Lithophylax trigeri A. Milne-Edwards & Brocchi, 1879

Guinot, Danièle & Breton, Gérard, 2006, Lithophylax trigeri A. Milne-Edwards & Brocchi, 1879 from the French Cretaceous (Cenomanian) and placement of the family Lithophylacidae Van Straelen, 1936 (Crustacea, Decapoda, Brachyura), Geodiversitas 28 (4), pp. 591-633 : 602-622

publication ID	https://doi.org/ 10.5281/zenodo.4665480
persistent identifier	https://treatment.plazi.org/id/03F52361-FFB6-114D-FD55-FC30FC86C51C
treatment provided by	Felipe
scientific name	Lithophylax trigeri A. Milne-Edwards & Brocchi, 1879
status

Treatment

Lithophylax trigeri A. Milne-Edwards & Brocchi, 1879 ( Figs 5-14 View FIG View FIG View FIG View FIG View FIG View FIG View FIG View FIG View FIG View FIG )

Lithophylax Trigeri A. Milne-Edwards & Brocchi, 1879: 117 . — Guillier 1886: 238 [and not 338], 244.

Palaeoplax Trigeri View in CoL – Guillier 1869: 32; 1886: 238 (and not 338) (Non Palaeoplax A. Milne-Edwards & Brocchi, 1879 ; type species: Ocypode incerta Desmarest, 1819 ).

Petrocarcinus Trigeri – Guillier 1869: 32; 1886: 244 (nomen nudum of A. Milne-Edwards).

Lithophylax trigeri – Glaessner 1929: 236, 431; 1969: R514. — Rathbun 1935: 52. — Van Straelen 1936: 43, pl. 4, fig. 9. — Juignet 1974: 669. — Vega et al. 1997: 619.

TYPE MATERIAL. — Despite Rathbun’s (1935: 52, and footnote) statement: “The holotype is not to be found in the Paris Museum, according to M. Boule et C. Gravier”, we found in the paleontological collections of the MNHN 36 specimens from Le Mans and Saint-Mars-sous-Ballon (Sarthe), registered as B16566 (3 specimens), B16567 (3 specimens), B16576 (9 specimens), B16577 (3 specimens), and B16595 (18 specimens) which very probably consitute the orginal material (of both sexes) studied by A. Milne-Edwards & Brocchi (1879), i.e. the type series. The topotype figured by Van Straelen (1936: pl. 4, fig. 9) has not been found; the specimen of the Bosquet’s collection ( RBINSB) is not the topotype figured by Van Straelen (1936).

Amongst the syntypes we select from B16566 one lectotype which receives the new number MNHN A25835 (original label “Cénomanien St Mars-sous-Ballon (Sarthe) Coll. A. Milne-Edwards, 1902-3”. Other label “Van Straelen determ. 1938”). We were not able to include the material of the type series in our Table 1 (Annexe) because of its very late re-discovery.

ADDITIONAL MATERIAL. — See Material examined and prepared, and Annexe: Table 1.

DESCRIPTION

Size moderate (maximum width 28 mm). Carapace much wider than long, transversally hexagonal to inverted trapezoidal, widest at position of the outerorbital teeth; length about 0.53 to 0.59 maximum carapace width in average. Dorsal carapace obviously lobulate; regions well distinct throughout surface. Rather deep grooves clearly differentiating regions, in particular the deeply impressed cervical, branchio-cardiac and gastro-cardiac grooves. A strong H-shaped mark on each side ( Figs 5 View FIG ; 6 View FIG ; 14 View FIG ), and a “human face” sometimes depicted (formed by attractor epimeralis muscle apodeme, posterior gastric muscle scar, median part of cervical groove, and branchio-cardiac groove).Two deep gastric pits along cervical groove.

Areolation ( Fig. 5 View FIG ) as follows: mesogastric (funnel-shaped and extending forward almost to the frontal) and metagastric regions well defined and forming an undivided, single plate; urogastric region elongated transversally, arched, crescent-shaped, separated from metagastric region by deep furrow; protogastric regions large, expanded laterally, may be crossed by elevated transverse ridge (that marks the maximum height of the carapace), with two more or less marked nodes on each side; hepatic region with two unequal lobes: anteriorly, a prominent, rounded bump; posteriorly, a narrow, elongated ridge reaching carapace lateral margin and separated from branchial region by shallow groove; paragastric region (or branchial lobe) semicircular, well differentiated, prominent; cardiac region with two elevations; epibranchial region with salient internal lobe and external marked ridge, posterior to hepatic swelling; mesobranchial region broad, consisting of narrow, salient internal lobe and inflated lateral part (may be ridged on three sides) with pronounced slope along carapace postero-lateral margin; metabranchial region depressed; intestinal region poorly separated and depressed posteriorly. Ornamentation of dense but fine granules. No trace of punctuations.

Lateral margins ( Fig. 6 View FIG ) strongly diverging anteriorly, converging posteriorly; junction of antero- and postero-lateral margins weakly marked; antero-lateral margins slanted, without tooth other than marked outer-orbital tooth, only followed by a blunt angle; postero-lateral regions steeply inclined, very oblique, with sharply defined margins; postero-lateral corners rounded; posterior margin rimmed, nearly straight to weakly concave centrally in dorsal view.

Rostrum ( Figs 5 View FIG ; 7 View FIG ; 10A View FIG ; 14 View FIG ) forming narrow, rather long, ungrooved process, inclined downwards, its distal half deflected at 90° to dorsal carapace surface; its extremity bluntly rounded. Frontal margin forming same line with supra-orbital margin, entire, except for small fissure at mid-part of supra-orbital margin; frontal margin lined by coalescent granules; supra-orbital margin lined by tubercules twice as long as frontal granules.

Orbits extremely deep, wide, long, extending from rostrum to outer-orbital tooth (total 80% carapace width; 55% for podophthalmite and corneal surface), directed obliquely rearwards, divided into two unequal fossae by small furrow, inner fossa being cylindrical, outer one being larger, elliptical and extending back to center of hepatic region; orbital margins very long, extending to outer-orbital teeth carapace corners. Infra-orbital margin lined by tubercules smaller than those of supra-orbital margin.

A pair of thick, stout, cylindrical and mobile eyestalks; basophthalmite ( Figs 5 View FIG ; 10A View FIG ; 14 View FIG ) present in four specimens, erected, their insertion close to base of rostrum clearly visible; rest of eye (podophthalmite and corneal surface) not present, interpreted however as sufficiently developed to fold back into orbit and to completely fill it; presumably inflated corneal surface probably lodged in external part of orbit ( Figs 5 View FIG ; 7 View FIG ; 8B View FIG ; 10A View FIG ; 14 View FIG ). Antennae and antennules not present, nor their cavities.

Below outer orbital fossa, suborbital region granular. Presence of a special plate, herein named “shutter” consisting of two parts: a quadrangular internal part, delineated by a slightly granulated to smooth rim, with marked right angle margins; and an external fan-like process ( Figs 7A, B View FIG ; 8A, C View FIG ). Relationships of the “shutter” with suborbital and antero-lateral margins, as well as pterygostomial region, remaining unclear. Pleural line ventrally located. Buccal cavern very wide. Presence of extremely strong mandibles. Mxp3 clearly diverging, directed obliquely, leaving large gap between them. Endopodite with coxa expanding laterally; ischion rectangular; merus widened and broadly rounded at postero-external angle; palp not preserved or barely preserved; exopodite well developed.

Thoracic sternum complete, wide, oval, flat, but last sternites 7 and 8 inclined; gynglymes of mxp3 and pereopods well visible. Suture 1/2 marked, complete; sternites 1 and 2 not fused, forming narrrow, relatively long and downward inclined triangle. Suture 2/3 marked. Suture 3/4 not present and replaced by sinuous ridge; sternite 3 wide, much more depressed than sternites 4 and 5. Sternite 4 well developed. Suture 4/5 extending along from the margins before turning sharply inwards, therefore interrupted axially; its extremities separated by rather wide gap and marked by deep depressions corresponding to the invagination of the two plates which form each phragma. Sutures 5/6 and 6/7 showing deep depressions, as suture 4/5; suture 5/6 (seen in four individuals) prolonging beyond depressions and thus (presumably) not interrupted axially. Suture 6/7 (weakly exposed in our material) not interrupted axially. Suture 7/8 incomplete, extremely short and thus only lateral, ending by small depressions. Sternites 4-6 normally exposed; most of sternite 7 covered by abdomen in both sexes; sternite 8 much reduced and subdorsal, covered by abdomen in both sexes. Episternites 4-6 well defined. Presence in both sexes of elongated cupules hollowed on sternites 4 to 6 on posterior margin, internal to episternites. A longitudinal median line present on sternite 7. Surface smooth, punctate.

Sterno-abdominal cavity triangular, extending the mid-length of sternite 4, not deeply excavated, without sharply delimited margins. Vulvae present on sternite 6, relatively far from each other, on oblique flanks of sterno-abdominal cavity.

Male abdomen with all somites free and telson, lodged in the cavity (except for its extreme anterior part), relatively wide, not much narrower than female one (weak sexual dimorphism), regularly triangular; its base broad; first abdominal somites covering completely space between coxae of last pereopods. A4 elongated transversally and joining level of P3 coxae. Carapace covering first abdominal somites, at least A1 (in some cases A2 and portion of A3 apparently covered by carapace, Fig. 9E View FIG ). A1-A4 ridged transversally. Telson triangular. Abdominal surface punctate, posterior axial part of each segment smooth.

Male gonopods preserved in 14 specimens at least (Annexe: Table 1); G1, or G2 or both cleared in three of these. G1 long, slender, each forming regular curve; punctuations being bases of setae. G2 relatively long, 1.4 times shorter than G1, with wide base, angled at distal fourth; tip pointed; G2 may be found in an horizontal position close to A1 or inserted inside G1.

Abdominal devices of press-button type: locking structures showing as developed and pointed prominences situated on sternite 5 about at its mid-part (but closer to suture 4/5 than to suture 5/6), remaining present and most probably effective in mature females; sockets (not seen) likely on abdominal segment 6, in postero-lateral angles.

All pereopods with their condyli often preserved (narrower for P1), their gynglymes often well visible. Chelipeds robust, with massive propodus, clearly heterochelous, right hand being distinctly larger (right hand varying from 10% to 25% wider than left, and 12-15% longer than left); palm outer surface not keeled but thickened in superior half; inferior border flat or slightly convex, not ridged; surface minutely granulate, without spines on borders. Fingers elongated, with surface minutely granulate; heterodonty rather weak but distinct, several teeth on both prehensile margins, slightly thicker at right.

Arthrodial cavities of P1-P5 not aligned, P4 and P5 being progressively nearer to the mid-line, P5 coxa being completely subdorsal ( Figs 11 View FIG ; 12B View FIG ). P2-P4 very long, thick; merus wide, elongated, markedly compressed dorso-ventrally, flattened ( Fig. 12D View FIG ), distal part more slender than proximal, with antero-dorsal carina lined by marked tubercles; surface minutely granulated; punctate traces on posterior margin may indicate existence of setae; carpus rarely preserved (only proximal part on P3 and P4); propodus and dactylus not preserved. Autotomy line visible on most pereopods, just distal to ischio-basis limit ( Fig. 12C View FIG ). P5 (only coxa and merus preserved) obviously reduced and subdorsal ( Figs 11 View FIG ; 12B View FIG ; 14 View FIG ); coxa very small, diameter and length half that P1-P4 coxae (Annexe: Table 2), and with male gonopore visible ( Fig. 11A View FIG , below); merus (present in only one specimen), presumably cylindrical, narrow, with a main distinctive row of pointing granules and another one with weaker granules. Right and left P5 coxae closer to each other than P4 coxae and more widely separated than P3 and P2 coxae.

Stridulatory apparatus in both sexes, consisting of a well specialised pars stridens, with about marked 29 striae on flattened area at inner surface of merus of chelipeds; and a plectrum on the suborbital region, represented by at least eight blunt tubercles. In one specimen (MHNH B 9199) ( Fig. 13 View FIG ) counterpart of five striae of the pars stridens situated in front of both suborbital blunt tubercles and tiny granules on posterior rim of quadrangular part of shutter.

See the reconstruction of dorsal carapace Figure 14. View FIG

REMARKS

Lithophylax trigeri was originally based by A. Milne- Edwards & Brocchi (1879) on material collected in the Butte de Gazonfier quarry, at Le Mans (Sarthe), Grès verts du Maine (or Sables à Rhynchonella compressa = Sables du Perche Fm ), middle-upper Cenomanian, Jukesbrownei and Guerangeri Biozones.

Most of the material studied herein comes from the Louvre quarry, at Lamnay (Sarthe), Sables et Grès de Lamnay Formation , lower Cenomanian , Mantelli (Saxbii) and Dixoni biozones. Thus this material is 2-3 ammonites biozones (c. 2 my) older than the original material (not seen) and the Guéranger’s topotypic specimens that we were able to examine. A few differences have been observed between the lower Cenomanian ( LC) and the middle-upper ( MUC) Cenomanian material. The carapace of MUC material is slightly larger and wider (average L/ W 0.53 for MUC and 0.59 for LC) (Annexe: Table 1) ; chelipeds of MUC material are more sturdy; areolation and granulation of the dorsal carapace of MUC material is more marked. Unfortunately, no new MUC material can be collected now, and the 19th century material is not located precisely enough to allow confident statistical measurements. Regardless, the differences between the MUC and the LC material are weak, and they do not seem sufficient, at this stage of our work, to distinguish two different species – or even subspecies – in the present paper.

REMARKS ON THE STRIDULATING APPARATUS

A Lithophylax specimen exhibits the counterpart of the pars stridens in situ ( Fig. 13 View FIG ), exactly just ventral to the tubercles of the plectrum: striae of the pars stridens are exactly perpendicular to the row of eight tubercles located posteriorly to the shutter ( Figs 8A, C View FIG ; 12B View FIG ). A similar mechanism to that of Lithophylax roughly involves the same areas in the Recent goneplacid Bathyplax typhlus A. Milne-Edwards, 1880 . In B. typhlus (see Tavares 1996: 416, fig. 3A, B) the pars stridens consists of a narrow area of transversal striae (about 40) on the inner face of P1 merus, and the plectrum consists of about 50 non-aligned, subhepatic granules: among these 50 granules, only those of a medial row appear to be blunt, thus are probably the only efficient ones. In Lithophylax trigeri the pars stridens is shorter and wider (29 striae), and the plectrum is a row of eight spaced tubercles. The anterior part of the pars stridens is supposed to be also in contact with the posterior rim of the shutter. This rim bears a minute granulation.We suggest, as an hypothesis, that this layout permitted a two-frequencies “diphonic” stridulation with a low frequency from the suborbital tubercles and a higher frequency from the shutter granules.

The stridulating apparatus in the Recent and fossil goneplacid Ommatocarcinus White, 1851 (not present in all species of the genus) (see Guinot- Dumortier & Dumortier 1960; Jenkins 1975), with striated suborbital ridge and crest on P1 merus, is different from the mechanism in Lithophylax . The goneplacid genus Psopheticus Wood-Mason, 1892 (see Guinot 1990) contains also stridulating species, but with tubercles on P1 merus (instead of striae in L. trigeri ) and pterygostomial crest and/or granules (tubercles in L. trigeri ).

Another case of friction of the P1 merus against the subhepatic region of the carapace is known in the Recent gecarcinid Gecarcinus quadratus Saussure, 1853 , but both parts consist of granules and/or tubercules ( Abele et al. 1973: 148, fig. 1). The rarity of known mechanisms for sound production in fossil crabs is explainable by our incomplete knowledge, and it is probable that other extinct species were able to produce sound. Examples of stridulating species are the Miocene Szaboa inermis (Brocchi, 1883) (= Matuta brocchii Glaessner, 1969 ) ( Müller 1984: 69; Müller & Galil 1998: fig. 2), Ommatocarcinus species (see Jenkins 1975: 36, 48, pl. 4, fig. 6, pl. 7, fig. 2c), and Stevea cesarii Beschin, Busulini, De Angeli & Tessier, 1994 from the Eocene of Italy (A. De Angeli, pers. comm.). The Maastrichtian Megaxantho zoque Vega, Feldmann, García-Barrera, Filkorn, Pimentel & Avendaño, 2001 shows striae on the inner surface of the distal portion of the palm suggesting a “possible stridulatory mechanism” ( Vega et al. 2001: fig. 5.3).

The stridulating apparatus, present in the fossil Hexapodidae (as in extant species), consists either of a prominent ridge of striae (e.g., in Goniocypoda edwardsi Woodward, 1867 and probably also in G. quaylei Crane, 1981 , both from the Upper Eocene of Hampshire, see Crane 1981: 6, 7, fig. 8D; and in Stevea cesarii ) or an area of pterygostomian striae (e.g., in Hexapu s pinfoldi Collins & Morris, 1978 , from Eocene of Pakistan; probably in other fossil Hexapus species). These striated ridges are rubbed by thin, closed striae on the inner surface of the dactylus of both chelipeds ( Guinot 2006).

OTHER BRACHYURAN CRUSTACEANS REPORTED FROM THE STRATOTYPIC CENOMANIAN WITH LITHOPHYLAX TRIGERI

According to Guillier (1886) the other brachyuran crustaceans reported from the stratotypic Cenomanian with Lithophylax trigeri are:

– Raninella trigeri View in CoL and R. elongata View in CoL , species names mentioned by A. Milne-Edwards (1862b: 493) in establishing the genus Raninella View in CoL ; the species were described and figured by Brocchi (1877). They belong to the Raninoidea De Haan, 1839, i.e. to the Podotremata Guinot, 1977.

– Caloxanthus formosus A. Milne-Edwards, 1862 View in CoL (A. Milne-Edwards 1862a: 44, pl. 9, fig. 1), type species of Caloxanthus A. Milne-Edwards, 1862 View in CoL , collected in the “grès verts du Maine by M.Triger”. An upper Santonian material of C. formosus View in CoL from Aude in France was figured by Wright & Collins (1972: 104, pl. 21, fig. 9). A second species of Caloxanthus View in CoL , C. americanus Rathbun, 1935 View in CoL , from the Albian, shows on the ventral surface a triangle inserted between the mxp3, interpreted by Rathbun (1935: 56, pl. 11, fig. 19) as “the terminal segment (female?) of the abdomen” (instead of, in our interpretation, the first thoracic sternites forming a narrow and triangular plate, inserted between the mxp3). Wright & Collins (1972: 56, 103) have suggested a derivation of Caloxanthus View in CoL from the Lower Cretaceous Diaulax Bell, 1863 View in CoL , “a stock that diverged from Dynomenidae View in CoL in the Later Jurassic” and, however, included Caloxanthus View in CoL in the Carpiliidae Ortmann, 1893 View in CoL . The Early Cretaceous Caloxanthus View in CoL was attributed to the Xanthidae View in CoL s.l. by Schweitzer et al. (2002a: 39, 40, table 4). At present we consider Caloxanthus View in CoL a podotreme crab, and prefer to assign it (with reservation) to the Diaulacidae Wright & Collins, 1972 View in CoL or to an undescribed family within the Podotremata.

Glaessner (1969: R488) included Diaulax View in CoL in the dynomenids, and Schweitzer et al. (2003a: 18, 20) synonymized the Diaulacidae View in CoL with the Dynonemidae Ortmann, 1892 (see also Schweitzer & Feldmann 2005: 38). The P5 dorsal location, mentioned in Diaulax carteriana Bell, 1863 ( Bell 1863: 7) View in CoL , sup- ports the hypothesis of a podotreme condition, similar to that of the Dynomenidae View in CoL . But the most reliable characters (sexual gonopores, ventral surface, and thoracic sternum/male abdomen relationships), allowing the certain assignment of a family (and more generally to distinguish podotreme crabs from heterotreme crabs), are still lacking, and it is why for the moment we prefer to consider the Diaulacidae View in CoL a separate family. According to Wright (1997: 135) re-examination of the crab fauna in Austria removes “the puzzling record of Diaulax View in CoL from the Jurassic” and “shows that it is Cretaceous, Cenomanian”.

We agree with Wright & Collins (1972: 54, 105) that the lower Senonian Creticarcinus Withers, 1928 ( Withers 1928: 461) is very close to Caloxanthus . Graptocarcinus Roemer, 1887 (see Stenzel 1944: 550; Wright & Collins 1972: 54; De Angeli & Garassino 2006: 279, fig. 6), with the Aptian G. texanus Roemer, 1887 and G. muiri Stenzel, 1944 , the Cenomanian G. bellonii Collins & Dieni, 1995 , shows also strong resemblance to Caloxanthus . Both Graptocarcinus and Caloxanthus are likely candidates to be podotreme. The lower Cenomanian Necrocarcinus avicularis Fritsch in Fritsch & Kafka, 1887 ( Fritsch 1887: 47 pro parte, pl. 10, fig. 12 only) might be included in Graptocarcinus Roemer, 1887 (see Glaessner 1929: 261; Wright & Collins 1972: 54, 55, 106; De Angeli & Garassino 2006: 280).

– Necrocarcinus inflatus A. Milne-Edwards , nomen nudum, in Guillier (1886: 244), was figured by Boule & Piveteau (1935: 392, fig. 570), that is not sufficient to validate the name (ICZN: Art. 13.1.1). Thus the author of the species is Van Straelen (1936: 37-39, pl. 4, fig. 8) who provided a long diagnosis and established the new genus Cenomanocarcinus Van Straelen, 1936 View in CoL for this species. This genus, considered invalid by Stenzel (1945: 447) and renamed ( Stenzel 1953: 214) “ Cenomanocarcinus Van Straelen, 1936 View in CoL in Stenzel 1945 ”, was later synonymised with Necrocarcinus Bell, 1863 View in CoL by Wright & Collins (1972: 62). Cenomanocarcinus View in CoL was re-attributed to Van Straelen, 1936 and rehabilitated by Schweitzer et al. (2002a: table 4, fig. 29) and Schweitzer et al. (2003a: 36). Anyway, C. inflatus ( Boule & Piveteau, 1935) belongs to the Necrocarcinidae Förster, 1968 View in CoL , an heterotreme family, appeared in Early Cretaceous (see Discussion).

– Necrocarcinus minutus A. Milne-Edwards , nomen nudum, in Guillier (1886: 244). Status unknown.

DISCUSSION (BY DG)

The abundant and extremely well preserved material at our disposal has allowed the rescue of most parts, so that the fossil crab studied herein is almost as complete as a living species. All characters of Lithophylax trigeri (except antennules, antennae, podophthalmites of eyes, distal articles of the P2- P5, details of G1, and part of the subhepatic area) can be described. The male gonopore is visible on the P5 coxa in two individuals. Vulvae and gonopods were found exposed in several individuals, thus it was possible to be sure of the eubrachyuran nature, and heterotreme condition, of Lithophylax . The sex of individuals was often easily defined. The gonopods were exposed after cleaning of the matrix and partial removal of the abdomen and carapace. Autotomy line was visible on P1-P4 ( Fig. 12C View FIG ) (see Legendre 1908). These characters are generally not preserved or poorly preserved in most fossil crabs, the vulvae being barely known and the gonopods present only exceptionally. Consequently, the Cenomanian material studied in the present paper, with specimens collected by one of us (GB) in particular, represents a precious paleontological document.

In Lithophylax trigeri (see Annexe: Table 1) thoracic sternal sutures show as follows: sutures 1/2 and 2/3 complete; suture 3/4 not marked; suture 4/5 interrupted and ending by deep depressions; sutures 5/6 and 6/7 also with deep depressions but presumably complete; suture 7/8 short, only lateral, and ending by small depressions. Suture 5/6, seen in four individuals, is presumably not interrupted axially; suture 6/7 (less exposed on our material) is not interrupted axially. In L. trigeri the depressions at level of sutures 5/6 and 6/7, similar to those located at the extremities of the suture 4/5, perhaps correspond to the median membraneous areas which are frequent at the point where both extremities of the sutures (especially in the case of suture 6/7) join medially.

At least five patterns of sternal sutures 5/6 to 7/8 are known to exist in extant eubrachyuran crabs. An interrupted suture 4/5 may be followed by: 1) sutures 5/6-7/8 complete; 2) suture 5/6 interrupted but sutures 6/7-7/8 complete; 3) sutures 5/6-6/7 interrupted but suture 7/8 complete; 4) suture 5/6 interrupted, suture 6/7 complete; suture 7/8 interrupted; 5) sutures 5/6-7/8 interrupted (see Guinot 1979: tables 2, 3; Guinot & Richer de Forges 1981). The extremities of the sutures may be sometimes close together (sometimes joining in a median membraneous area) so that they seem to be complete, and misinterpretations are possible. The Lithophylacidae shows another pattern: 6) suture 4/5 followed by complete sutures 5/6-6/7 and interrupted suture 7/8.

Considering the absence of thoracic sternal sutures in most ancient fossils, their condition cannot be used in our comparisons of the Lithophylacidae with extinct eubrachyuran families.

In the previous placements of Lithophylax , the carapace was the only character available to paleontologists, often obliged to take into account the traditional features commonly preserved and described in fossils (“proxy characters” as defined by Schweitzer 2003). Lithophylax , never considered a podotreme crab and finally referred to the Eubrachyura, already appeared at the limit between Lower and Upper Cretaceous. Our observations of the complete animal confirm such a placement. First referred to the Goneplacidae (see A. Milne- Edwards & Brocchi 1879) or “Goneplacidea” (see Van Straelen 1936) and then to the Portunidae Rafinesque, 1815 (see Rathbun 1935; Feldmann & Villamil 2002), Lithophylax was finally assigned, with more or less confidence, to the Carcineretidae (see Glaessner 1969; Bishop 1988; Vega & Feldmann 1991; Vega et al. 1995, 1997; Schweitzer et al. 2002a, 2003a).

The presence and shape of the last pereopod in L. trigeri , not mentioned by the preceding authors, was observed in the present material. Five individuals show a P5 coxa that is subdorsal and very small (Annexe: Table 2), in contrast to the wide coxae of preceding legs. The P5 coxae are close to each other, but by far to a lesser extent than in the Retroplumidae Gill, 1894 . Only one individual of L. trigeri showed a preserved merus, extremely slender and cylindrical in comparison with the size and shape of the large, compressed meri of preceding legs. The P5 carpus, propodus, and dactylus are not preserved in our material. A natatory condition for P 5 in L. trigeri is doubted, the narrow preserved (although incomplete) merus of L. trigeri not suggesting a paddle-shaped P5. Additionally, the subdorsal location of P5 is not reminiscent of a natatory pereopod.

Araripecarcinus ferreirai Martins-Neto, 1987 View in CoL , from the Brazilian Lower Cretaceous, exhibits, however, an extremely reduced P5 with a cylindrical merus and a wider and flattened article that has been presumed to be natatory (Martins-Neto 1987: 408, figs 1, 2). The small, damaged holotype specimen of 10 mm width does not provide enough information (M. Tavares, pers. comm.). The hypothesis that this crab belongs to the Podotremata Guinot, 1977 (perhaps to the Raninoidea De Haan, 1839) cannot be completely excluded. The Cretaceous Etyidae Guinot & Tavares, 2001 View in CoL , with a wide and areolated carapace, was considered podotreme ( Guinot & Tavares 2001), in contrast to the typically heterotreme Lithophylacidae .

LITHOPHYLACIDAE VS CARCINERETIDAE BEURLEN, 1930 View in CoL

The widespread Cretacous Carcineretidae View in CoL (type genus Carcineretes Withers, 1922 View in CoL ), within the superfamily Portunoidea , is suggested to have become extinct at the end of the Cretaceous (Feldmann et al. 1998; Vega et al. 2001; Schweitzer et al. 2002a; Schweitzer & Feldmann 2005). It includes: Branchiocarcinus Vega, Feldmann & Sour-Tovar, 1995 ,? Cancrixantho Van Straelen, 1934 , Carcineretes View in CoL , Mascaranada Vega & Feldmann, 1991 View in CoL , Ophthalmoplax Rathbun, 1935 , and Woodbinax Stenzel, 1953 ( Van Straelen 1934: 3, pl. 1, fig. 2; Stenzel 1953: 215, figs 6, 8, pl. 59, fig. 11; Beurlen 1958: 1, 6; Bishop 1988: 247; Solé & Vía 1989: 25; Vega et al. 1995: 345; 1997: 619; Fraaye 1996: 271; Feldmann & Villamil 2002: 720; Schweitzer et al. 2002a: 21, 36-40, fig. 29, table 4; 2003a: 44). The Carcineretidae View in CoL was already present in the Cenomanian with Woodbinax Stenzel, 1953 , and in the Turonian with Ophthalmoplax , Cancrixantho , and Carcineretes View in CoL .

Lithophylax was questionably included in the Carcineretidae View in CoL by several authors. In the remarks concerning a new carcineretid, Carcineretes planetarius Vega, Feldmann, Ocampo & Pope, 1997 View in CoL , Lithophylax was considered an “authentic carcineretid crab”, with too poor a preservation to give any paleoecological interpretation of its lifestyle ( Vega et al. 1997: 619). In a new description based on more complete new material of C. planetarius, Vega et al. (2001: 323) View in CoL did not mention Lithophylax among the members of the Carcineretidae View in CoL .

The Carcineretidae View in CoL are characterized by a square or transversally extended carapace (may be urn-shaped, flat to convex longitudinally); dorsal regions well marked by grooves and transverse ridges, posteriormost regions may be ornamented; lateral margins straight, converging posteriorly, more or less diverging anteriorly; outer-orbital spines may be directed antero-laterally; rostrum narrow to broad, straight, grooved or bifid; orbits wide and eyestalks long; suborbital margins long, with fissures and ending in forwarding spines; P5 propodus flattened, paddle-like, and dactylus oval or oblanceolate; P4 may be also flattened; thoracic sternum ovate; male abdomen with 6 free segments ( Withers 1922: 539, pls 16, 17; Glaessner 1969: R514; Schweitzer et al. 2003a: 44, 45).

The Carcineretidae View in CoL is a brachyuran radiation witnessing the rise of a natatory mode of life during the Cretaceous ( Fraaye 1996: 269). However, the flattened P5 propodus and oval dactylus may be also an adaptation for back-burrowing ( Morris 1993). At present taphonomical and morphological data do not indicate that Lithophylax may have been a swimming or a burrowing crab; it was probably a burying crab.

Resemblance of the Carcineretidae View in CoL with the Retroplumidae View in CoL has been evoked, but their relationships are obscure. Ophthalmoplax was considered close to the retroplumid Archaeopus Rathbun, 1908 (see Vía 1980: 5, 11, fig. 2) and a Cretaceous- Retropluma View in CoL ancestor ( Vía 1982: 118, fig. 2) (see also Vega & Feldmann 1992: 145-148, fig. 8). Branchiocarcinus cornatus Vega, Feldmann & Sour-Tovar, 1995 (type species of Branchiocarcinus ) resembles a retroplumid because of the shape and ridges of the carapace; it also has two sharp, anteriorly curved spines (forming the widest portion of the carapace) which probably correspond to the outer-orbital spines of Lithophylax . Archaeopus rathbunae Beurlen, 1965 ( Beurlen 1965: 271, fig. 4), from the Early Cretaceous (Albian) of Brazil, was not considered a retroplumid ( Vía 1980: 54, 64): it has a carapace that resembles that of the Carcineretidae View in CoL (see Vega & Feldmann 1992: 147), and Schweitzer & Feldmann (2001b: 202) suggested that it “may be a carcineretid”.

The Lithophylacidae differs from the Carcineretidae by numerous characters: carapace much wider, in the form of an inverted trapezoid; antero-lateral margins strongly diverging anteriorly; rostrum downturned; fronto-orbital border wide and continuous; supraorbital margin very long, without teeth or spines, only with a small fissure; orbits shape; thoracic sternum wider, with sternites 1-3 broad and flattened; male abdomen relatively wide; P5 reduced, (presumably) not natatory, with subdorsal coxae.

The relatively wide, fully lobulate carapace of Mascaranada (see Vega & Feldmann 1991: fig. 7.1) resembles that of Lithophylax but the antero-lateral margins are converging anteriorly and the P5 is paddle-like, not reduced. The narrow, not sulcate rostrum of Lithophylax fits the description of some Carcineretidae , such as the poorly known Cancrixantho Van Straelen, 1934 , from the Campanian, which was placed in the Carcineretidae ( Glaessner 1969: R514, fig. 325; Vega et al. 1997: 619; Schweitzer et al. 2002a: 21; 2003a: 44 with reservation). Cancrixantho pyrenaicus Van Straelen, 1934 (see Vía 1988: 351, fig. 339L; Solé & Vía 1989: 25) shows a subrectangular carapace that is deeply grooved and marked by transverse ridges, a narrow and spiniform rostrum, long orbits and eyestalks, characters which resemble those of L. trigeri . However, the carapace shape, the trilobate supra-orbital margin and the toothed postero-lateral margin of C. pyrenaicus distinguish it from L. trigeri .

In Lithophylax the developed outer-orbital teeth and the distance between these teeth (coinciding with maximum width) correspond to the condition found in some Carcineretes , Ophthalmoplax and Mascaranada , but the rest of the lateral margin is unarmed (only a small angle at the level of hepatic lobe) in Lithophylax instead of at least one more tooth or spine in the other genera. In the Carcineretidae (as Carcineretes , Ophthalmoplax ) the rostrum is broader than in Lithophylax trigeri ( Figs 5 View FIG ; 7 View FIG ; 10A View FIG ; 14 View FIG ) and more or less grooved or bifid.

Some of the genera previously included in the Carcineretidae have been removed from the family. Withersella Wright & Collins, 1972 ( Wright & Collins, 1972: 91, fig. 13, pl. 19, figs 4, 5; see also Fraaye 1996: 270, fig. 1.8) (type species: W. crepitans Wright & Collins, 1972 from the early Aptian), first regarded as the earliest known carcineretid, shows a rather concave frontal margin bounded by large outer frontal spines, supraorbital fissures, and a spine on each side of the bifid rostrum. Collins et al. (1995: 200) suggested close relationships to the Maastrichtian Binkhorstia Noetling, 1881 , and placed these two genera in the Carcineretidae . Van Bakel et al. (2003: 85-87, fig. 1) recently re-assigned Binkhorstia to the Torynommatidae Glaessner, 1980 ( Glaessner 1980: 180) , based on new material of the type species, B. ubaghsi (van Binkhorst, 1857) . Binkhorstia ubaghsi has a spatulate rostrum, all abdominal segments residing in a true sterno-abdominal cavity, broad thoracic sternum, flattened P2-P4, and a P5 coxa which is reduced and situated laterally to the first abdominal segment. The difficulty of the systematic placement of early crabs is well shown by the case of Binkhorstia , which was considered either a podotreme or an heterotreme, being successively assigned to the Dorippidae MacLeay, 1838 ( Quayle & Collins 1981: 738), the Cyclodorippidae Ortmann, 1892 ( Glaessner 1969: R492; Feldmann & Villamil 2002: 721), and to the Carcineretidae Beurlen, 1930 ( Wright & Collins 1972; Collins et al. 1995; Fraaye 1996: 272, figs 1.9, 1.10, 2; Wright 1997: 138, figs 12, 16; Jagt et al. 2000: 40, fig. 2). It is presently impossible to decide the status of all the Torynommatidae without verification of their sexual gonopores. The information at our disposal does not permit recognizing affinities of the Torynommatidae with the Lithophylacidae .

The Late Cretaceous genus Icriocarcinus Bishop, 1988 , initially assigned to the Carcineretidae with respect to its long eyestalks, transverse ridges on the dorsal carapace and heterochelous chelipeds, was included in the Goneplacidae ( Schweitzer et al. 2002a: 21, 28, 40). The P3-P5 meri are flattened, P5 are the smallest (but not markedly reduced and end with straight dactylus), and these characters (among others) are not carcineretid-like (see below).

Longusorbis Richards, 1975 View in CoL ( Richards 1975: 1850, figs 1-11), first included within the Carcineretidae View in CoL (see Bishop 1988: 251), then placed in the Xanthidae View in CoL (see Vega et al. 1997: 619; Schweitzer et al. 2002a: 21), was recently reassigned to the Carcineretidae View in CoL ( Schweitzer et al. 2003a: 44, fig. 15). The excellently preserved L. cuniculosus Richards, 1975 View in CoL , found buried and fossilized in its burrows (see Paleoecology and ethology), shows an almost complete eyestalk, as long as its supra-orbital margin, P2-P5 slightly compressed, “modified for crawling” ( Richards 1975: 1862), and P5 shorter than precedings, with paddle-like propodus and oblanceolate dactylus ( Schweitzer et al. 2003a: 45, fig. 15.1), that is a P5 very different from the reduced and subdorsal P5 of Lithophylax trigeri ( Figs 12B View FIG ; 14 View FIG ). See also Conclusion.

LITHOPHYLACIDAE VS NECROCARCINIDAE FÖRSTER, 1968 AND View in CoL ORITHOPSIDAE SCHWEITZER, FELDMANN, FAM, HESSIN, HETRICK, NYBORG & ROSS, 2003 View in CoL

The exclusively fossil family Necrocarcinidae Förster, 1968 View in CoL , with the first records occuring during the Early Cretaceous, was first considered a member of the Calappoidea De Haan, 1833 ( Förster 1968; Wright & Collins 1972; Schweitzer & Feldmann 2000a; Fraaije 2002), then attributed to the Dorippoidea MacLeay, 1838 ( Schweitzer et al. 2003a: 31, 32). According to Jagt et al. (2000: 40) some supposed necrocarcinids might represent parthenopids rather than calappids. The Necrocarcinidae View in CoL was assigned with reservation to the Podotremata ( Collins & Williams 2004: 34).

Six genera have been included in the Necrocarcinidae : Campylostoma Bell, 1858 , Cenomanocarcinus Van Straelen, 1936 ,? Corazzatocarcinus Larghi, 2004 (see below), Hasaracancer Jux, 1971 , Necrocarcinus Bell, 1863 , Paranecrocarcinus Van Straelen, 1936 , Pseudonecrocarcinus Förster, 1968 (for the family diagnosis see Schweitzer et al. 2003a: 32), and? Shazella Collins & Williams, 2004 ( Collins & Williams 2004). The earliest known occurrences include the Hauterivian Paranecrocarcinus hexagonalis Van Straelen, 1936 ; the Barremian P. kennedyi Wright, 1997 and Necrocarcinus ? olssoni (Rathbun, 1937); the upper Aptian Necrocarcinus undecimtuberculatus Takeda & Fujiyama, 1983 ; the Aptian- Cenomanian N. labeschii (Eudes-Deslongchamps, 1835, as Orithyia labeschii ); the Albian Paranecrocarcinus graysonensis ( Rathbun, 1935) , P. moseleyi ( Stenzel, 1945) , Necrocarcinus texensis Rathbun, 1935 , Cenomanocarcinus renfroae ( Stenzel, 1945) , C. oklahomensis ( Rathbun, 1935) , and Pseudonecrocarcinus stenzeli Bishop, 1983 ; the Albian-lower Cenomanian N. woodwardi Bell, 1863 ; and the Cenomanian P.libanoticus Förster, 1968 , P. digitatus Wright & Collins, 1972 , P. mozambiquensis Förster, 1970 , Paranecrocarcinus biscissus Wright & Collins, 1972 , and an indeterminate genus and species of Egypt ( Schweitzer et al. 2003b).

Necrocarcinus labeschii View in CoL , type species of the genus Necrocarcinus View in CoL ( Glaessner 1929: 282; 1969: R495, fig. 306.3; Wright & Collins 1972: 63, pl. 11, pl. 22, fig. 8a-c), is known by its male abdomen: six segments and telson in both sexes, a sharp dorsal rib on the first five segments, and segment 6 twice as long as segment 5 ( Wright & Collins 1972: 64).

As in many fossils, necrocarcinids have a long history of taxonomic problems and transferrals. The case of Necrocarcinus siouxensis Feldmann, Awuota & Welshenbaugh, 1976 ( Feldmann et al. 1976: pl. 1, fig. 5), from the Maastrichtian of North Dakota, is representative of the difficulty of the placement of crabs supposed to be necrocarcinids. Not referable to Necrocarcinus Bell, 1863 View in CoL ( Bishop & Williams 1991), of uncertain position ( Fraaye 1994: 264) or tentatively assigned to Cenomanocarcinus View in CoL ( Schweitzer et al. 2003a: 36-39, table 1), C. siouxensis was assigned to the Podotremata by Guinot & Quenette (2005: 329). The Cretaceous necrocarcinid Cenomanocarcinus vanstraeleni Stenzel, 1945 View in CoL ( Stenzel 1945: 447, fig. 15, pl. 44) also seems to be typically podotreme. The two species C. siouxensis and C. vanstraeleni View in CoL , as well as other necrocarcinids, might be referred to a new genus to be included in the Podotremata.

The hypothesis that certain members of the Necrocarcinidae actually prove to be non-eubrachyurans was already suggested by Larghi (2004: 529, 530), who questionably placed some of them within the Podotremata. When ventral parts exist in fossil records and have been cleared, the dimensonial relations between the male abdomen and the thoracic sternum are fundamental. A (basal) podotreme crab is characterized, besides its coxal female gonopores and the presence of spermathecae, by having most of the sternum laterally covered by male abdomen and in contact with coxae of pereopods ( Guinot & Tavares 2001).

In the supposed necrocarcinid Corazzatocarcinus Larghi, 2004 ( Larghi 2004: 530, figs 2-4; see also Larghi & Garassino 2000: 53, fig. 1), belonging to an “uncertain superfamily”, the sternum is missing, and the abdomen is probably that of a female; the dorsal and reduced P4 and P5 might support an assignation to the basal Podotremata (Dromiacea De Haan, 1833), otherwise to the Dorippoidea.

The Campanian Hasaracancer cristatus Jux, 1971 ( Jux 1971: figs 2A, C, pl. 17, figs 1, 2), with an incompletely folded abdomen and abdominal pleurae, was first referred to the Raninidae , thus supposed to be a podotreme crab, and then included in the Necrocarcinidae within the Heterotremata ( Schweitzer et al. 2003a: 32).

Some of the species first included in the Necrocarcinidae ( Schweitzer & Feldmann 2000a: 232, 246, fig. 1) were referred to the Orithopsidae Schweitzer, Feldmann, Fam, Hessin, Hetricks, Nyborg & Ross, 2003 ( Schweitzer et al. 2003a: 33). Both Necrocarcinidae and Orithopsidae have been associated within the Dorippoidea ( Schweitzer et al. 2003a: 39). The family diagnosis of the Orithopsidae must be completed in including the following characters: 1) thoracic sternum, known in Goniochele angulata Bell, 1858 and G. madseni Collins & Jakobsen, 2003 : sternum of G. angulata broadly ovate, two-fifths the width of the carapace, with “Episternum [corresponding to the anterior sternites] considerably longer than wide, much produced, pointed” ( Carter 1898: 23); female sternum of G. angulata and G. madseni with nodes: sternites 1-2 fused and forming a triangle between the mxp3; sternite 3 subcrescentic; sternite 4 subtrapezoidal; sternites 6-7 triangular ( Collins & Jakobsen 2003: pl. 3, figs 2a, 4a); 2) male abdomen of G. angulata very long, linear and narrow, with all free segments (see Bell 1858: 27, pl. 4, fig. 8); 3) female abdomen of G. angulata with segment 6 twice as long as any other of the anterior segments according to Carter (1898: 23, pl. 1, fig. 6), a feature not apparent in the figure of Bell (1858: fig. 9); and 4) subdorsal position of P4 and P5, coxae preserved in G. angulata (see Bell 1858: 26, 27, pl. 4, figs 4, 5). This last feature generally implies reduced P4 and P5, special arrangement (mobility), and carriage over the carapace. These pereopods have so far not been preserved in other necrocarcinid genera.

The familial status of some genera with apparent similarities to the Necrocarcinidae remains questionable. For example the Paleocene Camarocarcinus Holland & Cvancara, 1958, placed within the Necrocarcininae or Necrocarcinidae (see Jakobsen & Collins 1979: 63; Fraaye 1994: 264; Fraaije 2002: 914) or within the Calappidae (see Collins & Rasmussen 1992: 33, fig. 19; Schweitzer & Feldmann 2000a: 234, 246, fig. 3), was finally considered most closely related to the Leucosiidae Samouelle, 1819 ( Schweitzer et al. 2003a: 34). We have examined a cast of C. quinquetuberculatus Collins & Rasmussen, 1992 , with well preserved ventral parts. The mxp3 are elongate and of oxystomian type, the arthrodial cavities of the chelipeds are located close together and overhang the sterno-abdominal cavity, the thoracic sternum is narrow, with a flat, undivided bottom forming “a median furrow” ( Collins & Rasmussen 1992: 36) and lateral flanges oriented almost vertically. Anyway the combination of ventral characters (neither calappid nor leucosiid) does not allow a confident attribution, only suggests inclusion of Camarocarcinus in an extinct family.

The status of Shazella Collins & Williams, 2004 , from the upper Turonian, known by carapace only, remains uncertain.

LITHOPHYLACIDAE VS HEXAPODIDAE

MIERS, 1886

The P5, in the course of our first investigations on Lithophylax , were not visible in any of the specimens that were examined, and, finally, the discovery of a P5 subdorsal coxa prolonging into a narrow merus was an important discovery. The absence of discernible P 5 in fossil records makes the assignment to family and generic levels difficult, which may result in inaccurate placements. The absence of P5 may be due either to an evident absence as it is traditionally stated for the Hexapodidae (loss of all P5 articles except for the concealed, vestigial coxa in males, see Guinot 2006; Guinot, Tavares & Castro unpublished), or to a complete loss during fossilization. The fact that in many fossil crabs the legs do not readily fossilize, especially in the case of reduced, thin and subdorsal last pair, thus not preservable (see Glaessner & Rao 1960; Schweitzer & Feldmann 2001a; Larghi 2004), creates a problem for their identification. A reduced and subdorsal P5 which has been lost during fossilization risks being misinterpreted as missing. The only way of determining the presence of a P5 consists in carefully examining the P5 coxa (if preserved), the size and position of its arthrodial cavity, and the space that remains for this cavity at the posterior corners of the carapace. Reduced P5 coxae are often subdorsal (sometimes also P4); additionally, they have often a reentrant location on the carapace (also for the P4, but to a lesser extent), which corresponds to the progressive dorsal position of P4 and P5 related to the posterior thoracic curvature. For instance, in the case of the fossil Retroplumidae , preserved P5 coxae are much closer to each other than preceding ones so that they leave a very short distance between them. It is possible that some fossil crabs supposed to be Hexapodidae actually belong to the Retroplumidae . The characters of the thoracic sternum, when present, nevertheless, permit an identification since this region is quite different in these two families.

Lithophylax does not belong to the Hexapodidae View in CoL since cleaning from the matrix resulted in discovery of a P5, with a subdorsal, reduced coxa and a (presumably) narrow merus bearing a row of granules. Other features of Lithophylax are not at all those of an hexapodid, in particular the thoracic sternum. However, one characteristic shared by the Hexapodidae View in CoL (a part of them only) and the Lithophylacidae is the presence of a stridulatory apparatus. Extant Hexapodidae View in CoL ( Hexaplax Doflein, 1904 View in CoL , Hexapus De Haan, 1835 View in CoL , Stevea Manning & Holthuis, 1981 View in CoL , and Paeduma Rathbun, 1898 View in CoL ) as well as fossil (the Eocene Hexapus pinfoldi Collins & Morris, 1978 View in CoL and Stevea cesarii Beschin, Busulini, De Angeli & Tessier, 1994 View in CoL ; the upper Eocene Goniocypoda edwardsi Woodward, 1867 View in CoL and probably G. quaylei Crane, 1981 View in CoL ) show stridulating striae, arranged in two different patterns, either a suborbital row or a pterygostomial area (Guinot-Dumortier & Dumortier 1960; Manning & Holthuis 1981; Crane 1981: 6, 7, fig. 8D; Manning 1982; Glaessner & Secretan 1987: 8, pl. 1, fig. 5b, 6; Beschin et al. 1994: 194; Guinot 2006). However, the stridulating mechanism is different in the two families. In Lithophylax trigeri ( Figs 5B View FIG ; 13 View FIG ; 14 View FIG ), most of the inner part of the P1 merus bears a wide transverse area with 29 strong striae, and it is likely that the plectrum ( Fig. 8A, C View FIG ) consists of at least eight suborbital tubercles (versus striae on the ventral surface of the cephalothorax rubbed by striae on inner surface of P1 dactylus in the Hexapodidae View in CoL ).

The main differences between the Lithophylacidae and the Hexapodidae View in CoL are the presence of P5 with a subdorsal coxa and narrow merus in the Lithophylacidae ( Figs 11B View FIG ; 12B View FIG , below; 14) (partial loss of P 5 in males except for the vestigial coxa in male hexapodids, absent in female hexapodids; see Guinot 2006; Guinot, Tavares & Castro unpublished); thoracic sternum broadly triangular; sternites 4-8 unequal, sternite 7 being smaller, and sternite 8 subdorsal and smaller, covered by abdomen in the Lithophylacidae (very wide sternum, sternites 5-7 subrectangular, similarly developed and high, sternite 8 subdorsal, reduced, and generally partly exposed in the Hexapodidae View in CoL , see Glaessner & Secretan 1987; Beschin et al. 1994; Guinot 2006); sternal sutures 5/6 and 6/7 slightly oblique, presumably complete in the Lithophylacidae (sutures 4/5 to 6/7 nearly parallel, equidistant and interrupted in the Hexapodidae View in CoL ); male abdomen triangular in the Lithophylacidae (relatively narrow and with straight margins in Hexapodidae View in CoL ). A character shared by the Lithophylacidae and the Hexapodidae View in CoL is the mxp3 condition: they are strongly divergent anteriorly, leaving the mandibles exposed.

Palaeopinnixa perornata Collins & Morris, 1976 View in CoL , known from Miocene and Eocene (see Feldmann & Schweitzer 2004: 19, fig. 3B-E), has inflated swellings on the carapace which contrast with the smooth carapace of most fossil and extant hexapodids (the Recent Latohexapus Huang, Hsueh & Ng, 2002 View in CoL excepted) and shows a long male abdomen which is reminiscent of that of Paeduma cylindraceum (Bell, 1859) View in CoL and Stevea williamsi (Glassell, 1938) View in CoL (see Guinot 2006).

LITHOPHYLACIDAE VS RETROPLUMIDAE View in CoL

GILL, 1894

The relationships of the family Lithophylacidae with the Retroplumidae merit discussion. First considered “aberrant” Catometopes among the Goneplacidae ( Alcock & Anderson 1894: 180 under Archaeoplax Alcock & Anderson, 1894 ), then true Catometopes but of “an archaic type ” ( Alcock 1899: 78, 79 as Ptenoplacidae ), near Hexapus ( Alcock 1900: 285, 455 as Ptenoplacidae ), the Retroplumidae was left, more or less explicitely, in the Catometopes by many authors ( Tesch 1918: 29-33; Rathbun 1918: 15; Sakai 1976: 322, 592). The family was not included in the Thoracotremata by Guinot (1977), as stated by Schweitzer & Feldmann (2001b: 201). At present, the Retroplumidae appears as an heterotreme family ( Schram 1986; Guinot 1978; Saint Laurent 1989; Guinot & Bouchard 1998; McLay 2006), and is placed in proximity to the Dorippidae or Palicidae ( Balss 1957: 1633, 1662; Guinot 1978: 214, 249, 251, 284; Guinot & Quenette 2005: 334).

Retroplumid crabs are known from many fossils, and paleontologists are faced with a rich and diverse fauna, known only from the Late Cretaceous (the Early Cretaceous Archaeopus rathbunae was suspected to be a carcineretid by Schweitzer & Feldmann 2001b; see above). The family was supposed to have originated during the Cretaceous in the Americas ( Schweitzer & Feldmann 2001b: 202). The Retroplumidae was placed close to the Ocypodidae or within the Ocypodoidea ( Beurlen 1930: 350-352; Glaessner 1969: R531; Collins & Morris 1975: 823; Vía 1957: 553; 1969: 322- 328; 1980: 4, fig. 1, table 1; 1982: 115, figs 1, 2; Bishop 1983: 427), or considered a thoracotreme family ( Vega & Feldmann 1992: 139; Collins et al. 1993: 304; 1994: 29; Feldmann & Martins- Neto 1995: 610; Vega et al. 1995: 347; Feldmann et al. 1995: 16; 1997: 126). Beschin et al. (1996), Martin & Davis (2001), Collins et al. (2003), and Schweitzer et al. (2003a) followed Saint Laurent (1989) in considering a separate superfamily, Retroplumoidea. According to Saint Laurent (1989), Archaeopus Rathbun, 1908 ( Rathbun 1908: 346, pl. 47, figs 4-7, pls 48, 49, figs 2-4), considered one of the oldest known retroplumid fossil genus (or closely allied to) ( Beurlen 1930; Vía 1969; Collins & Morris 1975; Vega & Feldmann 1992; Schweitzer et al. 2003a) and supposed to have survived the end of the K/T event, must be removed from the Retroplumidae . According to McLay (2006) many fossil retroplumids (all Archaeopus species and several representatives of the genus Costacopluma Collins & Morris, 1975 ) are very likely members of the Palicidae (see below).

The differences between the Lithophylacidae and the Retroplumidae are too numerous to be enumerated. Nevertheless, the two families share some features: rostrum developed, projected (similarly downturned in Archaeopus rostratus Schweitzer, Feldmann, Fam, Hessin, Hetricks, Nyborg & Ross, 2003 ), thoracic sternum broad (sutures 4/5-7/8, however, are all interrupted in the retroplumids), mxp3 obviously diverging, sternite 8 reduced, and P5 dorsal, reduced, in the Lithophylacidae not as reentrant, however, as in the Retroplumidae .

LITHOPHYLACIDAE VS PALICIDAE BOUVIER, 1898 The View in CoL only obvious resemblances between the Lithophylacidae and the Palicidae Bouvier, 1898 View in CoL (see Castro 2000 for the date of publication) concern the P5, which are very reduced, nearly filiform, and dorsal in the Palicinae, instead to be represented only by a small, subdorsal coxa and a (presumed) narrrow merus in L. trigeri ( Figs 11A View FIG ; 12B View FIG ; Annexe: Table 2). In the Palicidae View in CoL the thoracic somite 8 is strongly modified, so that sternite 8 is reduced (although nearly aligned with preceding sternites) and also pleurite 8 is modified (Guinot, Tavares & Castro unpublished). In the Lithophylacidae the sternite 8 is subdorsal, reduced, covered by the abdomen in both sexes, and not visible dorsally.

The Palicidae View in CoL was regarded as a catometope or thoracotreme family ( Faxon 1895: 38; Alcock 1900: 290, 450; Borradaile 1907: 482; Rathbun 1918: 15, 182; Bouvier 1940: 303; Monod 1956: 387; Balss 1957: 1633, 1661; Serène 1965: 25; 1968: 97; Sakai 1976: 592; Manning & Holthuis 1981: 191; Schram 1986: 308), and was even included in the Ocypodoidea Rafinesque, 1815 ( Martin & Davis 2001: 75). Actually the Palicidae View in CoL shows a typical coxo-sternal condition of the male gonopores ( Guinot 1979: fig. 30G), thus is heterotreme. A relationship with the Dorippoidea was suggested ( Cano 1891; Bouvier 1897a, b; 1898; A. Milne-Edwards & Bouvier 1900, 1902; Gurney 1942; Bourdillon-Casanova 1960; Guinot 1978; Guinot & Quenette 2005). Affinities between the Palicidae View in CoL and Carcineretes woolacotti Withers, 1922 View in CoL (see Vega et al. 2001; Donovan et al. 2003), from the Maastrichtian, were evoked by Withers (1922: 541), despite the conflicting indications of their catometope/cyclometope conditions and the different shape of their P5.

The Palicidae Bouvier, 1898 View in CoL shares also with the Retroplumidae View in CoL a strong modification of somite 8 and dorsal location of P5. Fossil species of Archaeopus Rathbun, 1908 and Retrocypoda Vía, 1957 View in CoL , presently assigned to the Retroplumidae View in CoL , have been considered possible members of the Palicidae View in CoL ( Glaessner 1969: R531; McLay 2006). The carapace of the Eocene (lower Lutetian) palicine Spinipalicus italicus Beschin & De Angeli, 2003 View in CoL ( Beschin & De Angeli 2003: 7-12, figs 2-4) shows some similarities with the Cretaceous Archaeopus antennatus Rathbun, 1908 ( Rathbun 1908: 346, pl. 47, figs 4-7, pl. 48, pl. 49, figs 2-4), as stated by Beschin & De Angeli (2003: 12).

Several fossil Palicidae View in CoL are known: Eopalicus Beschin, Busulini, De Angeli & Tessier, 1996 View in CoL , from Eocene and Oligocene; Miocene species were attributed to the extant genus Palicus Philippi, 1838 View in CoL by Van Straelen (1938) and Müller (1984); see Beschin & De Angeli (2003). We agree with the hypothesis that ornamentation of transverse ridges in Eopalicus squamosus Beschin, Busulini, De Angeli & Tessier, 1996 View in CoL ( Beschin et al. 1996: pl. 1), with a typically palicid carapace-shape, is reminiscent of the “terraces” found in raninids and indicates a similar adaptation for burying. The condition of Spinipalicus italicus View in CoL , with a broad carapace and inflated, tuberculated dorsal regions, resembles some ancient retroplumids. The thoracic sternum, abdomen and (fragile) P5 are unfortunately unknown in fossil palicids. Possible relationship between the three eubrachyuran families Retroplumidae View in CoL , Palicidae View in CoL and Hexapodidae View in CoL , briefly evoked by Guinot & Quenette (2005: 334), needs further discussion (Guinot, Tavares & Castro unpublished).

LITHOPHYLACIDAE VS GONEPLACIDAE

MACLEAY, 1838

The family Goneplacidae is presently included within the Xanthoidea, by the neontologists ( Martin & Davis 2001) as well as paleontologists (Schweitzer 2000; Schweitzer et al. 2002a; Karasawa & Kato 2003). The definition of the Goneplacidae , as recently given by Schweitzer (2000: 717) and Karasawa & Kato (2003: 137), shows the difficulty (also encountered by neontologists) to reconcile all the characters attributed to the members of this heterogeneous family: carapace angular, rectangular, trapezoidal or rounded; dorsal regions of carapace relatively flattened; areolation and ornamentation either indistinct, or weakly distinct or developed; orbits wide and elongated; eyestalks short or elongated, cornea sometimes inflated; supra-orbital margin sometimes without fissures; supra-orbital angle sometimes present; outer-orbital spines attenuated or directed laterally; antero-lateral margins usually toothed; thoracic sternum broad; male abdomen with all segments free but segments may be fused in some; P2-P5 long, dactylus not modified, with or without corneous tip; G2 long or short.

The Late Cretaceous Icriocarcinus Bishop, 1988 , differs from other goneplacid members in possessing well developed carapace regions. The thin exoskeleton is consistent with a burrowing mode of life as evidenced by a specimen of I. xestos found preserved in a simple, oblique burrow ( Bishop 1988: 246, fig. 3D). Icriocarcinus xestos Bishop, 1988 , species of Ommatocarcinus (as O. corioensis (Cresswell, 1886)) , and of Orbitoplax Tucker & Feldmann, 1990 (as O. weaveri (Rathbun, 1926)) have deep grooves and inflations on the carapace, sometimes transverse ridges (Schweitzer 2000: 724). Fossil goneplacids as Branchioplax Rathbun, 1916 and Orbitoplax Tucker & Feldmann, 1990 exhibit a xanthid-like carapace (Schweitzer 2000: 718).

With respect to the recent revison of the Goneplacidae as divided into six subfamilies ( Karasawa & Kato 2003), Lithophylax shares with the Goneplacinae sensu stricto (type genus: Goneplax Leach, 1814 ) the following characters: front straight and prolonging without notch to supra-orbital margin, orbital borders without fissures (only a tiny supra-orbital fissure in Lithophylax ) nor angles, antero-lateral borders poorly toothed, thoracic sternum wide, sternal suture 7/8 interrupted, male abdomen wide, with all somites free, and long G2.

Lithophylax resembles Goneplax View in CoL sensu stricto (type species: G. rhomboides (Linnaeus, 1758)) View in CoL (see Guinot 1969b: 521, figs 64, 71, 72) and its allied genera (such as Ommatocarcinus View in CoL ) by having rectangular carapace that is much widened anteriorly; lateral margins strongly diverging anteriorly, with only one laterally directed tooth at the outer-orbital angle ( Fig. 5A View FIG ); usually developed orbits and eyes; wide thoracic sternum. The slanted and hollowed lateral margins of Lithophylax resemble those of the extant Ommatocarcinus macgillivrayi White, 1851 View in CoL (type species of Ommatocarcinus View in CoL ) and fossil members. The photographs and reconstruction of the dorsal surface carapace, mxp3 and abdomen of the well preserved O. corioensis View in CoL by Jenkins (1975: fig. 1) allow a clear comparison with L. trigeri .

However, Lithophylax markedly differs from the Goneplacidae View in CoL by several characters: rostrum extremely narrow and strongly downturned (wider and weakly inclined in Goneplax View in CoL ); front straight and continuous on the same line with orbital margin (sinuous in Goneplax View in CoL ); lobulation of dorsal surface of carapace marked (absent in Goneplax View in CoL ); G1 long, narrow and slightly tapering (rather stout and ending with a lobe in Goneplax View in CoL ); sternite 8 reduced, subdorsal and not exposed (more developed, not subdorsal and partly exposed in Goneplax View in CoL ); chelipeds short and stout (narrow and elongated in Goneplax View in CoL ); P2-P4 thick and massive (slender in Goneplax View in CoL ); P5 subdorsal and markedly reduced (“normal” position and size in Goneplax View in CoL ).

Goneplacidae View in CoL sensu lato seems, however, the closest family to the Lithophylacidae . Both families share: sternal thoracic sutures 4/5 and 5/6 that are interrupted and with depressions; suture 6/7 complete or weakly interrupted, with its extremities ending together so that it is (presumably) complete in Lithophylax , interrupted medially in Goneplax View in CoL ; suture 7/8 either complete (for instance in the Mathildellinae Karasawa & Kato, 2003) or interrupted (for instance in the Goneplacinae View in CoL ) (for sketches of the “goneplacid” thoracic sternum see Guinot 1969a-c; Guinot & Richer de Forges 1981). An important feature shared by the Goneplacinae View in CoL and the Lithophylacidae is the interrupted suture 7/8. The complete suture 5/6 of the Lithophylacidae indicates a less derived condition than that of the Goneplacidae View in CoL . It should be noted that in the Mathildellinae the sternite 8 is (nearly) completely covered by the abdomen, as in the Lithophylacidae , a character that (with the complete suture 7/8) corroborates the basal place of the Mathildellinae ( Karasawa & Kato 2003: 138). There is no close relationships, however, between the Lithophylacidae and the Mathildellinae, in which the sternite 8 is not subdorsal and the P5 are not reduced.

Lithophylax shows, like in the Recent goneplacid Bathyplax View in CoL , a stridulating mechanism which roughly involves the same parts, i.e. the P1 merus and the ventral surface of carapace. Bathyplax View in CoL seems to be the only crab to possess a specialized and wide striated area on P1 merus similar to that of Lithophylax , with striae that are similarly oriented ( Fig. 13 View FIG ). The subhepatic granules, however, are numerous and spaced in Bathyplax View in CoL , instead of being aligned and few (only 8) in L. trigeri ( Figs 8A, C View FIG ; 13C View FIG ). See Remarks on the stridulating apparatus.

Icriocarcinus xestos Bishop, 1988 ( Bishop 1988: 250, figs 2, 3A-D, table 1) exhibits some characters similar to those of L. trigeri : carapace shape, dorsal surface with plateau-like areolations, slanted lateral margins, strong outer-orbital spine, narrow and downturned rostrum, long and flattened P2- P4, broad sternum with barely visible somite 8, all abdominal segments free. According to Bishop (1988) in Icriocarcinus xestos the broadly rounded antero-lateral margins bear “three spines, one at the outer angle of orbit?”. Another hypothesis may be that the supra-orbital margin itself is spined, the external spine being actually at the outer-orbital angle, and the antero-lateral margin is slanted and unarmed, as in some goneplacids and in Lithophylax . L. trigeri shows a stout, thick basophthalmite, and the podophthalmite (absent) is presumably long and club-shaped, with an inflated cornea. Icriocarcinus xestos is decribed with an “eyestalk long and slim, apparently folding back into long orbit extending beneath front from rostrum to first anterolateral spine” ( Bishop 1988: 251, figs 2A, 3A), that might suggest a condition similar to that of Lithophylax . Icriocarcinus is perhaps the closest crab to Lithophylax by the preceding set of characters, but the two genera differ, however, by several features. Icriocarcinus xestos is characterized by: P5 being only smaller than preceding pereopods as usual, flattened and shaped similarly to preceding legs, with the merus as long as the broad carpus and propodus, with a straight dactylus (versus P5 subdorsal and much reduced in L. trigeri ); chelipeds elongate and heterochelous (versus massive and only slightly heterochelous in L. trigeri ).

Longusorbis Richards, 1975 View in CoL , considered close to Icriocarcinus (see Bishop 1988: 251) and with a similar burrowing mode of life ( Richards 1975: fig. 6), was transferred into the Xanthidae View in CoL ( Vega et al. 1997: 619; Schweitzer et al. 2002a: 21), finally into the Carcineretidae ( Schweitzer et al. 2003a: 44) View in CoL , based upon a position justified by the paddlelike P 5 in Longusorbis cuniculosus View in CoL .

Jenkins (1975: fig.7) suggested that Ommatocarcinus corioensis (Cresswell, 1886) View in CoL , a true goneplacid, was found in association with fossil burrow complexes which resembles the dwelling burrows of the extant Goneplax rhomboides (Pennant, 1777) View in CoL (see Rice & Chapman 1971: figs 5-7). In Ommatocarcinus View in CoL and Goneplax View in CoL , as in Icriocarcinus and Longusorbis View in CoL , the lateral extension and flattening of the body, the elongated pereopods, the developed orbits in which eyestalks may be folded are adaptative characters for burrowing. Conversely, Lithophylax , with its subdorsal and reduced P5, is not a burrower.

Placement of Icriocarcinus within the Goneplacidae has extended the geologic range of the Goneplacidae into the Late Cretaceous ( Schweitzer et al. 2002a: 21, 28, 40). Icriocarcinus does not belong to the Lithophylacidae .

The components of the exclusively fossil subfamily recently erected within the Goneplacidae , the Icriocarcininae Števčić, 2005, were unfortunately not quoted by the author (only mention of the type genus, Icriocarcinus ), in a similar way that was used for all the new taxa established by this author ( Števčić 2005: 69).