Revision of Mesozoic decapod crustaceans from Madagascar

ABSTRACT Decapod crustaceans from the Mesozoic of Madagascar have been studied by a number of authors during the last century. One of the largest sets of specimens was collected by the French General Maurice Collignon between 1930 and 1950; this was subsequently studied by Secrétan (1964) who described numerous species of macrurans, brachyurans, and axiideans. These crustaceans originate from the Upper Jurassic and the Upper Cretaceous of the Mahajanga Basin (NW Madagascar) and, especially, from the Morondava Basin (central-SW Madagascar). The purpose of the present study is twofold: to furnish an update of the stratigraphy and geology of the studied areas and, above all, to revise the Secrétan's species, employing current systematic nomenclature, supplying for each detailed geographic and stratigraphic data since such were either cursory or incomplete in previous papers. The present revision considers 13 species recorded by Secrétan as valid, namely: Enoploclytia collignoni Secrétan, 1964, Eryma granuliferum Secrétan, 1964, Pustulina spinulata (Secrétan, 1964) (all Erymidae Van Straelen, 1925); Hoploparia collignoni (Van Straelen, 1949), H. pusilla Secrétan, 1964 (both Nephropidae Dana, 1852 sensu Tshudy & Babcock 1997); Ctenocheles madagascariensis Secrétan, 1964 (Ctenochelidae Manning & Felder, 1991); Schlueteria menabensis Secrétan, 1964 (Axiidae Huxley, 1879); Linuparus bererensis Secrétan, 1964 (Palinuridae Latreille, 1802); Dromiopsis pulchella Secrétan, 1964 (Dynomenidae Ortmann, 1892); Caloxanthus simplex (Secrétan, 1964) (Etyidae Guinot & Tavares, 2001); “Xanthosia” robertsi Secrétan, 1982 (nomen novum pro Xanthosia elegans Secrétan, 1964, non Roberts, 1962), Titanocarcinus mamillatus Secrétan, 1964 (both Xanthoidea incertae sedis); and Secretanella arcuata (Secrétan, 1964) (indeterminate family). The three species described by Secrétan (1964) as belonging to the raninid genus Notopocorystes McCoy, 1849, N. bituberculatus, N. australis and N. denisae will be revised in a forthcoming paper. Our comparative studies have also revealed a number of synonymous taxa, as follows: Eryma granuliferum Secrétan, 1964 (junior synonym: E. madagascariensis Secrétan, 1964); Linuparus bererensis Secrétan, 1964 (junior synonym: L. bererensis multispinosus Secrétan, 1964); Hoploparia collignoni (Van Straelen, 1949) (junior synonym: H. intermedia Secrétan, 1964 and H. sculpta Secrétan, 1964); Schlueteria menabensis Secrétan, 1964 (junior synonym: S. tuberculosa Secrétan, 1964). Lastly, we have doubts about the systematic validity of Enoploclytia armata Secrétan, 1964, Eryma australe (Secrétan, 1964), and Coleia incerta Secrétan, 1964, because of the absence of main diagnostic features, useful for their systematic ascription.


GeoGraphic distribution and GeoloGical settinG
Three principal sedimentary basins have been recognized in Madagascar: the Diégo Basin (in the north-northeast of the island), the Mahajanga Basin (along the northwestern coast), and the largest Morondava Basin, which extends from the central to the southwestern coast (Besairie 1972) (Fig. 1). All specimens studied for the present paper originate from the Mahajanga and Morondava basins.

Mahajanga Basin
This basin extends along the northwestern part of the island covering a crescent-shaped area around the town of Mahajanga and delimited by the Ampasindava peninsula in the north and the Cape Saint-André Anticline in the south (Fig. 2). Developed in the basin are sediments of Late Palaeozoic (Sakamena Formation) to Holocene age. We have divided the Mahajanga Basin into two geographic portions to simplify location of the fossiliferous levels, as follows: 1) the northern limb of the Mahajanga Basin, covering the area north of the town of Mahajanga, close to the delta of the Betsiboka River; and 2) the southern limb of the basin, including localities south of Mahajanga and to the west of the Mahavavy River (Mahajanga Province) (Fig. 2).

Morondava Basin
This basin covers a distance of about 1000 km on the western coast of the island, between Cape Saint-André in the north and Cape Sainte-Marie in the south, forming a band running subparallel to the central-southern coast along the Mozambique Channel ( Fig. 1). In view of its sheer size, it has been subdivided into geographic sectors, almost delimited from the north to south by the most important rivers which cross the basinal area more or less parallel from east to west (Besairie 1972; see Figure 3).

Southern
To date, Cenozoic decapod faunas from Madagascar are poorly known. In fact, the sole study is that by Collignon & Cottreau (1927), who described a marine Miocene (Burdigalian-Aquitanian) assemblage from Mahakamby Island, in the Mozambique Channel, south of the town of Mahajanga (northernwest Madagascar), comprising some incomplete brachyurans, referred to the portunoid Achelous. This fauna is currently the subject of review by Charbonnier et al. (in press). Finally, Rakotozafy & Goodman (2005) noted the presence of some indeterminate brachyuran chelipeds ("mangrove crabs") in subfossil deposits in the extreme southwest of the island, associated with a terrestrial fauna and archaeological remains.

MATERIAL AND METHODS
All specimens studied are preserved three-dimensionally, slightly compressed, and partially exposed on the surface of subnodular, irregular calcareous concretions. They are rarely articulated, occurring mostly as incomplete parts, floating in sediment due to the natural erosion. The epicuticle generally is poorly preserved. Some samples are preserved as a three-dimensional casts inside nodules. The  peculiar state of preservation and anatomical position of some macrurans suggests that these may represent moults. For a more detailed description some of the best-preserved specimens were partially prepared manually or, were the matrix proved too hard, with the help of mechanical engravers. Secrétan (1964: 53) used unpublished stratigraphic data, supplied by General Collignon, who collected and studied cephalopod faunas from the Morondava and Mahajanga Basins. Subsequently, papers on these assemblages (Collignon 1966(Collignon , 1969(Collignon , 1970a(Collignon , 1971) enabled some revised stratigraphic and geological interpretations, as summarized by Besairie (1972).Here, we follow the most recent stratigraphic data.The Belo-sur-Tsiribihina region (Berere, central Morondava Basin) (Figs 3; 5) in particular has been prospected by different researchers for many years. Different cross-sections ("Coupes"), more or less directed east to west, between different villages and different ages were named, each including beds ("couches"), with numerous layers ("gisements"), all numbered and marked on specimens collected from them. The same "Coupe" may thus include layers of different age along the section. In the descriptions below, we report the age for the type species and for some certain specimens, listing in the discussion the stratigraphic range of the taxon. Many of the specimens recorded in the present paper were not included in the list supplied by Secrétan (1964).
For some specimens it is difficult to determine the exact provenance, because layer numbering occasionally results in double sets amongst the various scholars who collected and catalogued material over the years; for complete references we refer to Besairie (1972). Moreover, some small villages were deserted or displaced, leading to changes in the original tracks and names during recent years (G. Pasini, pers. comm. 2010). For each specimen, we list the registration number, locality or section ("coupe"), number of level ("gisement"), age and zonation (when known or recognized), region and provenance (basin). Biozonations are those of Collignon (1959Collignon ( , 1960Collignon ( , 1962Collignon ( , 1963Collignon ( , 1964Collignon ( , 1965aCollignon ( , b, 1966Collignon ( , 1969Collignon ( , 1970aCollignon ( , 1971, based on cephalopod (ammonite) faunas (see also Walaszczyk et al. 2004).

discussion
We concur with Förster's (1966) diagnosis of Enoploclytia. Based on our revision of the Malagasy material referred to Enoploclytia by Secrétan (1964), we subscribe to this generic attribution, but wish to note that the line drawings in Secrétan (1964: figs 43, 44)

discussion
The type material included only fragmentary chelipeds that preclude a detailed characterisation of the species. The only features that can be recognized are: propodus massive, very large and compressed with inner and outer surfaces strongly tuberculate, dorsal margin finely tuberculate, ventral margin strongly tuberculate with two rows of more or less aligned strong tubercles; dactylus and index triangular in cross section; dactylus with two lateral strong basal spines. We question the attribution of this material to Enoploclytia in view of the fact that diagnostic characters of this genus are found mainly on the cephalothorax. Moreover, the chelae of Enoploclytia usually have a globose propodus with very elongate fingers, as seen in specimens of E. collignoni. Only the discovery of more complete specimens can resolve the systematic position of these specimens. diaGnosis. -Cephalothorax with deep cervical and antennal grooves; postcervical and branchiocardiac grooves parallel and slightly marked; branchiocardiac groove connected with deep hepatic groove; postcervical groove not joined to branchiocardiac groove and interrupted prior to the junction with the hepatic groove; inferior groove not well marked.
description Cephalothorax subcylindrical (CL = 25 mm, CH = 17 mm); very small rostrum (length c. 2 mm) with rounded distal extremity and smooth margins; ocular incision rather shallow; gastric and antennal region well delimited between cervical and antennal grooves; inflated antennal region; very narrow cardiac region; wide, rounded branchial region; cervical groove near-right with a weak depression at mid-height, slightly inclined intercepting dorsal margin at an angle of c. 65°; postcervical and bran- chiocardiac grooves strongly inclined intercepting dorsal margin at an angle of c. 40°; hepatic groove strongly rounded; inferior groove not well marked; posterior margin with small carina; ornament uniform with small tubercles. Abdomen with subrectangular somites of uniform length; pleurae 1 and pleurae 4 and 5 with spiny ventral margin; tail fan not preserved; surface of somites smooth with sparse pits.
Pereiopods fragmentary (three different segments preserved, difficult to attribute to pereiopod).

discussion
We concur with Förster's (1966) diagnosis of Eryma as well as with Secrétan's (1964) reference of the Malagasy material to this genus. Secrétan (1964) differentiated Eryma granulifera from E. madagascariensis on the presence of an anterior branch of the postcervical groove interrupted prior to joining the hepatic groove. Our re-examination of the type specimen of E. granuliferum has clearly shown that the postcervical groove is not divided into two branches. Moreover, the branchiocardiac groove is connected with the hepatic groove and not with the postcervical groove, as assumed by Secrétan (1964). Finally, the postcervical and branchiocardiac grooves are parallel, not fused. This configuration is similar in both E. granuliferum and E. madagascariensis, which is why we consider them to be conspecific, the name E. granuliferum being chosen as the valid name, on the basis of the Principle of the First Reviser (ICZN 1999: art. 24.2).  diaGnosis. -Cephalothorax with deep cervical, antennal, postcervical and hepatic grooves; branchiocardiac groove almost absent; postcervical groove joined to hepatic groove; well-marked inferior groove; pterygostomial region strongly inflated; well-developed antennal spine; two strong spines in the antennal region aligned with the antennal spine; short rostrum with smooth margins; fusiform area on the dorsal margin, posteriorly closed to the rostrum.
description Cephalothorax subcylindrical (CL = c. 36 mm, CH = c. 25 mm); rostrum very small (length c. 2 mm) with rounded distal extremity and smooth margins; well-marked ocular incision, ventrally delineated by a strong antennal spine; eyestalk well developed with globose eye; gastric and antennal region well defined between cervical and antennal grooves, antennal region with two well-developed spines directed forwards and aligned with antennal spine; cardiac region wide, bounded by cervical and postcervical grooves; pterygostomial region strongly inflated and tuberculate; branchial region wide and flat; cervical groove almost straight, slightly inclined intercepting dorsal margin at an angle of c. 55°; postcervical groove markedly deep and inclined, intercepting dorsal margin at an angle of c. 65°; hepatic groove strongly curved; inferior groove well marked; ornament uniform with well-developed tubercles. Pereiopods fragmentary (only two fragments of chelae, probably short); propodus globose with inner and outer margins tuberculate; index and dactylus with circular cross section, occlusal margin of a fragmentary dactylus with a small basal tooth, occlusal opening probably narrow. We here adopt Förster's (1966) diagnosis of Pustulina (= Phlyctisoma) and, after re-examination of the Malagasy material studied by Secrétan (1964), also agree with her generic attribution. However, we wish to point out that her reconstructions (Secrétan (1964: figs 37-39) are not correct in that: 1) the rostrum described in the text and illustrated as pointed and trifid is in fact rounded; 2) in the antennal region, only two spines (not three) are aligned, the first one is in fact the strong antennal spine; 3) the cervical groove does not show the small depression in the median portion; and 4) the groove "a", as shown in the line drawings is not the branchiocardiac groove, but the postcervical one, located not so close to the posterior margin of the cephalothorax; the short groove "c", as shown in the line drawings is the postcervical, which, in reality is more elongate, joining the hepatic groove.     margin up to mid-height, after becoming very weak, not really joining the weak and rounded hepatic groove; weak longitudinal branchial carina, parallel to the dorsal margin and located between the posterior margin and the postcervical groove; posterior margin slightly sinuous with a thin marginal carina; ornament: branchial region smooth, cardiac, gastric, and antennal regions finely tuberculate.
Abdomen with subrectangular somites; somites 1 and 6 smaller than others; somites 2-5 of uniform length; two different types of ornament on abdomen: 1) somites 1-6 with terga and pleurae completely smooth; and 2) somite 1 completely smooth, somites 2-5 with two pairs of dorsally aligned tubercles, somite 6 with dorsal surface uniformly tuberculate; pleurae 2 and 3 with two proximal tubercles; pleurae 4 and 5 with a single proximal tubercle; pleura 6 with one median tubercle. Two different kinds of telson: one with dorsal surface completely smooth and the other with two median tubercles. Uropods fragmentary, exopod with distal diaresis bearing small spines on the upper margin. Pereiopods fragmentary, only the propodi are well preserved; all are elongate, truncated conical, without apparent heterochely; specimen MNHN.F.R03941 shows ventrally two chelipedes associated with the carapace; two types of ornament occur on the lateral margins of chelipeds: 1) outer margin reinforced by a slightly tuberculate carina, inner margin by a carina of two rows of aligned strong spines; and 2) outer margin reinforced by a carina of two rows of aligned small spines, inner margin reinforced by a carina of two rows of aligned, strong spines; specimen MNHN.F.R03941 bears propodi of the first type.

discussion
We adopt the generic features outlined by Feldmann et al. (2007) for Hoploparia, and subscribe to Secrétan's (1964) assignment of the Malagasy material to it. Secrétan (1964) distinguished Hoploparia collignoni from H. sculpta and H. intermedia principally by: 1) the fact that cephalothoracic grooves were more or less well marked; and 2) the occurrence of different ornament of the inner and outer margins of the propodus and of the abdominal somites. However, we question the validity of Secretan's (1964) species for the following reasons: 1) all specimens of the three species come from the same outcrops; 2) the configuration of the cephalic grooves are quite similar in all three species, and also the location of the different spines on the gastric and antennal regions are similar; and 3) the propodus presents the same ornament of the inner and outer margins in these three species, irrespective of propodus size.
For these three reasons and also in view of the incomplete preservation of material of H. sculpta and H. intermedia, we here refer specimens assigned to those species to H. collignoni. Moreover, the differences observed in the ornament of the abdominal somites and of the propodus could either represent intraspecific variation and/or sexual dimorphism (Fig. 16). However, the specimens studied are not complete enough to be able to determine which of these interpretations is correct.  diaGnosis. -Cephalothorax with deep cervical groove starting at mid-height and joining the antennal groove; deep postcervical groove joining the well-marked hepatic groove; branchiocardiac and inferior grooves absent; gastric region with small spine; antennal region with tuberculate antennal carina; rostrum dorsally flat, flanked by two smooth lateral carinae, one strong, welldeveloped basal spine.
description Cephalothorax subcylindrical (CL = c. 19 mm, CH = c. 10 mm); rostrum dorsally flat with smooth margins and flanked by two smooth lateral carinae; one strong, well-developed basal spine; ocular incision shallow, delimited ventrally by the antennal spine; gastric region not delimited by the cervical groove, bearing one small spine located in the centre of the ocular incision; antennal region well demarcated by the cervical and antennal grooves, bearing a single postantennal spine aligned with the tuberculate antennal carina distally prolonged by the antennal spine; deep cervical groove starting at mid-height and joining the antennal groove; postcervical groove deep, slightly inclined, intercepting dorsal margin at an angle of c. 70°, starting in the first posterior ⅓ of the dorsal margin and clearly joining the well-marked hepatic groove; pterygostomial region strongly inflated; posterior margin slightly sinuous with a thin marginal carina; ornament uniformly finely tuberculate. Abdomen with subrectangular somites; somites 1 and 6 smaller than the others; somites 2-5 of uniform length; terga and pleurae uniformly tuberculate; pleurae with weak depression; telson subrectangular with two pairs of carinae converging distally, ornament similar to that of somites. Pereiopods fragmentary, only one propodus with carinate inner margin.
discussion Upon re-examination of the Malagasy material studied by Secrétan (1964), we agree with her generic attribution and with the erection of a new species. As pointed out by Secrétan (1964) and confirmed by our new observations, H. pusilla differs from H. collignoni in having a deep postcervical groove joining an equally deep hepatic one. Moreover, there is a difference in stratigraphic age between the two species.

discussion
Following the definition of ?Coleia by Van Straelen (1923, 1925, we doubt the generic attribution of the specimen studied by Secrétan (1964). In fact, the most important diagnostic characters of the genus are to be found on the cephalothorax and uropods. Since the single Malagasy specimen preserves only a small fragment of the posterior carapace and the abdominal somites, it is impossible to recognize the diagnostic generic characters. Although Secrétan (1964) pointed out that the three fragmentary carinae on the cephalothorax hint at ?Coleia, this feature can also be found in other palinuran taxa, such as Linuparus. The abdomen of the present specimen, considered typical of the eryonids by Secrétan (1964), can equally be assigned to other palinuran taxa. In conclusion, the characters displayed by the holotype are too general to ascribe it any fossil genus known to date.  Secrétan, 1964 (Figs 19-22) Linuparus bererensis Secrétan, 1964: 122-128, pl. 8, fig. 1 diaGnosis. -Cephalothorax with cephalic region smaller than branchial one. Cephalic region with: 1) two very strong supraorbital teeth directed upwards with two posterior basal spines; 2) wide frontal region without rostrum; 3) two dorsal carinae converging forwards into one strong spine; 4) two well-developed semicircular dorso-lateral carinae and strongly serrated; 5) straight lateral carina well developed, strongly serrated with very strong antennal spine delimiting the ocular incision; and 6) straight, smooth ocular incision; branchial region with: 1) three carinae well developed and strongly serrated; and 2) posterior margin of the cervical groove marked by small spines; posterior margin with submarginal groove of equal width both medially and laterally; abdominal somites with terga bearing two median spines; subrectangular, very strong somite 6 with two lateral depressions forming a rounded central part bearing two median spiny lines as well as two lateral parts slightly carinate with one small proximal spine and one very strong distal spine; subrectangular telson with two weakly curved lateral depressions forming a small rounded central part with two small median spines, lateral margins with strong distal spine, inferior margin near-straight and denticulate.
Branchial region with three well-developed, strongly serrated carinae: median carina (c. 7-9 spines, sometimes second proximal spines doubled) and lateral carinae (c.14-15 spines); posterior margin of cervical groove marked by small spines; posterior margin with submarginal groove of equal width medially and laterally and with median dorsal spine; pleural region subdivided by one longitudinal groove extending from the posterior margin to the end of the cervical groove: inferior portion strongly tuberculate, superior portion near-smooth excepted for the proximal part which is mostly tuberculate; lateral margins with possible stridulatory apparatus (= appareil stridulant sensu Secrétan 1964), located just posterior to cervical groove; dorsal surface finely tuberculate.
Antennal articles very strong and spiny. Proepistome with two median, aligned tubercles and one slight median depression.
Triangular mandibles well developed. Abdomen slightly convergent distally, somite 1 highly reduced, somites 2-5 of equal size, somite 6 wider than others; terga of somites 2-5 with two transverse grooves: proximal groove slightly sinuous medially, distal groove triangular in shape anteriorly; one strong spine close to the median concavity of the proximal groove, one strong spine located at the top of the posterior margin; median spines of somites 1-4 directed forwards, median spines of somite 5 backwards; pleurae of somites 2-5 with two aligned tubercles and trifid ventral margin: proximal spine slightly directed forwards, median spine larger than the others, distal spine slightly directed backwards; subrectangular somite 6 very strong, with two lateral depressions forming a rounded central part bearing two median spiny lines and two lateral, weakly carinate portions with a single small proximal spine and one very strong distal spine; subrectangular telson with two weakly curved lateral depressions forming a small rounded central portion with two small median spines, lateral margins with strong distal spine, inferior margin almost straight and denticulated; uropodal exopod and endopod equal in length with smooth lateral margins and rounded inferior margin, basipodite of uropods subrectangular.

discussion
We adopt the diagnosis of Linuparus given by Bruce (1965). After re-examination of the Malagasy material studied by Secrétan (1964), we subscribe to her generic attribution. Secrétan (1964) erected the subspecies Linuparus bererensis multispinosus, based on the number (nine) of spines on the median carina and a more elongate cephalothorax. We question the validity of this subspecies for the following reasons: 1) in her description of L. bererensis (Secrétan 1964: 124), nine spines were recorded on the median carina, which fact leads to confusion in distinguishing these taxa; 2) the median carina in small and large specimens of L. bererensis shows a variable number of spines (ranging from 7 to 9); and 3) the ratio between cephalothoracic length (CL) and width (CW) measured on several specimens and the spine count on the median dorsal carina of L. bererensis and subspecies multispinosus do not show distinct values to confirm Secrétan's proposition. The CL/ CW ratio of Malagasy specimens of Linuparus can be described as a linear model as has been observed for numerous fossil and extant crustaceans. Our plot diagram is relatively homogeneous and does not show any separate clusters that might be interpreted as distinct populations (Fig. 21). Moreover, we stress that the majority of the most complete specimens bear seven spines (68%) while only 18% and 14% have eight and nine spines respectively. In conclusion, with the number of spines of the median carina being so variable in the specimens studied and with cephalothorax proportions relatively homogeneous, these features are best considered to represent intraspecific variation and/or sexual dimorphism. Thus, they are not real diagnostic characters on which to erect a subspecies. In addition, all material stems from the same outcrops.   Secrétan, 1964 ( Fig. 18B-D) Ctenocheles madagascariensis Secrétan, 1964: 149-152, pl. 19, figs 1-4, 11-14;pl. 20, figs 9-16. -Collignon 1970b: 35. -Förster & Mundlos 1982: 154. -Feldmann et al. 1995: 8. -Feldmann et al. 2010: 340. -Schweitzer et al. 2010 discussion Upon re-examination of the Malagasy material studied by Secrétan (1964), we agree with her generic attribution. As noted by Secrétan (1964), the preservational state of the specimens studied allows confirmation of the presence of two different propodus types: one globose with a very elongate index and the other flat with an unknown index. Dactyli invariably are always fragmentary. However, because two kinds of propodus are typical of Ctenocheles, the lack of diagnostic characters, such as the number of teeth along the occlusal margins of specimens studied precludes a detailed diagnosis of the Malagasy species and a comparison with other Cretaceous forms. Although C. madagascariensis is still considered to be the oldest fossil species known to date (Schweitzer & Feldmann 2002), only the discovery of more complete specimens can refine its morphological features.  Secrétan, 1964 (Figs 23-25) Schuleria [sic] menabensis Secrétan, 1964: 138-143, pl. 13, fig. 1;pl. 14, figs 1-6;pl. 15, figs 1-6;pl. 16, figs 1-4;pl. 17, figs 3, 4. Schlüteria tuberculosa Secrétan, 1964: 143-146, pl. 12, figs 6, 7;pl. 13, figs 2-5;pl. 15, fig. 7;pl. 16, figs 5, 6;pl. 17  diaGnosis. -Cephalothorax with deep cervical groove delimiting a very wide branchial region and very narrow gastric region; markedly elongate rostrum, carinate dorsally; gastric region with two dorsally tuberculate carinae; P1 chela with short, stout palm and elongate dactylus and index incurved distally; small area covered by strong rounded tubercles at level of the propodus-dactylus articulation; occlusal margin of dactylus with one strong, very elongate proximal tooth, one strong median tooth, and a few pointed distal teeth in random arrangement; occlusal margin of index with a row of strong, pointed teeth in a uniform arrangement; P2 chela with short, inflated palm with smooth occlusal margin; uropodal exopod without diaresis.

Schlueteria menabensis
description Cephalothorax subrectangular (CL = 36-56 mm, CH = 23-40 mm); very elongate rostrum with smooth median carina; weak ocular incision delimited only by rounded antennal margin; gastric region with two dorsally tuberculate or smooth carinae curved, and converging to the base of the rostrum, the inferior one joining the lateral margin of the rostrum; area between the two curved carinae covered by aligned transverse small tuberculate lines; smooth antennal region; deep cervical groove arising in the first anterior quarter of the dorsal margin, strongly inclined, intercepting dorsal margin at an angle of c. 50°; cervical groove reaching the anterior ventral margin creating a strong inflection; cervical groove delimiting a very broad branchial region and very narrow gastric region; ornament of small pits uniformly distributed.
Abdomen with subrectangular somites; somite 1 smaller than others; somite 6 larger than preceeding ones; somites 2-5 of equal length; pleurae with smooth inferior margins; two different types of ornament on the abdomen: 1) terga of somites 1-6 without dorsal carina, pleurae with one lateral carina inducing a small ventral depression; and 2) terga of somites 1 and 6 without dorsal carina, terga of somites 2-5 with dorsal carina, subtriangular, very wide subtriangular pleurae with two carinae delimiting central depression; subrectangular telson with slightly sinuous lateral margins bearing two small spines, dorsal margin with two aligned tubercles.
P1 bearing very massive homochelous chelae; propodus with strongly tuberculate dorsal margin and ventral margin bearing a carina of strong aligned spines directed forwards and decreasing in size anteriorly; very strong spine located at the base of articulation of the dactylus; small area covered in strong, rounded tubercles at level of propodusdactylus articulation; occlusal margin of dactylus with one strong, very elongate proximal tooth, one strong median tooth, and a few pointed distal teeth in random arrangement; occlusal margin of index with a row of strong, pointed teeth in uniform arrangement; opening very broad and curved with occlusal margins of dactylus and index not in contact; dorsal margin of dactylus bearing a median tuberculate carina; outer surface of propodus with uniformly distributed small and large tubercles; inner surface of propodus with aligned, large tubercles alternating with small ones; P2 chela well developed, with short, inflated palm; short dactylus and index with smooth occlusal margins; very elongate, thin P4-5 achelate. discussion Following re-examination of the Malagasy material studied by Secrétan (1964), we agree with her generic attribution. Secrétan (1964) distinguished Schlueteria tuberculosa from S. menabensis principally by (1) spiny carinae in the gastric region and (2) a dorsal carina on the terga of somites 2-5 in the former. We doubt the validity of S. tuberculosa for the following reasons: 1) all specimens of the two "species" originate from the same outcrops; 2) the configuration of the cephalic grooves is quite similar in both; and 3) P1 chelae are morphologically similar in both, irrespective of propodus size, as confirmed also by the type material of S. tuberculosa which has identical chelipeds to those of S. menabensis.
For these three reasons and in view of the incomplete state of preservation of material of S. tuberculosa, we synonymise both. Moreover, differences in ornament of carinae on the gastric region and on abdominal somites could be seen as reflecting intraspecific variation and/or sexual dimorphism. However, the material is not enough complete to determine which of these interpretations is correct. Because S. carinata Taylor, 1979, was described on the basis of a poorly preserved specimen, retaining only the P2 (Taylor 1979: 23, pl. IVa), Schlueteria menabensis represents, together with the type species S. tetracheles, the most complete member of the genus known to date. The original type of the latter, housed in the collections of the National Museum, Prague (Czech Republic), has allowed us to compare this with the Malagasy species, so as to outline features in common as well as different characteristics. The two share the following characters: two longitudinal, parallel ridges on the cephalic region of the carapace; deep cervical groove strongly inclined forwards, creating a cephalic region which is smaller than branchial region; P2 chela short, stout, with inflated palm; pleurae of abdominal somites ending in a point; smooth subrectangular telson; smooth uropods; uropodal exopod without diaresis. Although the original sample of S. tetracheles comprises a few more or less complete specimens we have been unable to observe the rostrum that could have been thin and elongate as in the Malagasy species. The main difference between the two species is the structure of P1 chela: in S. tetracheles (Fig. 24I), it is strong with a short, stout palm and elongate dactylus and index; the dorsal margin of the chela has a few variably sized spines in a random arrangement, while the ventral margin has a row of strong spines in a uniform arrangement, disappearing prior to the distal extremity; the occlusal margin of the dactylus has two strong conical teeth located proximally and medially and some strong pointed teeth located distally; the occlusal margin of the index has a row of uniformely arranged strong, pointed teeth; the surface of the chela is strongly tuberculate, while that of the dactylus has two rows of aligned tubercles; 4-5 tubercles aligned at the level of propodus-dactylus articulation; in S. menabensis the chela is strong with short, stout palm and elongate dactylus and index; the dorsal margin of the chela is smooth, while the ventral margin has a carina of strong, aligned spines directed forwards and decreasing in size anteriorly; the occlusal margin of the dactylus has a single strong, very elongate proximal tooth, one strong median tooth, and a few pointed distal teeth randomly arranged; the occlusal margin of the index has a row of strong, pointed teeth in uniform arrangement; the surface of the chela is strongly tuberculate with a row of tubercles aligned along the occlusal margin of the index; a small area is covered by strong, rounded tubercles at the level of the propodus-dactylus articulation. The main features observed in both species confirm the reconstruction of S. tetracheles by Fristch in Fristch & Kafka (1887).
In view of the above, we consider it important to select a lectotype for Schlueteria menabensis because Secrétan (1964) did not designate a holotype for it. This is MNHN.F.R03920, from Berere (gisement 208, Middle Santonian) which shows almost all the diagnostic characters of both species and genus. discussion Secrétan (1964: 187-191, text-figs 112, 114) described and illustrated a single incomplete specimen of this species. Judging from the original description and the line drawing (reconstruction), in comparison with the specimen itself, we noted inconsistencies such as the orbital area and the interpretation of the outline of the dorsal bosses. Recently Schweitzer et al. (2007) have discussed and reviewed Titanocarcinus, listing all species referred to this genus and their current status. Those authors pointed out that T. mamillatus showed a development of carapace regions which was similar not only to that of Titanocarcinus, but also to members of other Cretaceous xanthoid families, such as the Palaeoxanthopsidae. Concomitantly, they noted that the anterolateral margin in the Malagasy species was incomplete and that the fronto-orbital width was considerably less than the maximum width in other species of Titanocarcinus, which led those authors discussion Secrétan (1964: 178) described a single fragmentary carapace, as a new species of Xanthosia Bell, 1863, X. elegans. As this name appeared to be preoccupied, the replacement name X. robertsi was later proposed by Secrétan (1982). However, as noted by Schweitzer Hopkins et al. (1999)

discussion
Although Secrétan (1964) gave a full description of this specimen, its state of preservation precludes systematic assignment, other than "xanthid".

CONCLUSIONS
We have revised 13 species described by Secrétan (1964) on the basis of a collection of more than 1500 specimens and presented an updated systematic list of Malagasy decapod crustaceans (Table 1). It is apparent that assemblages of Mesozoic crustaceans from Madagascar are remarkable for their diversity and relative abundance. Unfortunately, these faunas are known exclusively from samples in relatively old collections; collection of new material appears highly rewarding in this respect.