Both mitochondrial DNA and genome-wide SNP data analysis support a species complex in the brown-banded bamboo shark. By sampling comprehensively across a part of its range, we have revealed greater genetic variation and potentially more species than suggested by a previous study (18) that included fewer sampling locations. Obtaining representative samples across a species’ distributional area can be challenging for elasmobranchs. There are potential difficulties in obtaining a comprehensive sample set for a species, due to; rarity in nature or in markets; difficult to catch; financial cost; large specimen sizes; wide geographical spread and species protections, which limits access to samples or requires special permits (32). Still, many new species of elasmobranch have been described based on limited geographic sampling over the last decade (e.g. (33–35)).
Our analyses on the genetic data of Chiloscyllium punctatum support the hypothesis of Naylor et al. (18) that a potential cryptic species occurs in Australia. However, by including a more comprehensive suite of sampling sites that captured more genetic variation, we were able to identify four OTUs within C. punctatum. Yet, as the OTUs are allopatric, it is difficult to discern the population-species boundary on genetic data alone. Depending on the nature of the supporting information, C. punctatum species complex may consist of three or four putative species. Based on the genetic data alone, the four OTUs can be interpreted as four distinct species supported by both phylogenetic and species delineation (BFD*) analyses.
The type locality for the original description of C. punctatum was described from a geographic location associated with OTU1 in the waters off Jakarta, Indonesia (36). Specimens from Jakarta were clearly grouped in OTU1, which comprises individuals that inhabit shallow waters (20–50 m) within the Indo-Malay Archipelago (22, 23). Individuals from OTU1 are not only genetically separated to those in the other OTUs, but the distinction is also supported by biological, ecological and geographical data. Individuals from OTU1 have a relatively smaller body size with the maximum TL at around 1 m, compared with other OTUs that have a maximum TL more than 130 cm (see Table 2). That Indo-Malay individuals mature at a smaller size than do those from the Australasia further supports species delimitation (24, 27, 30,31). However, no information on size at maturity is available for individuals from the Indian Ocean region (OTU2 and OTU3) and given this biological feature is likely important for species delimitation in this group, further studies are needed. Ecologically, we observed that C. punctatum from the Indo-Malay region is commonly found on soft-bottom habitats (e.g. sand flat, muddy substrate). Geographically, all locations that were linked by shallow waters within the Sunda Shelf region (from Phuket, Thailand to East Kalimantan, Indonesia) showed strong connectivity except for South Sulawesi, which is geographically the furthest east location included in OTU1. South Sulawesi is also separated from the Indo-Malay region by a deep trench that could restrict gene flow.
Individuals from the Indian Ocean and Australian regions may be grouped into a putative species based on similarities of their biological characters, such as the larger maximum body size than for the Indo-Malay region, or may be separated into two distinct species based on ecological and geographical perspectives. However, the BFD* analysis lent support to two separate groups. Therefore, individuals from Australian (OTU4) and the Indian Ocean regions (OTU2, OTU3) can be designated as an incipient species if there are any supportive morphological characters that can distinguish them.
In the Indian Ocean region, individuals from the west coast of Sumatra (OTU2) were clearly separated from Lombok (OTU3) based on their fixed allele differences. However, this may reflect population level difference, given continuity of habitat between their geographic locations in the outer margin of the Sunda Islands region. Further, the mtDNA phylogram grouped both locations in the same clade together with those from Muncar in the south of the Bali Strait (see Fig. 2a). The high fixed allele differences, FST values and marked separation in the phylogenetic tree of specimens between Lombok and other locations may be caused by the unique geographical position of this location in the Lesser Sunda Island region. Lombok is separated from other locations by deep water (The Bali Sea and the Flores Sea in the north, and the Indian Ocean in the south, each with depths of ca. 1500–3000 m) and the strong Indonesian Throughflow current in the west (the Lombok Strait). Together, these likely represent significant barriers to gene flow between Lombok and both the Indo-Malay and the Australian regions. The geographical barrier in the Lombok Strait is an important feature for speciation in particular groups such as the separation of two blue-spotted maskray species, Neotrygon caeruleopunctata in the west and N. australie in the east (37, 38). In contrast, this barrier is just resulting in population structure for other species such as the Indonesian wobbegong, Orectolobus leptolineatus (33) and Indonesian speckled catshark, Halaelurus maculosus (39) that occur along the eastern Indian Ocean. Additional sampling at locations along the south coast of Java and the west coast of Sumatra is required to bridge the geographic gap that exists in our data between Sumatra and Lombok to further resolve whether the separation demonstrated here is representative of speciation or population differences. Nevertheless, in terms of habitat preference, C. punctatum from along the west coast of Sumatra, south of Java to Nusa Tenggara in eastern Indonesia (including Lombok), inhabit similar coral and rocky reef habitats.
The bamboo shark species from the Australian region (OTU4) can be found in various habitats, including coral reefs, seagrass beds, mangroves and estuaries (24). The tectonic plate boundary between the Sahul Shelf (South PNG, Western Australia, and South East Queensland) and the Sunda Shelf (Greater and Lesser Sunda Islands) is considered a major barrier causing a genetic break that separates populations and potential species within this complex. Examples of the influence of this barrier on species level differences can be found in blackspot sharks and banded eagle rays (40, 41). Even though species in the Australian region formed a distinct clade, there are some fixed differences between specimens from South East Queensland and those from South PNG and Western Australia. Nevertheless, there is no physical barrier, such as deep water or strong currents that might be a driver of genetic segregation between these locations. Therefore, we suggest that locations in OTU4 represent a single species with population structure, and the differences within the clade may reflect a variation by distance effect (42). This should be further tested by intermittent sampling locations in northern Australian waters.
Our study also revealed that geographic distance does not directly correspond to genetic distance. For instance, Aceh (OTU2) is separated from Phuket (OTU1) by only about 470 km while some sites within the OTU1 cluster (such as between Phuket and West Java) are separated by more than three times the distance. Similarly, Muncar and East Java are separated by only 300 km and connected by the narrow and shallow waters of the Bali Strait, yet showed strong genetic separation. A possible reason for that separation is habitat type. In addition to the role that deep water and tectonic plate barriers play, habitat preferences are likely to be important in either structuring populations or delimiting species (43, 44). Based on habitat preferences, the four OTUs can be further delimited into three groups with OTU2 and OTU3 combined. The BFD* analysis also lent high support to this hypothesis.
The decision to delineate species complexes may vary among taxa. In some taxa, genetic differences may not be reciprocal with either morphology (phenotypic plasticity), biology (reproductive traits), or ecological characteristics (habitat preferences) (45). In elasmobranchs, genetics has been used to can delineate cryptic species in a species complexes in the genera Carcharhinus, Aetobatus and Neotrygon (38, 40, 46, 47), as well as coalesce two genera into one genus due to genetic indistinctiveness, such as in Mobula (48).
For some taxonomic groups such as birds and some mammals, the term 'subspecies' is used to define a population or group of populations that are distinctive yet insufficiently different to constitute a separate species based on subtleties in appearance and/or in genetic makeup (49–51). The term 'subspecies' is also used to differentiate a species complex based on ecological speciation for populations without multigene discontinuity (52). Moreover, 'subspecies' is applied to geographically isolated populations driven by biodiversity and conservation purposes, such as in some freshwater fishes (50). However, the use of subspecies in some taxa is not preferable due to confusion with the population term (53, 54), and has been rarely used in the past few decades for marine fishes (55). For elasmobranchs, this term has only been applied to few taxa, e.g. in catsharks (56–59), dogfishes (60), smooth-hounded sharks (61), hammerheads (62), eagle rays (63), and skates (64). Some of those subspecies remain valid such as the smooth-hounded shark (Mustelus canis canis and M. c. insularis) and for several species of skates (from Genus Raja and Leucoraja), while others were considered junior synonyms or have been elevated into distinct species (65). Therefore, the use of subspecies for the bamboo shark OTUs is plausible if they cannot be definitively classified at either the species or population level.
Conservation implications of the species delimitation
Splitting cryptic species in a complex into one or more distinct species may provide advantages for the species, not just for formal scientific recognition, but also to assess conservation risk (3, 4, 6, 66). Thus, from a conservation perspective, separating C. punctatum into two or three species is desirable as differences in biological features, spatial distribution, habitat occupancy, and type of fishery that operates in an area could influence how each species should be managed for sustainability. For instance, due to the intensive trawl operations in the Indo-Malay region (26, 67), the species based on OTU1 would be subjected to higher fishing pressure than individuals within OTU2 and OTU3 where bottom longline fisheries operate due to unsuitable substrates for bottom trawling, and compared to OTU4 where they are neither targeted nor caught as bycatch. This situation may lead to differences in threat profiles that necessitate revision among the OTUs of their conservation status.
In terms of fisheries management, each OTU with distinct fisheries characteristics can be treated separately. For instance, limits on minimum size or permitted fishing gear may differ between OTUs due to the nature of the fishery. For countries that implement ecosystem or species-based conservation management such as in Indonesia (referring to Regulation of the Minister of Forestry P.57/Menhut-II/2008 and Regulation of the Minister of Marine Affairs and Fisheries 3/2010), regulating two different options of fisheries management for one species is challenging, compared with countries that implement conservation strategies on a population-basis. Therefore, splitting C. punctatum into at least two or three different species for management purposes is appropriate for countries that apply species-based management, as long as diagnostic morphological characters are available. Especially when they are marketed in the same place with a lack of traceability, as occurs in Indonesia.
In contrast, splitting a cryptic species complex into several species based purely on genetics can cause problems if there are no strong supportive diagnostic characters to differentiate them in the field. An example is in the Neotrygon kuhlii species complex (38, 68) where several species are sympatric yet cannot be diagnosed morphologically, which complicates management. Policymakers may find it difficult to implement conservation and management actions, especially if the species are sympatric, inhabit similar ecotypes, or occur in one fishing region. Without the ability to differentiate among the members of a complex, studies on their biology, ecology, or population stock for management purposes are rendered problematic due to the likelihood of misidentification and possibly overlapping information (37). This is particularly challenging in countries or regions where access to genetic analysis is still limited and costly (69–71).