Bioacoustics of Sesarmidae

Genetic sequences by Schubart (2011) , with Varuna litterata as outgroup, and by Ip et al. (2015) indicate a close relationship of Leptograpsodes to Xenograpsidae , even a possible sister-taxon relationship with Xenograpsu s. The question is therefore to test whether the xenograpsid morphology provides the accurate traits to corroborate the DNA data or whether there is another case of incongruence between the currently adopted classification and molecular phylogenetics. Xenograpsus testudinatus was shown to have distinct gene order ( Ki et al. 2009 ). According to the gene order analysis of brachyuran mitochondrial genomes by Basso et al. (2017) , in which Brachyura appears to be a hot-spot of gene order diversity within the phylum Arthropoda, XenGO is an apomorphy currently known only for X. testudinatus . According to Wang et al. (2018b: 39 , fig. 6), based on the complete mitochondrial DNA (mtDNA) sequences, Xenograpsus testudinatus and Sesarmidae was found forming a group with close relationships. The morphological characters of Xenograpsus testudinatus have been studied and illustrated by N. K. Ng (2006) and N. K. Ng et al. (2007) . The comparative characters are: proepistome inserted as thin tongue into lower margin of front (N. K. Ng et al. 2007 : figs 3B, 6A) (also thin tip in Leptograpsodes ); orbit almost complete, in the form of deep sunken cavity (N. K. Ng et al. 2007 : fig. 5A) (partially opened laterally in Leptograpsodes ); suborbital margin not stridulatory (N. K. Ng et al. 2007 : fig. 5G) (suborbital margin with stridulatory function in Leptograpsodes ); mxp3 broad, operculiform, lacking rhomboidal gap (N. K. Ng et al. 2007 : fig. 7A) (gaping in Leptograpsodes ); typical pleonal-holding mechanism present: rather big sternal button and deep pleonal socket (as in Leptograpsodes ); G1 strongly calcified, short and stout (long, quite slender, with blunt, horny tip and flap on sternal surface in Leptograpsodes ).

COMPARISON WITH LEPTOGRAPSODIDAE N. FAM The traditional characters that may be used to differentiate Leptograpsodidae n. fam . from Varuninae could be, e.g. carapace rounded, with very convex margins in Leptograpsodidae n. fam . (in Varuninae varying from quadrate to quadrangular, lateral margins varying from subparallel to moderately convex); front about half width of carapace (vs less than half maximum carapace width); two pairs of postfrontal lobes (vs without four postfrontal lobes); pterygostome, subbranchial, subhepatic regions setose (vs glabrous); suborbital crest stridulatory (vs not stridulatory); mxp3 largely gaping when closed, with antero-external angle of merus not produced, (vs small, rhomboidal gape; antero-external angle of merus strongly produced and auriculate).

Other characters of the subfamily are: sternites 1 and 2 fused (no suture 1/2), sternite 1 narrow and distinctly tapering at tip; suture 2/3 present but thin ( Fig. 9 J-K); suture 3/4 absent; sutures 4/5-7/8 interrupted; sternite 8 developed; median line visible on long portion of sternite 8 and extending on sternite 7 ( Fig. 9K, L ); episternites 5 and 6 very narrow and long; male gonopore in middle of sternite 8, far from suture 7/8 and from P5 coxa, a large part of sternite 8 being intercalated between gonopore and P5 coxo-sternal condyle; proepistome not forming developed structure ( Bouchard et al. 2013 : fig. 26B, D); stridulatory apparatus: suborbital crest separated in at least three portions (pars stridens) rubbing against marked thickening (plectrum) on inner edge of cheliped merus ( Fig. 11B, C ) ( Tweedie 1954 ; Davie & N. K. Ng 2007 ); press-button mechanism typical. COMPARISON WITH LEPTOGRAPSODIDAE N. FAM Leptograpsodidae n. fam . shares with Thalassograpsinae the presence of a stridulatory apparatus, however different. In Leptograpsodes suborbital crest is composed of homomorphic granules and the plectrum is a delimited ridge on P1 merus ( Figs 1G ; 2C, D ; 3D ); in Thalassograpsus harpax the suborbital crest consists of heteromorphic tubercles and the plectrum is only a thickening on P1 merus, see Fig. 11B, C ).

DIAGNOSIS Carapace subquadrate, lateral margins varying from subparallel to strongly convergent, usually armed with two epibranchial teeth behind exorbital tooth. Front much narrower than half exorbital width; exorbital width not markedly exceeding carapace length. Orbit long, not closed laterally. Eyestalks not nearly as broad as front. Proepistome as wide triangular or pentagonal plate, its very narrow tip inserted under front ( Tavares 1989 : fig. 21, as Chasmagnathus granulatus Dana 1851 , now Neohelice granulata ). Stridulatory apparatus usually present (e.g. Helice , Helicana , Chasmagnathus ): pars stridens formed by suborbital ridge composed of granules or tubercules, sexually dimorphic (ridge often heteromorphic in males), rubbed by prominent horny ridge on inner margin of cheliped merus; may be absent. Reticulation on pterygostomial regions poorly developed. Thoracic sternum wide; sternite 1 separated from sternite 2 by convex ridge; suture 2/3 straight; sternite 8 developed, exposed medially, not deeply notched by posterior emargination. Deep median line usually extending from sternite 8 to sternites 6 and even 5, but intermittently. Male gonopore in posteriormost location in relation to suture 7/8. G1 with suture of sperm channel usually torsioned towards dorsal face, except in Chasmagnathus .

So varied is the subfamily Cyclograpsinae that it is difficult to make general statements. Cyclograpsus seems to be the closest member to Leptograpsodes , thus our study will first focus on this genus. As Cyclograpsus , a genus that probably exceeds 20 worldwide distributed species (Davie & N. K. Ng 2007 ; Ng et al. 2008 ; Naruse & N. K. Ng 2012; Naruse 2015 ), is probably not monophyletic, we will restrict our study to the type species C. punctatus (by subsequent designation of Rathbun 1918: 328 ) and only its allies. They show a small orbit, a real infraorbital margin lacks, and a variously ornamented (granules, tubercles, elongated structures) suborbital crest is present. In cyclograpsines, the strongly developed and suborbital crest can be stridulatory in conjunction with cheliped merus, which is either only thickened and without a differentiated structure, or bears a horny ridge, usually sexually dimorphic ( Tesch 1918 ; Guinot-Dumortier & Dumortier 1960 ; Sakai et al. 2006 ; Davie et al. 2015a ). See Stridulatory structures in grapsoids. The biology of Cyclograpsus punctatus , a species restricted to southern Africa (see Rathbun 1918 ; Barnard 1950 ; Garth 1957 ; Griffin 1968 ; Fagetti & Campodonico 1971 ; Emmerson 2016 ), provides some similarities with that Leptograpsodes octodentatus : it develops aerial respiratory adaptations, so is highly resistant to desiccation, lives semi-terrestrially, enters estuaries and diggs more or less elaborate burrows in the muddy banks of the rivers ( Broekhuysen 1941 ; Alexander & Ewer 1969 : fig. 3b; Emmerson 2016 ). Leptograpsodes octodentatus lacks the fine reticulated pattern of short curved setae that entirely covers the pterygostome and results in a fine film of water over the setae, characteristic of Cyclograpsus punctatus and sesarmids ( Alexander & Ewer 1969 : fig. 4). COMPARISON OF C. PUNCTATUS AND LEPTOGRAPSODES The comparaison between Leptograpsodes octodentatus (especially when small specimens are examined, see Fig. 2 A-C) and Cyclograpsus punctatus , which is a relatively small species, reveals a lot of distinguishing characters. Leptograpsodes differs from C. punctatus by, e.g. carapace margin strongly convex and with three weak teeth (only slightly convex, almost straight, with margin subcristate, entire or indistinctly notched in C. punctatus ); anterolateral margin normal (carinate and ventrally hollowed in C. punctatus ); presence of several postero-lateral striae (at most 2-3 faint ridges obliquely connecting to posterolateral margin in Cyclograpsus ); two pairs of frontal lobes (absent in C. punctatus ); carapace dorsal surface grooved (flat and smooth in C. punctatus ); developed suborbital tooth (absent in Cyclograpsus ); sunken, faint, finely granulose suborbital margin, continuous to near inner limit of orbit and paralleled by developed stridulatory suborbital crest (a thickening encircling and surrounding base of eyestalk in middle of orbit, eclipsed by prominent suborbital crest having a stridulatory role in C. punctatus ); shape of epistome; mxp3 merus without oblique setose crest (with oblique setose crest in C. punctatus ); pterygostome with shallow, simple vertical groove, parallel to buccal cavity (without vertical groove in C. punctatus ); subhepatic, pterygostomial and subbranchial regions covered by pubescence lacking reticulate pattern (reticulated network of short hooked setae in C. punctatus ). The shape of the orbit, with a stridulating suborbital crest in both genera, merits careful comparison: a close scrutiny reveals their discrepancy. In Cyclograpsus punctatus the orbit is small and closed laterally, a real infraorbital margin lacks and there is only a strongly developed, coarsely granulated suborbital crest that acts in conjunction with variously thickened inner margin of cheliped merus. In Leptograpsodes the large orbit is opened laterally, there is a faint but very distinct suborbital margin, with, in addition, subparallel to and below, a suborbital crest consisting of numerous, very fine and regular granules; this crest (pars stridens) is stridulatory in conjunction with a smooth, delimited crest (plectrum) at inner face of cheliped merus. The status of all other Cyclograpsus is uncertain. That Cyclograpsus is not monophyletic was shown by the specific disparity of larval features ( Gore & Scotto 1982 ; Cuesta & Rodríguez 1994 ) and of locking-pleonal mechanism ( Guinot & Bouchard 1998 ; Köhnk et al. 2017 ). As a result, all species assigned to Cyclograpsinae need a reevalution, as Naruse & N. K. Ng (2012) has done for Cyclograpsus lophopus Nobili, 1905 that revealed to be a sesarmid and became Cyclorma lophopus (Nobili, 1905) . C. granulosus H. Milne Edwards, 1853 , a conspicuous species occurring on rocky shorelines in southeastern Australia including Tasmania ( Campbell & Griffin 1966 ; Griffin 1971 ), and known by its completely sequenced mitochondrial genome ( Tan et al. 2016 ), is morphologically quite distinct from Leptograpsodes octodentatus . The junction of the proepistome to the front is rather similar in the two genera. In Leptograpsodes sternal suture 1/2 strongly marked, sternites 1 +2 semi-ovate; suture 2/3 straight (in C. punctatus sternite 1 individualised, with thick ridge at the location of suture 1/2, see Fig. 9A, B ); in Leptograpsodes sternite 8 only developed laterally, a minute portion visible posteriorly when pleon is folded ( Fig. 3C ) (in C. punctatus sternite 8 developed, medially exposed, and with large portion visible dorsally when pleon is folded); in Leptograpsodes male gonopore close to P5 coxa and far from suture 7/8; penis long, emerging on sternite 8, not very far from suture 7/8, see Fig. 3B ) (in C. punctatus male gonopore far from P5 coxa; penis short, see Fig. 9C ).

The East African C. eulimene and C. ortmanni are unique in that the dorsal margin of the male palm has a distinct oblique pectinated ridge and the dorsal margin of the chelar dactylus has differentiated granules, presumably for stridulation. These characters are consistent and are here considered to be generically significant. To this effect, we propose to establish a new genus, Cristarma , for these two species. With regard to these two characters, Cristarma is similar to the two known species of Selatium Serène & Soh, 1970 , viz. S. brockii ( De Man, 1887 ) and S. elongatum (A. Milne-Edwards, 1869 ) (cf. Schubart et al., 2009 ). Cristarma can easily be distinguished from species of Selatium in that the pectinated ridge on the male palm is markedly oblique to almost transverse in orientation ( Fig. 15D, F ) (versus longitudinal), the outer lower surface of the palm has two short, smooth, submedian oblique ridges ( Fig. 15C, E ) (versus absent), the lateral margins of the ambulatory dactyli are not lined by dense short black setae ( Fig. 13E, F ), and the surfaces between the first to third ambulatory coxae have dense tufts of long setae ( Fig. 18C, D ) (versus very short, scattered, or no setae). In addition, the vulva is very low, with the two sternal vulvar covers bracketing the non-projecting opening ( Fig. 43G, H ). Our molecular phylogenetic tree also shows that there is no close relationship between Selatium and Cristarma or those genera with two transverse pectinated crests, e.g., Parasesarma and Perisesarma ( Fig. 59 ). The other two species endemic to Africa, but from the Atlantic coast, C. angolensis and C. buettikoferi , are referred to the new genera, Trapezarma and Platychirarma , respectively. Like Chiromantes s. str. , their epigastric and postfrontal crests are sharp and almost continuous ( Figs. 13G, H, J , 14G, H ), a character shared with no other group of species. Unlike the West African C. eulimene and C. ortmanni that have pectinated crests on their chelae and granules on the dactylar fingers, the chelae of C. angolensis and C. buettikoferi are smooth and not specially ornamented. What is unusual about their adult male chelae is that the outer surface is gently convex to flat (including the pollex) ( Figs. 16 , 17 ), being most extreme in C. buettikoferi . The prominently flattened outer surface of the chela of C. buettikoferi ( Fig. 17 ) is a character shared only with Pseudosesarma bocourti and C. cheirogonum from Southeast Asia ( Fig. 28 ) (both referred to a new genus, see later). The carapaces of C. angolensis and C. buettikoferi are quite distinct with a more trapezoidal appearance, with a wider frontal margin and strongly converging lateral margins towards the posterior carapace margin ( Fig. 13G, H, J ). While the two species are genetically defined as sister species (despite long individual branches, Fig. 59 ), the suite of cheliped, male pleonal and G1 characters argue against including them in the same genus (see discussion under Trapezarma ). This gains further support from the very differently shaped vulvae, those of Trapezarma widely spaced and the projecting opening directed obliquely posteriorly ( Fig. 43I ), while the ones of Platychirarma are closer together and the slightly projecting opening is directed obliquely anteriorly ( Fig. 43J ). Ng & Liu (1999: 230) had noted that the generic placement of Chiromantes obtusifrons should be reappraised. The trapezoidal carapace of this species is diagnostic, although it shares other characters with Orisarma , new genus . However, C. obtusifrons has a peculiar frontal margin that is very broad and bends downwards so strongly that the antennae and antennules are effectively covered ( Fig. 14A–D ), the outer surface of the palm has no longitudinal ridge ( Fig. 15A ), and the surfaces between the first to third ambulatory coxae have distinct dense tufts of long setae ( Fig. 18A, B ) (a character otherwise only present in C. eulimene and C. ortmanni ). In addition, the vulvae have a totally different structure, with the sternal vulvar covers being plate-like and the opening not projecting ( Fig. 43E, F ). Davie & Ng (2012) re-examined Chiromantes obtusifrons and recognised four new closely allied species from the Indian Ocean and Western Pacific, and also commented on their generic status. All five species are here referred to a new genus, Danarma . Ng & Liu (1999) had already described Sesarma ( Sesarma ) stormi De Man, 1895 , in detail and explained at length why it should be placed in its own genus Stelgistra Ng & Liu, 1999 . There is no need to elaborate here, as all the characters discussed then are still valid in defining Stelgistra . Stelgistra stormi has since been found in Guam ( Paulay et al., 2003: 508 ). The revised classification of the above-discussed species is summarised in Table 2 .