Tubuca alcocki, a new pseudocryptic species of fiddler crab from the Indian Ocean, sister to the southeastern African T. urvillei (H. Milne Edwards, 1852) (Crustacea, Decapoda, Brachyura, Ocypodidae)

Abstract A new pseudocryptic species of fiddler crab, Tubuca alcocki sp. n., is described from the northern Indian Ocean. The new species was previously identified with T. urvillei (H. Milne Edwards, 1852), but can be distinguished by the structures of the anterolateral angle of the carapace and male first gonopod. The molecular data of the mitochondrial cytochrome oxidase subunit I gene shows that both are sister taxa and the divergence time is estimated at 2.2 million years ago, around the beginning of the Pleistocene. While the new species is widely distributed in the northern part of Indian Ocean, occurring from the Red Sea to India and the Andaman Sea; T. urvillei sensu stricto has a more restricted range, and is known only from southeastern Africa.

Of these taxa, Tubuca urvillei is a large-sized species, which has been widely reported from throughout the Indian Ocean and is the only Tubuca Bott, 1973 species known in the western Indian Ocean (Crane 1975;Shih et al. 2016). Aspects of its biology has also been investigated in southeastern Africa (e.g., Macnae 1963;Hartnoll 1975;Litulo 2005;Peer et al. 2015), Thailand (e.g., Jaroensutasinee et al. 2003;Jaroensutasinee and Jaroensutasinee 2004), and Pakistan (Ghory and Siddiqui 2006).
In this study, specimens from the range of Tubuca urvillei, including the type specimens, were examined. There are two clades, with small but consistent morphological differences supported by DNA evidence from cytochrome oxidase subunit I (COI). The material from the northern and eastern parts of the Indian Ocean is herein described as a new pseudocryptic species, T. alcocki.

Materials and methods
Specimens of Tubuca urvillei sensu lato collected from southeastern Africa, India and western Thailand examined (including the types) are deposited in the Muséum national d'Histoire naturelle, Paris, France (MNHN); Zoological Collections of the Department of Life Science, National Chung Hsing University, Taichung, Taiwan (NCHUZOOL); Senckenberg Museum, Frankfurt am Main, Germany (SMF); and Zoological Reference Collection of the Lee Kong Chian Natural History Museum (formerly Raffles Museum of Biodiversity Research), National University of Singapore, Singapore (ZRC). The abbreviation G1 is used for male first gonopod. Measurements, all in millimeters (mm), are of the maximum carapace width (CW), carapace length (CL) and pollex length (PL). The terminology used essentially follows Crane (1975) and Davie et al. (2015).
Sequences of COI were obtained following the method described by Shih et al. (2016), after verification with the complimentary strand. Sequences of the different haplotypes have been deposited in the DNA Data Bank of Japan (DDBJ) (accession numbers in Table 1). According to Shih et al. (2016), T. urvillei is sister to the clade composed of T. dussumieri (H. Milne Edwards, 1852), T. paradussumieri (Bott, 1973) and T. capricornis (Crane, 1975). As a result, the sequences of these three species, as published in Shih et al. (2016) (LC150436, LC053373 and LC150430), are used as outgroups in this paper.
Remarks. In his revision of the genera and subgenera of the fiddler crabs of the world, Bott (1973) established Tubuca and designated Gelasimus urvillei H. Milne Edwards, 1852 as the type species from the lectotype (Bott 1973: fig. 11). The type specimens of Tubuca urvillei were supposedly collected from "Vanikoro" (an island between Solomon and Vanuatu) in the western Pacific. Crane (1975) queried this type locality noting that the species as she understood it did not occur outside the In-   dian Ocean. As such, Crane (1975) considered the data on the label to be wrong. Of the three extant type specimens of Gelasimus urvillei H. Milne Edwards, 1852, Crane (1975 selected the male as the lectotype, the other two females becoming paralectotypes ( Fig. 1F-G). Crane (1975) considered the male to be an immature specimen (CW 18.5 mm) but its G1 is actually already developed (present study). According to Litulo (2005), the smallest ovigerous female from Mozambique is only CW 10.0 mm. This suggests that the lectotype male, while small is probably already mature. In any case, the G1 of the lectotype of T. urvillei (Crane 1975: fig. 9D) agrees well with the species as is now understood from southeastern Africa (cf. Fig. 5E-H). They also agree in all other morphological characters.
A note on Gelasimus dussumieri H. Milne Edwards, 1852 (at present Tubuca dussumieri) is necessary. The type material of Tubuca dussumieri include specimens from Samarang (Java, Indonesia) and Malabar (Mumbai, India) (H. Milne Edwards, 1852), and as no holotype was originally selected, Crane (1975) designated a male from Samarang as the lectotype of T. dussumieri. The paralectotype male from Malabar, however, she reidentified as T. urvillei instead. She also found that T. dussumieri and T. paradussumieri were sympatric in the western Pacific and eastern part of Indian Ocean. She reidentified all the records (including "T. acuta") from western Indian Ocean as T. urvillei, with one exception -the record of G. dussumieri by Hoffmann (1874: pl. 3(22)) and De Man (1891) from Nossy Faly, northern Madagascar, which was referred to T. paradussumieri instead. As no other record of T. paradussumieri from eastern Africa has been reported since 1874 (Crosnier 1965), Crane (1975) regarded this specimen's provenance as questionable. Another record of "T. dussumieri" from Bombay, western India (Krishnan 1992) will also need to be confirmed in the future as well. In summary, Crane (1975) emphasized the westernmost distribution of the genus Tubuca (= Deltuca Crane, 1975) should be T. urvillei from southeastern Africa (Tanzania, Madagascar and South Africa), with the species also present in Pakistan and western India. Later, the species was reported from the Red Sea by Hogarth (1986) and Price et al. (1987).
With regard to the records of T. urvillei and T. acuta in Alcock (1900), Crane (1975: 61) considered only those from Pakistan and western India as belonging to true T. urvillei (shown as "(part)" behind these records). That is, she did not think or was uncertain if the records from the Bay of Bengal and the Andaman Sea (e.g. Madras; Sunderbunds; Mergui; Andamans and Nicobars) by Alcock (1900) were also T. urvillei. Lundoer (1974) added a new record of "U. angustifrons (De Man, 1892)" from Phuket, Thailand, but this was later reidentified as T. urvillei by Frith et al. (1976) and Frith and Frith (1977a) (see also Frith and Frith 1978;Brunenmeister 1980, 1983;Jaroensutasinee et al. 2003;Jaroensutasinee and Jaroensutasinee 2004 (Figs 3A, 4A, 6A, C, G, 7A, C, E) trapezoidal, smooth; front narrow, with distinct, narrow median groove; anterolateral angle acutely triangular, directed obliquely anteriorly; anterolateral margin short to moderately long; dorsolateral margin long, definite, strongly converging; one posterolateral stria. Floor of orbit with row of 5-11 tubercles, sometimes with blunt ridge. Major cheliped (Figs 4C, 6B, D) with dactylus usually longer than palm, outer surface of dactylus and pollex each with 1 long groove proximally extending beyond midlength. Fingers of minor cheliped (Figs 3B, D, 6B, D) without conspicuous tooth on either finger. G1 (Fig. 5A-D) with distal tube slender, slightly curved to almost straight, distal and proximal parts subequal in width; thumb of moderate length, extending beyond base of distal tube.
Colouration in life. Adults with carapace and legs brown or dark brown, posterior part gray, especially in females (Fig. 6A, C, E). Some females with anterolateral angles orange (Fig. 6E, F) or with dark blotches on blue carapace (Fig. 6G). Major cheliped with fingers white; lower palm deep yellow in large individuals, orange in young individuals; upper palm brown (Fig. 6B-D). Females sometimes with minor chelipeds orange, sometimes with tint of blue (Figs 3F, 6F, G).
Etymology. This species is named after Alfred William Alcock, who first recorded this species from India and Pakistan as "Uca urvillei" (cf. Alcock 1900).
Distribution. Western Thailand, India, Pakistan, and the Red Sea (see Remarks). Remarks. Although the number of tubercles on the floor of orbit and thumb morphology of G1 are sometimes useful for distinguishing species of fiddler crabs, they are too variable in Tubuca alcocki sp. n. and T. urvillei (Crane 1975: 58-59; this study) to be used. The two species are similar, but can be morphologically distinguished by the characters of the anterolateral angle of the carapace and G1. The anterolateral angle in male T. alcocki is acutely triangular and directed obliquely anteriorly (Fig. 7A, C, E) (vs. relatively broadly triangular in shape and directed more laterally in position in T. urvillei; Fig. 7B, D, F). In female T. alcocki, the anterolateral angle is broadly triangular in shape and the anterolateral margin is relatively longer and curves gently to join the dorsolateral margin (Fig. 7G) (vs. anterolateral angle acutely triangular in shape with the anterolateral margin short and merging with the dorsolateral margin in an almost straight line in T. urvillei; Fig. 7H). The G1 structure is also different. The distal tube of the G1 of T. alcocki is proportionately more slender, being slightly curved to almost straight, with the widths of the distal and proximal parts subequal (Fig. 5A-D) (vs. distal tube relatively stouter, more distinctly curved and gently tapering towards the tip, with the distal part distinctly narrower than the proximal part in T. urvillei; Fig. 5E-H). Crane (1975) figured specimens of what she referred to as T. urvillei from southeastern Africa and western India, and they agree with the characters of T. urvillei and T. alcocki, respectively. The anterolateral angles of the male lectotype of T. urvillei (Fig. 1A;Crane 1975: pl. 9E) and the male specimen from Tanzania (Crane 1975: fig. 7A) are both broadly triangular. In addition, the G1 distal tubes of the lectotype of T. urvillei as well as those from Somalia and Madagascar figured by Vannini and Valmori (1981: fig. 6F) and Crosnier (1965: figs 195-196) are all relatively stout, curved and tapering towards the tip. As such the material from Tanzania, Somalia and Madagascar should all be referred to T. urvillei sensu stricto.
The specimen from Malabar, western India, and one of the paralectotypes of Gelasimus dussumieri (see discussion earlier), have the G1 distal tube relatively more slender, almost straight, with the distal and proximal parts subequal in width (Crane 1975: fig. 9E) and are thus is clearly referable to T. alcocki. The G1 structures of specimens from Pakistan (Saher 2008: fig. 2.2;Tirmizi and Ghani 1996: fig. 39) also match that of T. alcocki. Interestingly, Hogarth (1986) reported "Uca (Deltuca) urvillei" from the Red Sea, which was a new record of this species for this region, but without any figure or description. The first author has examined specimens from the Red Sea and they are clearly T. alcocki as well (H-T Shih and BA Kumar, in preparation). The distribution of T. alcocki thus stretches from the northern part of the Indian Ocean (Red Sea) to the Arabian Sea and Andaman Sea.
There are also colour differences between T. urvillei and T. alcocki. While the colouration of females, young males, and juveniles are variable in Tubuca species, the colouration of the adult male carapace is generally more useful (Crane 1975;von Hagen and Jones 1989;Beinlich and von Hagen 2006). Adult male T. urvillei sensu stricto have various degrees of blue on the carapace and ambulatory legs (Fig. 2D, E), with the palm of the major cheliped ochraceous to apricot brown (Fig. 2D); while young and females sometimes have pale and dark blotches on a blue background (Fig. 2F). In adult male T. alcocki, the dorsal surface of the carapace is always dark brown (Fig. 6A, C) whereas in T. urvillei, it is always blue (Fig. 2D, E).

DNA analyses and discussion
The molecular analyses include 12 specimens of Tubuca alcocki sp. n. from western Thailand and western India; and three specimens of T. urvillei from southeastern Africa ( Table 1). The phylogenetic tree (Fig. 8) based on COI shows that specimens from southeastern Africa form a distinct clade, sister to another clade with material from western India and western Thailand. The genetics therefore supports the recognition of two species. Only one haplotype is found from T. urvillei from southeastern Africa, with two haplotypes from T. alcocki.
Tubuca alcocki and T. urvillei are quite similar in general morphology, but can be still distinguished by characters of the carapace and G1 (see Remarks under T. alcocki), which is supported by molecular evidence (Fig. 8). As a result, T. alcocki can be considered a pseudocryptic species (i.e., minor morphological difference, only after other methods have unveiled their existence), which is not uncommon in marine organisms (Knowlton 1993(Knowlton , 2000, including brachyuran crabs (Ragionieri et al. 2009(Ragionieri et al. , 2012Shih et al. 2013a;Ng and Shih 2014;Lai et al. 2017).
According to Shih et al. (2016) and this study, the clade of T. urvillei and T. alcocki is sister to the clade composed of T. dussumieri, T. paradussumieri and T. capricornis. From the distributional patterns of these two main clades, while T. urvillei and T. alcocki are found only in the Indian Ocean, another main clade, composed of T. dussumieri, T. paradussumieri and T. capricornis , is primarily western Pacific, although T. alcocki and T. paradussumieri are sympatric in the eastern Indian Ocean Frith 1977a, 1978;present study). Because T. dussumieri has been recorded from Surin Islands, Phang Nga Province, Thailand Frith 1977a, b, 1978), T. alcocki is probably also sympatric with it.  Table 1 and Materials and methods. Hogarth (1986) has emphasized that the Red Sea population of "T. urvillei" is discontinuous with other populations and represents a significant extension of the known range. His Red Sea material is now recognized as T. alcocki (unpublished data; see Remarks under T. alcocki), and the northernmost distribution of T. urvillei sensu stricto is in southern Somalia (Crane 1975;Hogarth 1986). Tubuca alcocki thus has a wider range, which includes most of the northern Indian Ocean, from western Thailand (facing the Andaman Sea), through the Bay of Bengal and India, to the Red Sea. Tubuca urvillei, on the other hand, is known with certainty only from southeastern Africa. This distributional pattern is probably caused by the major oceanographic circulation systems at around 10°S, which limit the dispersal of larvae to the southeastern African coastline (Tsang et al. 2012).
Based on the pairwise divergence rates of 1.66 % per million years for COI of marine coastal crabs (Schubart et al. 1998), T. urvillei and T. alcocki diverged 2.2±0.4 million years ago (mya) (with uncorrected p-distance divergences of 3.65 % ± 0.71 %) around the beginning of the Pleistocene. The divergence between them is probably caused by the change of larval dispersal routes through ocean currents (e.g., the Equatorial Counter Current for a biogeographic barrier of barnacles; Tsang et al. 2012), which was likely influenced by the extreme climate during the glaciation periods in the Pleistocene (Shih et al. 2013a).