1 slightly larger than others; P1 dactylus and pollex (= fi xed fi nger) slightly curved; stocky P1 palm (portion of propodus excluding pollex), slightly longer than dactylus; P1 carpus short and subtriangular; P1 merus slightly longer than palm, with parallel margins; P2-P4 similar to P1, decreasing in size from the second to the fourth; P5 ending in a short dactylus, without claw in observed specimens.  PLEONAL APPENDAGES. Pleopods 1-5 poorly preserved, with subrectangular basipod; uropods with a stocky basipod carrying very rounded endopod and exopod, both with thin serration of their margins; uropodal exopod with no visible diaeresis.
ORNAMENTATION. Carapace and pleonites and fi rst pereiopod propodus covered with coarse tubercles. Garassino, Teruzzi & Dalla Vecchia, 1996. Please note that due to the preservation conditions of specimens, the precise outline of the carapace and shape of grooves are diffi cult to assess. Image hand-drawn by D. Audo.

Comments
Rosenfeldia triasica is known from 39 specimens. Most of these specimens are incompletely preserved, only appendages and pleon are often clearly visible. Among specimens which preserve the carapace, some appear to have a carapace that is longer than wide, such as in the holotype ( Fig. 1A-B), others exhibit a very wide carapace, such as in specimen MFSNgp 40312 ( Fig. 1C-D). As a result, R. triasica was reconstructed as very wide by Garassino et al. (1996: fi gs 6, 8A-B). Indeed, Garassino et al. probably considered the more abundant wide specimens to be typical. Alternatively, Audo (2016: fi g. 3M) and Audo et al. (2017: fi g. 3M), based upon the holotype only, represented R. triasica as narrower. Our reinvestigation of all available material leads us to affi rm these reconstructions. Indeed, the specimens with an apparently wide carapace are undoubtedly exuvia. Their carapaces are split open ( Fig. 1C-D) and consequently seem superfi cially wider than they were (see also Audo 2016 for a short discussion of moulting in Polychelida). For this reason, we disagree with the reconstruction of R. triasica with a very wide carapace, proposed by Garassino et al. (1996). On the other hand, reconstructions by Audo (2016) and Audo et al. (2017) represent the carapace with indistinct cervical and postcervical incisions, which are, however, clearly represented by Garassino et al. (1996) and visible on some specimens ( Fig. 1C-D). Clearly, these reconstructions, based only upon the holotype, which does not preserve the lateral margins well, are inaccurate. We therefore propose an emended catalogue of the carapaces of all genera of fossil Polychelida, each based on its type species (Fig. 2), and a new, complete reconstruction of R. triasica (Fig. 3), both based on most of the available specimens. However, it should be noted that grooves on a carapace are rarely preserved, so their shape on the reconstruction is based on available information and comparison with other polychelidans.

Diagnosis
Extremely thin exoskeleton; ovoid, slightly pear-shaped cephalothorax longer than wide, wider in its anterior half; very narrow frontal margin; small, shallow cervical and postcervical incisions; cervical groove strongly marked near cervical incision and near lateral margin, shallower between these two AUDO D. et al., Rosenfeldia triasica and Rogeryon oppeli gen. et comb. nov.
areas; postcervical groove strongly marked from postcervical incision toward median line, longer than ⅓ of carapace width; telson and uropods rounded; eyes with hexagonal ommatidia; median margin of third maxilliped ischium with a few proximal large teeth and numerous more distal small teeth (serrated aspect), fi rst pereiopods about as large as succeeding ones.

Etymology
Dedicated to Roger Frattigiani, an enthusiast amateur fossil collector who co-operates with researchers since several years. Gender of the genus is masculine.

Remarks
Rogeryon gen. nov. is represented by a single, very rare species: Eryon oppeli Woodward, 1866, from the Late Jurassic of Germany. The study of this species is complicated due to the preservation of many specimens: (1) the holotype does not preserve its carapace, but can fortunately be reliably associated with other specimens by the peculiar shape of tail fan and pereiopods; (2) most known specimens have a very thin carapace that was variably deformed; (3) some specimens (SMNS 66004/2, SMNS 70102), whose carapaces are divided across median line, probably correspond to exuvia (see Audo 2016). We remark that the thinness of the carapace is unlikely directly linked to the preservation of exuviae in the case of some specimens. Exuviae occur abundantly in Solnhofen-type outcrops (Schweigert 2007a), but in polychelidans they do not present a particularly thin carapace, not as thin as that of E. oppeli or Knebelia bilobata (Münster, 1839) (see Audo et al. 2014b).
Moreover, the redescription of Eryon oppeli allows the recognition of several characters almost unique to this species among other polychelidans: (1) an extremely thin exoskeleton, reminiscent of the thinness of the exoskeleton of Knebelia bilobata, K. schuberti (Meyer, 1836) and that of some extant species, but unlike that of other fossil species, which are distinctly thicker; (2) an ovoid, almost pear-shaped exoskeleton wider in its anterior half, distinct from other polychelidans (clearly ovoid, subcircular or pear-shaped, larger in posterior half in almost all other species); (3) a rounded telson, as in Rosenfeldia, rare in polychelidans (generally triangular); (4) a fi rst pereiopod about as large as succeeding ones, generally larger in other polychelidans, except in the poorly preserved Wrangelleryon perrates Feldmann, Schweitzer & Haggart, 2013;(5) fi nally, as indicated by Audo et al. (2016), Eryon oppeli possesses hexagonal ommatidia, contrary to all other polychelidans, for which preserved quadratic ommatidia were documented. Considering only these points, E. oppeli could resemble W. perrates, but these two species do not seem to be closely related as they differ in many aspects: W. perrates is narrower, has European Journal of Taxonomy 367: 1-23 (2017) petaloid uropodal endopod and exopod, and its P5 is larger in proportion. Their resemblance is therefore superfi cial and possibly linked to similar constrains on the shape of their pereiopods.
For these reasons, we consider that Eryon oppeli cannot be ascribed to any currently described genera of Polychelida and we propose the erection of Rogeryon to accommodate it.
The familial assignment of Rogeryon is not straightforward. Indeed, its morphology is very different from all other polychelidans. Its relatively wide carapace and the lack of a diaeresis on uropodal exopod could suggest an assignment to Eryonidae; however, both these characters occur regularly within polychelidans. Furthermore, its hemicircular Mxp3 is clearly different from the subtriangular one of Eryonidae (and narrow one of Polychelidae

Type age
Late Jurassic, Early Tithonian, Hybonotum ammonite biozone (after Schweigert 2007b). European Journal of Taxonomy 367: 1-23 (2017) Description SHAPE OF CARAPACE. Extremely thin exoskeleton; pear-shaped cephalothorax in outline in dorsal view, longer than wide, wider in its anterior half; very narrow and rather shallow frontal margin; side of frontal margin (anterolateral angle) forming a long spine; smooth lateral margin cut by ocular, cervical and postcervical incisions; deep ocular incision, hemicircular, opening laterally and closed ventrally; long anterolateral margin, almost straight, oblique compared to longitudinal axis; small and shallow cervical and postcervical incisions, both opening in the anterior half of lateral margin; straight mediolateral margin, distinctly shorter than anterolateral margin; triangular posterolateral angle, projecting along s1; posterior margin concave, distinctly wider than anterior margin.

Diagnosis
CARAPACE GROOVES AND CARINAE. Cervical groove extending from cervical incision, strongly marked medially and near lateral margin, probably cutting median line, and medially shallower between these two areas; postcervical groove longer than one third of carapace width, not connected to cervical groove, strongly marked from postcervical incision toward median line; shallow branchiocardiac groove extending obliquely in the posterior half of carapace; dorsal carinae (branchial, postorbital, median carinae) not visible; raised submarginal carina, extending ventrally parallel to lateral margin in the fi rst half of carapace. Fig. 6. Reconstruction of Rogeryon oppeli (Woodward, 1866). A. Dorsal aspect. B. Known parts of the ventral aspect, thoracic appendages omitted on the right side. Please note that due to the softness of the carapace, its precise outline is uncertain. Image hand-drawn by D. Audo.
European Journal of Taxonomy 367: 1-23 (2017) PLEON AND TELSON. Pleon slightly shorter and one-half width of carapace; pleonite 1 poorly preserved, shorter than others; s2-s6 with terga crossed by two transverse grooves converging medially and by an axial carina placed posteriorly to anterior transverse groove and cutting posterior transversal groove to join with posterior margin; s3-s5 tergopleura with concave lateral margin, which appears hooked; s2-s5 tergopleura with a small anterior process articulating with a small notch in the preceding pleonite; sharp s6 tergopleuron; telson rounded distally, with fi nely serrated margins, a pair of serrated longitudinal carinae and, possibly, a median carina.
EYE AND CEPHALIC APPENDAGES. Large spherical eye with hexagonal ommatidia; mandible with a massive triangular coxal body and a subtriangular incisor process, the edge of which forms several teeth (poorly preserved); other cephalic appendages too poorly preserved to be described.
THORACIC APPENDAGES. Third maxilliped with a hemicircular ischium and short following podomeres; median margin of third maxilliped ischium inner margin serrated with numerous thin spines (crista dentata); stocky pereiopods, decreasing in size from the fi rst to the last, with slightly stocky and curved dactylus and pollex; P1 only slightly larger than P2.
PLEONAL APPENDAGES. Elongated petasma (modifi ed pleopods 1 of male), poorly preserved; uropods with a short basipodite carrying a rounded endopod and a rounded exopod with a small spine on outer margin; uropodal endopod and exopod both strengthened by a longitudinal carina; uropodal exopod with no visible diaeresis.

Occurrence
Early Tithonian of southern Germany.

Comments
All available specimens seem to differ from each other by the shape of their carapace. The shapes of both sides of the carapaces do not often match. These differences are undoubtedly due to the thinness of the carapace, which was deformed easily during fossilisation. In addition, due to its carapace which is probably divided across its median line, the specimen SMNS 70102 (Fig. 4D) is likely an exuvia; it shows a deformation possibly consistent with that interpretation. Due to these constraints, we base our reconstruction (Figs 2M, 6) of the outline of the carapace on specimens that seemed the less deformed such as SMNS 65545 (Fig. 4E) and SMNS 66004 (Fig. 5).

Palaeobiology of Rosenfeldia triasica
Contrary to polychelidans which might have been more nektobenthic such as Willemoesiocaris ovalis (Van Straelen, 1923) (Audo et al. 2014c) or well-adapted to swim (Audo et al. 2014b), Rosenfeldia triasica seems to possess a rather thick carapace and stout pereiopods. It was therefore undoubtedly mostly benthic.
Rosenfeldia triasica possesses rather stout P1. Contrary to many polychelidans with long slender claws, which are probably ambush predators, being able to deploy their fi rst pereiopods to catch preys (Ahyong 2009), R. triasica probably had to move close to its food or prey to catch it. It is therefore likely that R. triasica fed on slow moving and sedentary preys and / or was a scavenger. The redescription of R. triasica permitted us to identify for the fi rst time a mandible molar process in a fossil Polychelida (Fig. 1D). Extant Polychelidae do not possess this character and thus tend to swallow food items completely (Cartes & Abelló 1992). As a consequence, R. triasica, compared to extant Polychelidae, was probably more able to eat large or resistant food items.  argued that polychelidans could feed on molluscs as palinurids do. Although we consider that polychelidans are probably not closely allied to palinurids and scyllarids (see Scholtz & Richter 1995;Bracken-Grissom et al. 2014 and references therein). It is possible that molluscs were part of the diet of Rosenfeldia triasica, perhaps more so than in the diet of other polychelidans, especially extant ones, whose mandibles seem to lack a molar process.

Palaeobiology of Rogeryon oppeli gen. et comb. nov.
Rogeryon oppeli gen. et comb. nov. possesses an unusual morphology among polychelidan lobsters: (1) a very thin, probably soft, carapace, (2) eyes with hexagonal ommatidia (probably apposition eyes: Audo et al. 2016), (3) P1 almost as large as P2-P4. Since none of these traits are common among polychelidan lobsters, most of them are likely autapomorphies (unique to Rogeryon oppeli) and unlikely to be ancestral characters retained in Rogeryon oppeli, but lost in most or all other polychelidans. As such, replacing R. oppeli in its palaeoenvironmental context and comparison of its traits to that of other polychelidans and arthropods might shed some light on its possible life-habits.

Palaeoenvironment
Rogeryon oppeli gen. et comb. nov. was discovered in the early Tithonian outcrops of southern Germany. These outcrops are considered to have been formed in relatively shallow water (20 to 60 m) in a subtropical, probably semi-arid zone (Barthel et al. 1990;Keupp et al. 2007). These outcrops preserve a diverse marine fauna and some terrestrial taxa, including tetrapods, insects and plants. In this fauna, Rogeryon oppeli is one of the rarest species: only 7 specimens are known to us, including specimens in private collections. Preservation of the specimens is generally excellent. Most specimens, except perhaps SMNS 65545 (Fig. 5D) and SMNS 66004 ( Fig. 5A-C), seem to correspond to exuviae, the carapace either missing or partially disarticulated. It seems unlikely that these specimens were transported over a large distance. Indeed, the entire animal had a thin carapace and its exuviae were undoubtedly quite fragile, since they do not contain muscle and tissue limiting the disarticulation. Therefore, R. oppeli probably lived in a shallow water palaeoenvironment.

Vision
The eyes of Rogeryon oppeli gen. et comb. nov. exhibit hexagonal ommatidia. These hexagonal ommatidia suggest R. oppeli had apposition eyes . Apposition eyes provide a greater resolution but at the expense of poorer performance in dim-light conditions compared to superposition eyes (see Cronin & Porter 2008 for a full review of eye-types in crustaceans). Since most other polychelidans seem to have had superposition eyes , it is possible that the apposition eyes of R. oppeli correspond to an adaptation to a shallow water palaeoenvironment. This character seems congruent with the shallow water palaeoenvironments attributed to the Solnhofen-type outcrops. However, it should be noted that other polychelidans from Solnhofen-type outcrops with preserved ommatidia, had refl ection superposition eyes (quadrate ommatidia: Audo et al. 2016). Therefore, the palaeoenvironment alone cannot explain the eye-type of R. oppeli. As a specialisation, or ecological differentiation, R. oppeli could have been diurnal, while other polychelidans were more active when light declined or during the night, as several extant subtidal species are (Moller & Naylor 1980;Kanciruk 1980;De Grave & Turner 1997). Unfortunately, due to the life-habits and reduced eyes of extant polychelidans, it might not be possible to draw stronger conclusion from this particularity of R. oppeli.
European Journal of Taxonomy 367: 1-23 (2017) While so far we have considered the ultimate factors for the presence of hexagonal ommatidia in R. oppeli gen. et comb. nov., we can also address the proximate ones. Quadrate facets (and with this refl ective superposition eyes) are widespread among adult decapods, larvae usually possess apposition eyes with hexagonal facets. It seems, therefore, likely that the presence of hexagonal facets and a supposed apposition eye is a simple retention of a larval condition into the adult phase, hence most likely a case of paedomorphic heterochrony (evolutionary shift in developmental timing). Unfortunately, our knowledge of the exact timing of this transition is rather limited in general and, due to the reduced eyes in modern forms, in polychelidans especially. The only case in which eyes are preserved in a fossil polychelidan larva is an eryoneicus type-larva with quadrate facets preserved (Haug et al. 2015). Yet, the apposition eye might more likely not be coupled to larval decapods in general, but to zoea-type larvae, and the later eryoneicus (and the eryoneicus-type) larvae are megalopa-type larvae.
Given the size of the eyes in relation to the body and number of ommatidia in R. oppeli gen. et comb. nov. we can furthermore suggest that the mode of paedomorphosis is neoteny. The other types of paedomorphosis (progenesis and post-displacement) would have led to comparably smaller eyes.
Also in other instances within Polychelida paedomorphosis has been suggested, e.g., for species of Knebelia Van Straelen, 1922(Audo et al. 2014b). Yet, in Knebelia other structures were affected than in R. oppeli gen. et comb. nov., leading to quite different ecological adaptations (see also further below). This demonstrates that heterochrony played an important role in the diversifi cation of Polychelida.

Relation to substrate
The thin exoskeleton of R. oppeli gen. et comb. nov. is reminiscent of that of Knebelia bilobata, which was possibly well-adapted for swimming (Audo et al. 2014b). In contrast to K. bilobata, R. oppeli does not seem to have had particularly developed uropods: the mean ratio between uropod length and carapace total length is only 0.21 (see method of measurement in Audo et al. 2014b: measurements on specimens SMNS 65545, SMNS 66004, SMNS 70102). Due to the small size of the sample, this comparison cannot be tested statistically. If R. oppeli did not actively swim, it may have fl oated. Extant eryoneicus larvae and juveniles of Palaeopentacheles roettenbacheri (Münster, 1839) seem to use their abundant spines to provide additional buoyancy (see Gurney 1942;Eiler et al. 2016). However, the exoskeleton of R. oppeli appears devoid of spines, except perhaps very small ones. Therefore, it seems unlikely that R. oppeli was well adapted to fl oat in the water column. Besides, the structure of the pereiopods in R. oppeli is more consistent with benthic life-habits: indeed, its pereiopods seem more massive than those of Knebelia bilobata¸ Willemoesiocaris ovalis (Van Straelen, 1923) (Audo et al. 2014c) and even that of extant Polychelidae. Considering the relatively small uropods and massive pereiopods, a thin carapace does not seem suffi cient to propose a swimming life-habit for Rogeryon oppeli. The thinness of carapace could be: (1) an adaptation to its life-style, but not linked to improbable developed swimming capacities; (2) or be plesiomorphic (that is being primitive for R. oppeli and its closest relative), which is unlikely as within all polychelidans, only Knebelia bilobata and perhaps Wrangelleryon perates seem to have a thin carapace; (3) or, be linked to the evolution of other characters. This third possibility is in our opinion plausible. Indeed, we remark that the carapace thinness may be the result of R. oppeli having evolved through neoteny (see Audo et al. 2014b). This assertion would also be supported by the apposition eyes in R. oppeli (see Cronin & Porter 2008), and its relatively wide carapace compared to the pleon (see Bravi et al. 2014). Despite its thin carapace, Rogeryon oppeli seems, therefore, to have been a benthic animal.
As a benthic animal, Rogeryon oppeli gen. et comb. nov. does not seem to present adaptations to bury itself in sediments: (1) although antennulae and antennae are poorly preserved, the preserved parts do not seem to be enlarged to form a respiratory canal, unlike those known in polychelids (Ahyong 2009) and some eryonids (Audo et al. 2014a); (2)  suited for burrowing: it is very thin, possibly soft, wider than the pleon (see Ahyong 2009; note that this last character probably does not preclude burying, see Audo et al. 2014a); (3) the eyes of Rogeryon oppeli do not seem adapted to burying, indeed, they are neither surrounded by ocular incision as in many eryonids nor raised on stalks as in Eryon. Taking into account all these characters, it seems likely that Rogeryon oppeli was an epibenthic animal.

Diet
The fi rst pereiopods of R. oppeli gen. et comb. nov. are almost as large as the succeeding pereiopods, contrary to other polychelidans, where the P1 is generally larger than the succeeding ones and generally held folded against the carapace sides so that they can strike prey from a buried position (Ahyong 2009). Contrary to the preceding characters, a short P1 does not seem to be linked to paedomorphosis since eryoneicus larvae (Bernard 1953;Eiler et al. 2016), and eryoneicus-like larvae (Haug et al. 2015) do have a larger P1. The size of pereiopods is therefore likely closely linked to the palaeoecology of Rogeryon gen. nov. The P1 probably has a predatory role in both extinct (Garassino et al. 2012) and modern polychelidans (Gore 1984) as well as a defensive role (Santucci 1933). Gut content may yield information on the diet of fossil polychelidans. However, even in exceptionally well-preserved specimens, food items were not recognizable (Jauvion et al. 2016). Gut contents of modern polychelidans have shown they are either predators of small invertebrates or scavengers (Firth & Pequegnat 1971;Lagardère 1973;Gore 1984;Cartes & Abelló 1992). With its relatively short P1, R. oppeli was probably not able to catch preys at a distance by reaching out to them.
Such a disposition is in fact reminiscent of that of slipper lobsters which also have subequal pereiopods, notably P1 (pers. obs. D.A.). Feeding habits of slipper lobsters are generally considered to consist mostly of epifaunal and endofaunal molluscs and worms, as well as scavenging (Suthers & Anderson 1981;Johnston & Alexander 1999;Johnston 2007). Slipper lobsters seem particularly suited to open bivalve shells (Johnston 2007). Rogeryon oppeli might have had a similar diet, especially since extant polychelidan lobsters prey on small invertebrates; it might have dug them up from a soft substrate. Slipper lobsters, however, are achelate and possess strong dactyli that can be used as wedges to open bivalve shells. The claws of R. oppeli were probably not as effective to open shells; however, it might have fed on other epifaunal and infaunal animals, including thin-shelled bivalves, as is the case for extant horseshoe crabs (xiphosuran chelicerates), which also have four chelate appendages. Horseshoe crabs also have chelicerae and gnathobases to reduce the food item they catch (Walls et al. 2002 and references therein), but the function of these parts is performed by mandibles, maxillulae, maxillae and three pairs of maxillipeds in decapod crustaceans.
It seems therefore likely that R. oppeli gen. et comb. nov. had a diet somewhat similar to that of slipper lobsters and horseshoe crabs. Besides, prey such as polychaete worms and bivalves were available in the environment, since fossilized specimens have been discovered in the Solnhofen-type outcrops (e.g., Frickhinger 1994Frickhinger , 1999. It could be noted that another species might have had a similar diet: Wrangelleryon perrates, which also possesses similar subequal pereiopods.

Conclusion
Rogeryon oppeli gen. et comb. nov. is an unusual species of polychelidan lobster. Its inclusion in Rosenfeldia overlooked many of its unique characteristics. Although the reconstruction of its palaeobiology relies mostly on comparison with distant modern taxa, it is still obvious that it had lifehabits unique among polychelidan lobsters.
During the Late Jurassic, the polychelidan diversity reached a peak (Audo et al. 2017). Although Solnhofen-type outcrops are clearly favourable to the preservation of crustacean remains, the occurrences of unusual polychelidan lobsters such as Rogeryon gen. nov. or Knebelia suggest that these settings also European Journal of Taxonomy 367: 1-23 (2017) represented an opportunity for crustaceans to diversify and exploit new environments and resources. The disappearance of these palaeoenvironments at the end of the Jurassic may be one of the factors leading to the decline of polychelidan lobsters immediately after this important diversifi cation.