Delimitation of Tigertooth Croaker Otolithes Species (Teleostei: Sciaenidae) from the Western Arabian Gulf Using an Integrative Approach, with a Description of Otolithes arabicus sp. nov.

© 2019 Academia Sinica, Taiwan Open Access Delimitation of Tigertooth Croaker Otolithes Species (Teleostei: Sciaenidae) from the Western Arabian Gulf Using an Integrative Approach, with a Description of Otolithes arabicus sp. nov. Yu-Jia Lin1,*, Mohammad A. Qurban1, Kang Ning Shen2, and Ning Labbish Chao3 1Marine Studies Section, Center for Environment and Waters, King Fahd University of Petroleum and Minerals, Kingdom of Saudi Arabia; Marine Studies Section, Center for Environment and Waters, King Fahd University of Petroleum and Minerals, Kingdom of Saudi Arabia. *Correspondence: E-mail: [email protected] (Lin). E-mail: [email protected] (Qurban) 2Department of Environmental Biology and Fisheries Science, National Taiwan Ocean University, Taiwan. E-mail: [email protected] 3Sciaenidae Red List AuthorityIUCN, SSC. E-mail: [email protected] Received 18 September 2018 / Accepted 24 April 2019 / Published 24 June 2019 Communicated by Hin-Kiu Mok Two species, Otolithes ruber and Otolithes cuvieri, are currently recognized in the sciaenid genus Otolithes. Recent findings suggest that Otolithes ruber likely has multiple genetically and morphologically distinct lineages and one of them, Otolithes sp. West Indian Ocean II group (WIO II group), has been previously identified in the Arabian Gulf. In this study, the specimens of Otolithes sp. collected from the western Arabian Gulf were examined using an integrative approach by combining mitochondrial cytochrome c oxidase 1 gene, morphological characteristics, and otolith-shape analyses. Three groups were found to have small within-group and large between-group genetic distance: the Otolithes sp. Western Arabian Gulf (WA) group, and the Otolithes sp. WIO II groups type A and type B. Accordingly, three primary species hypotheses were proposed. Evidence from conventional morphological comparisons, multivariate statistical analysis, geometric morphometric landmark analysis on morphological characteristics, and otolith shape analysis based on wavelet transformation all favor the hypothesis that the Otolithes sp. WA group is a distinct lineage. For this new species, the name Otolithes arabicus sp. nov. is proposed. A detailed description of Otolithes arabicus sp. nov. and a key to identifing species in the genus Otolithes are also provided. However, the primary species hypotheses for Otolithes sp. West Indian Ocean II group type A and type B cannot be fully supported because of partial congruence, which may result from recent divergence. Key words: DNA barcode, Integrative taxonomy, Morphology, New Species, Otolithes, Otolith shape, Sciaenidae. Citation: Lin YJ, Qurban MA, Shen KN, Chao NL. 2019. Delimitation of tigertooth croaker Otolithes species (Teleostei: Sciaenidae) from the western Arabian Gulf using an integrative approach, with a description of Otolithes arabicus sp. nov. Zool Stud 58:10. doi:10.6620/ZS.2019.58-10. BACKGROUND Correct species delimitation is central to the exploration and proper measurement of biodiversity. Multiple data sources, such as morphology, genetics, biochemical characteristics, life history traits, ecological roles, and behavior patterns, have been applied to delimit species. However, none of these is perfect, and they each have their own limitations. Therefore, an integrative approach that takes advantage of complementary data sources has been suggested to solve taxonomic problems, avoid failures inherent to single disciplines, and increase rigor in species delimitation (e.g., Dayrat 2005; Schlick-Steiner et al. 2010; Pante et al. 2015). A number of studies have proposed protocols for determining taxonomy using an integrative approach. Zoological Studies 58: 10 (2019) doi:10.6620/ZS.2019.58-10 1 © 2019 Academia Sinica, Taiwan Generally, it involves first proposing several primary species hypotheses based on one data source, such as a single genetic marker, and then testing these primary species hypotheses in a hypothesis-test manner using other data sources. When all lines of evidence reach agreement or congruence in favor of some primary species hypotheses, these hypotheses are turned into secondary species hypotheses, a name is given, and a formal species description is conducted (Schlick-Steiner et al. 2010; Pante et al. 2015). The genus Otolithes is a group of marine teleost fishes from the family Sciaenidae, which are commonly called croaker or drum fish and commonly have large otoliths. The genus is characterized by having two exposed canines in each jaw, from which it derives the common name ‘tigertooth croaker’. According to Trewavas (1977), two species, Otolithes ruber and Otolithes cuvieri, are currently recognized in the genus. The former differs from the latter by having a shorter head depth, more gill raker numbers at the lower limb of the first gill arch, and more numerous swim-bladder appendage pairs. As indicated by recent findings based on genetic tools, there may be multiple lineages in the genus Otolithes, especially for the species Otolithes ruber. Zemlak et al. (2009) found deep divergences in the cytochrome c oxidase 1 gene (CO1) between Otolithes ruber populations from South African and Australian waters that exceeded the threshold value of species level divergence. Lo et al. (2017) further investigated the evolutionary relationships, species diversity, and taxonomic status of Sciaenidae within the Indo-West Pacific region using multiple genetic markers. In addition to confirming the validity of 32 currently recognized species, they further identified several operational taxonomic units within Otolithes ruber and concluded that the taxonomy of Otolithes ruber requires further investigation to identify potential cryptic and new species. The Arabian Gulf is one of several places where potential cryptic and new species in the genus Otolithes may occur. Lo et al. (2017) found that one group inside Otolithes ruber, identified as Otolithes sp. West Indian Ocean II group (WIO II) collected from Dubai in the southern Arabian Gulf, differed from other Otolithes ruber lineages in the Pacific based on genetic distance, strong support in phylogenetic analyses, and a distinct metristic trait (low number of dorsal fin spines). However, only 28 Otolithes individuals were examined in Lo et al. (2017) from their entire range, and it is not yet clear if Otolithes ruber from the Arabian Gulf is allopatric to other lineages in the Pacific. A comprehensive taxonomic examination with more specimens is therefore helpful to elucidate whether the Arabian Gulf lineages are allopatric. Having higher sample sizes also enables the construction of reliable estimates of ranges for conventional comparisons of metristic traits and morphological measurements, as well as enabling modern multivariate statistical and geometric morphometric analyses. In this study, all Otolithes specimens collected were called Otolithes sp. to be in accordance with the terminology in Lo et al. (2017). Then, the taxonomic status of Otolithes sp. collected from the western Arabian Gulf was examined using an integrative approach following methodological guidelines proposed by Dayrat (2005), Schlick-Steiner et al. (2010), and Pante et al. (2015). First, a discovery-driven approach was conducted to derive possible primary species hypotheses by analyzing the phylogenetic relationship of mitochondrial cytochrome c oxidase 1 gene (CO1) for Otolithes sp. individuals collected from the western Arabian Gulf, as well as Otolithes ruber from the Indo-Pacific region and Otolithes cuvieri from the Indian Ocean, over the known distribution range of the genus. Based on their phylogenetic relationships, several primary species hypotheses about possible species lineages were proposed. Then a hypothesisdriven approach was applied by testing those primary species hypotheses against specimens collected from the western Arabian Gulf, using other data sources: the threshold criteria of DNA barcoding suggested by Hubert and Hanner (2015), somatic morphological data, and otolith shape outline data. Four types of analyses were applied to the somatic morphological data: visual examination of the specimens, conventional comparison of meristic traits and morphometric measurements, multivariate statistical analysis from morphometric measurements, and geometric morphometric analysis on pre-determined landmarks. As an alternative source of information, otolith shape analysis is an emerging tool for identification of fish species (e.g., Christensen et al. 2018). Lastly, a primary species hypothesis supported by all lines of evidence was identified, which is considered as a secondary species hypothesis with congruence. Then, this group was described and an updated key was provided to identify species in the genus Otolithes in the Western Indo-Pacific region. MATERIALS AND METHODS Sample collection and measurement A total of 126 fresh samples of tigertooth croaker (Otolithes sp.) were collected from the Jubail Fish Market in the Eastern Province of Saudi Arabia from 29th November 2016 to 23rd February 2018 and preserved in a -18°C freezer. After defrosting, the page 2 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan total weight was measured to the nearest 0.1 g, 32 continuous morphological measurements were made to the nearest 0.01 mm by a digital caliper, and 16 meristic traits were counted (details shown in Appendix 1). A small piece of the left pectoral fin was clipped and preserved in absolute ethanol for DNA sequencing and a phylogeny was reconstructed. After measurements, fish were dissected, and the sex was determined by macroscopic examination of the gonads. Sagittal otoliths were removed, washed in freshwater, and stored dry. After dissection, each specimen was given a unique tag number (e.g., KFUPM-LKR001), fixed in 10% formalin, and then preserved in 70% ethanol. DNA extraction, PCR and sequencing Genomic DNA was extracted using the DNA purification kit of Bioman (Taipei), preserved in TrisEDTA buffer, and then quantified and diluted to 100 ng/μl prior to polymerase chain reaction (PCR). DNA extraction failed in a total of 3 specimens due to poor preservation conditions. A fragment of the mitochondrial CO1 gene was amplified with the universal primer sets FishF1/ FishR1 and FishF2/FishR2 (Ward et al. 2005). PCR was performed using the genomic DNA extract as a template with the following temperature program: heating at 94°C for 5 min, 35 cycles of denaturation at 94°C for 30 s, annealing at 53°C for 30 s, and extension at 72°C for 30 s. The PCR products were electrophoresed in a 2% agarose gel (Bioman) and stained with ethidium bromide (0.5 µg/ml) to characterize the size of bands with ultraviolet transillumination. The PCR products were sequenced directly using the primers used for PCR in an automated ABI Prism 377 sequencer (Applied Biosystems, Foster City, CA) at the Taiwan Genomics Company Sequencing Facility (Taipei). Analysis of DNA sequences In addition to specimens from the western Arabian Gulf, 48 sequences of individuals in the genus Otolithes were collected from the GenBank nucleotide database (https://www.ncbi.nlm.nih.gov/ genbank/) to infer the full phylogenetic relationship within this genus. Protonibea diacanthus was set as the outgroup species (Guo et al. 2017). Accession ID and collection sites of the collected sequences from GenBank are referred to in appendix 2. New sequences were blasted against available sequences from the GenBank nucleotide database. A 678 bp-long sequence alignment was performed using the package Clustal Omega (Sievers and Higgins, 2014). Although blind use of Kimura 2-parameter distance (K2P distance) in DNA barcoding studies has been criticized (Srivathsan and Meier 2012), in our preliminary examination, nearly identical results were obtained when using K2P distance and uncorrelated p-distance. Therefore, the results from K2P distance are reported in this study so as to be comparable to other studies. Different nucleotide substitution models were tested, and the model GTR + G was selected for having the smallest Akaike information criterion value. Then, a phylogenetic tree was reconstructed under the best-supported nucleotide substitution model using the partitioned Maximum Likelihood method (Felsenstein 1981). The reliability of the tree was tested by a bootstrap approach with 5000 iterations (Felsenstein 1985) with 80% agreement regarded as sufficient support. These analyses were completed using the packages ape (Paradis et al. 2004) and phangorn (Schliep 2011) in R (version 3.4.2, R Core Team, 2017). An empirical approach to species delimitation (Hubert and Hanner 2015) based on a threshold value of betweenand within-group genetic distance was applied across the following four cases: case 1, withingroup distance smaller and between-groups distance larger than the threshold: this group is likely a species achieved by reciprocal monophyly; case 2, both withinand between-groups distance larger than the threshold: these are likely composite species or several lineages; case 3, both withinand between-groups distance smaller than the threshold: the species or group has recently diverged from its sister species, and either ancestral polymorphism or introgressive hybridization has occurred; case 4, within-group distance larger and between-groups distance smaller than the threshold: specimens may have been misidentified or mislabelled, and a reassessment is required. Ward et al. (2005) found that the betweenspecies K2P distance of CO1 greater than 90% of 207 Australian fish species did not exceed 2%, which was proposed by Hubert and Hanner (2015) as the cutoff for species delineation. However, the fish from the Western Indo-Pacific region likely exhibited a higher degree of cryptic species diversity (Zemlak et al. 2009). Hubert et al. (2017) analyzed 3174 DNA barcodes from 805 fish species at three locations across the Western Indo-Pacific region and found that more than 80% of the mean intra-specific genetic differences were lower than 3%. Therefore, a more conservative threshold of 3% genetic distance was applied for the empirical approach to species delineation. Once classified into a potential taxonomic group, the average, standard deviation, and range of withinand between-group K2P distance of CO1 was calculated using the function meandist in the package vegan (Oksanen et al. 2018) in R. Several primary species hypotheses were proposed page 3 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan based on the analyses of CO1. These hypotheses were then tested against morphometric measurements, meristic traits, and otolith shape. Analysis of morphometric measurements and meristic traits Four approaches were applied to analyze morphometric and meristic trait data. First, specimens were visually examined for differences in general appearance. The second approach applied conventional taxonomic comparison of the range of morphometric measurements and meristic traits. Data from Trewavas (1977), Talwar (1995), and Lo et al. (2017) were used to construct the baseline range of morphometric and meristic traits for Otolithes ruber and Otolith cuvieri. Information on the phylogenetically distinct WIO II group (Lo et al. 2017) also was included. Then, the traits and measurements from different groups were compared with those in the baseline group. The third approach included a multivariate analysis of morphometric measurements (Claude 2008). A total of 30 morphometric measurements, except total and standard length, were standardized by a general linear model using standard length as the covariate to remove the interference effect of body size (Lleonart et al. 2000; Berner 2011). To test the primary species hypotheses, the constrained origination, canonical analysis of principal coordinates (CAP, Anderson and Willis 2003) on the Euclidean distances among the standardized measurements was conducted. The significance of the differences between groups over the constrained axes was tested by a canonical test with 4999 permutations. The data were plotted on the first two canonical axes to show the patterns of different groups. The scores of different variables on the canonical axes were used to show the contribution of different measurements to the separation of groups. The canonical analysis of principal coordinates and canonical test were done using the function capscale and anova in the package vegan (Oksanen et al. 2018) in R. The fourth approach to morphometric measurements used geometric morphometric landmark analysis based on the Procrustes paradigm. This approach utilizes the coordinates of pre-determined landmarks to record the relative positions of morphological points, boundary curves, and surfaces as the basis of shape quantification (Claude 2008; Adams and Otárola-Castillo 2013). A total of 17 landmarks were identified to represent the general body shape, based on D’Anatro and Lessa (2011) and Zhang et al. (2016). The specimens were oriented in a fixed position, and digital photos were taken under a white background with a scale bar. The coordinates of these 17 landmarks on the digital photos were read and digitized. Then, a generalized Procrustes analysis (Claude 2008) was applied for the superimposition and the minimization of shape variation (represented by the Procrustes distance) among specimens. The shape variation among groups was tested by the Procrustes regression with 4999 permutations. Body size was used as a covariate in the Procrustes regression to control the shape-size covariation. Changes in body shapes were visualized by plotting the mean locations of the landmarks with the vectors representing the direction and magnitude of the shape changes between groups. The geometric morphometric analysis was performed with the package geomorph (Adams and OtárolaCastillo 2013) in R. Analysis of otolith shape Otolith shape analysis was performed on 2D projections of the lateral face of sagittal otoliths. Otoliths were photographed on a black background with a reflected light source, using a dissection microscope with an Olympus DP72 (Olympus©) digital camera attached. Images were captured and digitalized with the software CellSens Standard, ver. 1.5 (Olympus©) including a scale bar of 2 mm. The closed otolith outline was automatically detected and captured, successive coordinate points were smoothed with 100 iterations, and otolith images were rotated so that they were all oriented along the longest axis. When all the outlines were captured, smoothed, and positioned, otoliths shape coefficients were extracted using wavelet transformation. Shape coefficients were standardized by fish size by removing allometric relationships with fish lengths (Lleonart et al. 2000; Berner 2011). A canonical analysis of principal coordinates on the wavelet coefficients of the otolith shape was applied using Euclidean distances. A canonical test with 4999 permutations was used to test for significance of differences under the different primary species hypotheses. Lastly, the mean shapes of otoliths among different groups were plotted to visualize the differences in shape. Tracing of otolith outlines and contour mapping for each group were performed with the package shapeR (Libungan and Pálsson 2015) in R. RESULTS Phylogenetic relationship of the genus Otolithes based on the mitochondrial CO1 gene Figure 1 shows the maximum likelihood phylogenic tree based on 102 CO1 sequences (678 bp) page 4 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan Fig. 1. (a) Full phylogenetic tree of the genus Otolithes with specimens from the Indo-Pacific, Western Arabian Gulf, and South Africa by Kimura2P distance and maximum-likelihood distance using the GTR + G nucleotide substitution model and Protonibea diacanthus as the outgroup; (b) subtree showing three clades: the Otolithes cuvieri complex, the Otolithes ruber Western Indo Pacific clade, and the Otolithes sp. Arabian Gulf/South Africa clade. The black numbers at nodes are bootstrap supporting values (%). Only values larger than 80% are shown. Grey numbers in italics on the branches represent branch length, with only those longer than 3% plotted. from the genus Otolithes across their range from South Africa to Southeast Asia. In this tree, three major clades with high bootstrap support values (BS) were identified: Otolithes ruber Indo-Pacific clade (BS = 100%); Otolithes sp. Arabian Gulf/South Africa clade (BS = 96%); and the Otolithes cuvieri complex (BS = 99%), which contained specimens identified as both Otolithes ruber and Otolithes cuvieri (Fig. 1a). Furthermore, three large groups within the Otolithes sp. Arabian Gulf/ South Africa clade with strong bootstrap support were identified: the Otolithes ruber South Africa clade (BS = 100%), the Otolithes sp. WIO II group (BS = 100%), and the Otolithes sp. West Arabian Gulf group (WA, BS = 100%, Fig. 1b). Inside the Otolithes sp. WIO II group there were two clear groups, one (BS = 100%) of which had nearly identical CO1 sequences to the sequence of Otolithes sp. WJC-2017 published in Lo et al. (2017). Therefore, their terminology, with reference to the West Indian Ocean Group II (WIO II group), was kept in this study. The two clusters in the Otolithes sp. WIO II group (BS = 100%) had substantial genetic differences and were separated into type A (WIO IIA) and type B (WIO IIB, Fig. 1c). According to the observed groupings based on CO1 sequences, a primary species hypothesis was proposed that Otolithes sp. collected in the western Arabian Gulf (the WA Otolithes sp. group) consists of Otolithes sp. WIO II group type A (WIO IIA) and type B (WIO IIB) and are distinct lineages with different characteristics. page 5 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan The primary species hypothesis about Otolithes sp. in the western Arabian Gulf was tested by examination of the mean withinand between-group Kimura-2-parameter (K2P) distances (Table 1) under a threshold value of 3%. The primary species hypothesis for the Otolithes sp. WA group, Otolithes sp. WIO II group type A and type B is supported by their small mean within-group K2P distances from 0.05 to 0.33% and large mean between-group distances from 3.01 to 17.63%. In addition, large withinand between-group genetic distances further suggest that the Otolithes ruber Indo-Pacific clade might have different lineages and that the Otolithes cuvieri complex is likely composed of multiple species. Comparison of morphology and analysis of morphometric measurements and meristic traits for specimens from western Arabian Gulf WIO IIA individuals were the most common of the specimens from the western Arabian Gulf (52%), followed by the WA group (28%) and WIO IIB individuals (20%, Table 2). The WA group had significantly smaller standard lengths and total weights than the WIO IIA group (generalized linear model, t122 = 3.571 and 2.827, p = 0.0005 and 0.0055, respectively) and WIO IIB group (t122 = 7.650 and 5.664, p = < 0.0001 and < 0.0001, respectively). By visual inspection of the general appearance of each group, the primary species hypothesis for Otolithes sp. WA group has the following Fig. 1. (c) one subtree in the Arabian Gulf/South Africa clade showing two clusters: the South Africa clade and the Western Arabian Gulf type (WA) clade; (d) subtree in the Arabian Gulf/South Africa clade showing one group that is identical to the West Indian Ocean II (WIO II) group from Lo et al. (2017), WIO II type A, and WIO II type B. The black numbers at nodes are bootstrap supporting values (%). Only values larger than 80% are shown. Grey numbers in italics on the branches represent branch length, with only those longer than 3% plotted. Samples collected from the Arabian Gulf are abbreviated with LKR. page 6 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan distinct morphological characteristics: a truncated caudal fin (that in the WIO IIA and WIO IIB groups is slightly rhomboid or rounded), two rows of spots on the second dorsal fin, deeper body depth, and larger eye (Fig. 2, lower panel). In contrast, the species hypothesis to distinguish the WIO IIA and WIO IIB groups is not supported due to their similar appearance (Fig. 2, upper and middle panel). Primary species hypotheses were tested by comparing morphometric measurements and meristic traits of Otolithes sp. individuals collected from the western Arabian Gulf, which differed from those of Otolithes ruber and Otolithes cuvieri from published literature (Tables 3, 4) in which non-overlapping ranges indicate support for the hypotheses. Compared to Otolithes ruber, individuals in the WA group had higher body depth (relative to standard length 23.5 to 29.8% compared to 21.0 to 26.2%) and significantly higher numbers of gill rakers on the lower part of the first gill arch (ANOVA on log-transformed data, t136 = 12.86, p < 0.0001). Compared to Otolithes cuvieri, the WA group individuals have more pairs of swim bladder appendages (32-35 pairs) and longer lower jaw bone length relative to head depth (106 to 110%). WIO IIA and WIO IIB differed from Otolithes ruber in having fewer spines on the spinous dorsal fin (mostly 8-9) and from Otolithes cuvieri in having more pairs of swim bladder appendages and longer lower jaw bone length (Tables 3, 4). When testing species hypotheses within the WA individuals, the WA group can be distinguished from the WIO II group (type A and type B) by having more gill rakers on the lower part of the first gill arch (11-13 vs. 7-10) and more spinous dorsal fin spines (10 vs. 8-9, Table 4). However, individuals from WIO IIA and WIO IIB have overlapped morphometric measurements and meristic traits that cannot be used to deliminate these two types (Tables 3, 4). Figure 3 summarizes the results of canonical Table 1. Mean and ranges (in parentheses) of genetic distances (Kimura 2 parameter distance, %) within groups (diagonal values in bold) and between groups (lower triangle values) for the Otolithes sp. western Arabian Gulf group (WA), West Indian Ocean Group II type A and B (WIO IIA and WIO IIB), Otolithes ruber South Africa clade (O. ruber SA), Otolithes ruber Indo-Pacific clade (O. ruber IP), Otolithes cuvieri complex (O. cuvieri), and the outgroup Protonibea diacanthus (OG) WA WIO IIAWIO IIBO. ruber SA O. ruber IP O. cuvieri OG WA 0.05 (0.00-0.38) WIO IIA12.28 (11.78-13.47) 0.33 (0.00-3.15) WIO IIB12.34 (11. 80-13.00) 3.01 (0.00-3.76) 0.28 (0-.97) O. ruber SA 12.27 (12.02-12.50) 7.10 (6.47-7.33) 6.61 (6.25-7.12) 0.00 (0.00-0.00) O. ruber IP 16.48 (15.77-17.87) 18.03 (16.20-19.36) 16.95 (15.95-18.31) 15.88 (15.16-17.21) 2.41 (0.00-5.02) O. cuvieri 17.63 (15.87-20.52) 16.60 (13.87-19.64) 15.54 (13.63-18.06) 16.89 (15.34-18.3) 14.56 (10.31-17.68) 4.25 (0.00-9.25) OG 18.96 (18.93-19.20) 21.54 (20.51-21.86) 20.56 (20.24-20.79) 20.49 (20.49-20.49) 17.66 (16.95-18.51) 18.79 (17.61-20.60) NA Table 2. Sample size (n) and mean (± SD) values of the standard length (SL) and total weight (TW) of two types in the West Indian Ocean Group II (WIO IIA and WIO IIB) and western Arabian Gulf (WA) groups of Otolithes sp. collected from the western Arabian Gulf Type WIO IIAWIO IIBWA Sex Female Male Female Male Female Male n 37 29 16 9 16 19 SL (mm) 302 ± 38 328 ± 51 392 ± 90 365 ± 57 289 ± 26 271 ± 17 TW (g) 517 ± 230 645 ± 284 1043 ± 318 801 ± 318 481 ± 156 369 ± 75 page 7 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan analysis of principal coordinates on 30 morphometric measurements standardized to standard length. The canonical tests indicate that measurements are significantly different among WIO IIA, WIO IIB and WA individuals (canonical test with 5000 permutations, permuted-F = 13.92, p < 2 × 10-4). Pairwise comparisons indicate that WA individuals are significantly different from WIO II individuals (type A and B combined, permuted-F = 35.87, p < 2 × 10-4). WIO IIA individuals also differ significantly from WIO IIB individuals (permuted-F = 3.92, p = 0.0002). WIO IIA individuals are mostly distributed over the central region and WIO IIB individuals over the region where values on the 1st and 2nd canonical axis are negative. WA individuals have a distinct distribution around the region with positive values on the 1st canonical axis (Fig. 3a). The contribution of the measurements on the CAP1 and CAP 2 axis (Fig. 3b) indicate that the body depth (BD), horizontal and vertical eye length (EhL and EvL), length of the spinous dorsal fin (D1L), and pectoral length (PecL) make the highest contribution to separating the WA and WIO II groups. WIO IIB individuals may have a longer length of the soft dorsal fin (D2L) and vent-to-anal fin-origin length (AVL). Figure 4a shows the locations of 17 landmarks for geometric morphometric analysis. The landmark coordinates differ significantly among these three groups (Procrustes regression, permuted-F = 10.06, p < 10-4, 4999 permutations), considering the shapesize covariation (permuted-F = 6.78, p = 2 × 10-4). Pair-wise comparisons indicate that the WA group has significantly different landmark coordinates than the WIO II group (types A and B combined, permuted-F = 5.64, p < 10-4), and the WIO IIA group differed significantly from WIO IIB (permuted-F = 3.49, p = 0.0043). WA group individuals differed from the WIO II group by having deeper body depth, larger heads, and longer eye horizontal length (Fig. 4b). WIO IIB individuals differed from WIO IIA individuals by having narrower and more slender body depth and smaller Fig. 2. Fresh specimen photos for West Indian Ocean II Group type A (WIO IIA) and West Indian Ocean II Group type B (WIO IIB), and western Arabian type (WA), and the posterior part of the soft dorsal fin. Scale bar = 5 cm. page 8 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan heads (Fig. 4c). Comparison of otolith outlines for specimens from western Arabian Gulf Left sagittal otoliths, sulcus side down, from WIO IIA, WIO IIB, and WA groups are shown in figure 5a. The summary of the canonical analysis of principal coordinates using the wavelet coefficients from the otolith outline is in figure 5b. The differences in the mean otolith shapes among these three groups are visualized in figure 5c. Otolith shapes differ significantly among the three groups (canonical test, permuted-F = 26.53, p < 2 × 10-4, 4999 permutations). Pair-wise comparisons indicate that members of WA have a significantly different otolith shape than the WIO II group (type A and B combined, permuted-F = 48.75, p < 2 × 10-4). Within the WIO II group, the otolith shape also differs significantly between type A and B (permuted-F = 2.89, p = 0.0106). Specifically, WIO IIA individuals are distributed around the central region; WIO IIB individuals have negative values in the 1st and 2nd canonical axes and the WA individuals usually have positive values in the 1st canonical axis (Fig. 5b). WA individuals have rounder otolith lateral faces and are relatively shorter in the anterior-posterior axis, while WIO IIB individuals have more pointed and longer otoliths in the anterior axis. DISCUSSION Between-group comparisons are made with a good total sample size (N = 126 specimens); intraand inter-group variation is addressed, and two species currently recognized in the genus Otolithes are included. Moreover, all Otolithes sp. specimens, i.e., whole specimens, fin clips, and otoliths, are preserved and deposited in multiple museums, which enables further phylogenetic and morphological re-examination (Dayrat 2005). Large sample size coverage in space and time is desirable to cover possible spatial and Table 3. Comparison table of counts and ratios of body measurements for species in the genus Otolithes: Otolithes ruber, Otolithes cuvieri, West Indian Ocean group II (WIO II) from Lo et al. (2017), and two types in WIO II (WIO IIA and WIO IIB), and the western Arabian Gulf group (WA). SL: standard length, HL: head length, HD head depth (the distance between the sphenotic ridge and the angle of lower jaw); SP and SO dorsal fin indicates spinous and soft dorsal fin, respectively. References: R1: Trewavas 1977; R2: Talwar 1995; R3: Lo et al. 2017 Species Otolithes ruber Otolithes cuvieri WIO II WIO IIAWIO IIBWA Counts R1 R2 R3 R1 R2 R3 This study SP dorsal fin spine number 10 9-10 10 10 9 8-10 9 10 SO dorsal fin spine number 1 1 1 1 1 1 SO dorsal fin soft ray number 27-30 27-30 27-29 29-32 29-32 30 28-32 27-30 28-31 Anal fin spine number 2 2 2 2 2 Anal fin soft ray number 7 7-8 7 7 7 Pectoral fin ray number 16 16 16 16 16 Pelvic fin spine number 1 1 1 1 1 Pelvic fin soft ray number 5 5 5 5 5 1st gill arch gill raker number, upper part 4 3-4 6 5 3-5 3-4 4-6 1st gill arch gill raker number, lower part 8-11 8 - 11 9-10 12-17 11-16 12 7-10 8-10 11-14 Swim bladder appendages pairs 30-37 32-36 36-40 25-28 25-28 35 32-42 33-40 32-38 Measurement ratios Body depth (% SL) 21.0 - 26.2 21.0 - 25.6 25.4-30.0 26.2-30.0 23.0-27.8 21.3-26.3 23.5-29.8 Head length (% SL) 27.6 - 32.5 29.0 - 34.0 30.0-34.3 30.0-33.5 27.3-32.6 27.0-31.7 28.1-33.3 Pectoral fin length (% SL) 21.0 - 22.0 21.0-22.2 16.5-21.1 16.1-19.0 17.8-23.5 Eye horizontal diameter (% HL) 13.5 - 23.1 15.4 - 23.5 22.0-26.0 23.0-26.0 12.3-19.0 12.4-18.3 14.8-20.0 Eye vertical diameter (% HL) 9.9-15.1 9.8-14.6 13.0-17.6 Snout length (% HL) 23.0 - 29.0 23.0 - 27.6 23.0-27.0 23.0-26.5 20.5-26.9 20.7-28.6 20.8-25.2 Interorbital length (% HL) 21.6 - 25.0 20.5 - 24.3 24.0-28.4 24.5-28.4 21.8-27.9 20.0-25.8 22.5-29.7 Lower jaw length (% HD) 100 - 124 95-100 103-133 110-122 105-121 Comparison Snout length > eye horizontal length Yes Yes No Yes Yes Yes Canines Large Moderate Large Large Large page 9 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan apparent or disappear. Distribution: Presently this new species is known to waters off the western Arabian Gulf to the Gulf of Oman. Otolithes arabicus had a rare occurrence inside the Arabian Gulf, and the author (YJL) has seen it infrequently during regular visits to the Jubail Fish Market. Only a total of 6 out of 112 Otolithes sp. specimens were recorded from landings in the Western Arabian Gulf. However, a box imported from Oman to Saudi Arabia containing 37 Otolithes arabicus sp. nov. individuals was collected on the 1st of February, 2018, suggesting that this species might mainly occupy the Gulf of Oman, and enter the Arabian Gulf only occasionally. Etymology: This species is named based on the locality of discovery, the Arabian Gulf and the Gulf of Oman region. Arabian Tigertooth Croaker is proposed as the common name of this new species. Key to the species of Otolithes (see also Tables 3, 4, Figs. 2, 5 to 7) 1a. Caudal fin truncated, two to three black spots on the fin membranes between the soft rays of the soft dorsal fin from the base to the middle of the ray ............. Otolithes arabicus sp. nov. 1b. Caudal fin rhomboid or rounded, part between the soft rays of the soft dorsal fin dusky without dots .............................................. 2 2a. 8 to 9 spines on the spinous dorsal fin ........................................... ..................................... Otolithes sp. West Indian Ocean Group II 2b. 10 spines on the spinous dorsal fin ............................................. 3 3a. Gill rakers on the lower part of the 1st gill arch 8 to 11, the distance between the sphenotic ridge and the angle of the lower jaw less the lower jaw bone length, body depth 21.0 to 26.2% of standard length, 30 to 40 pairs of swim bladder appendages, canines very strong ............................................... Otolithes ruber 3b. Gill rakers on the lower part of the 1st gill arch mostly 12 to 17, the distance between the sphenotic ridge and the angle of the lower jaw as long or longer than the lower jaw, body depth 25.4 to 30.0% of standard length, 25 to 28 pairs of swim bladder appendages, canines moderate .......................... Otolithes cuvieri CONCLUSIONS The Otolithes sp. specimens collected from western Arabian Gulf are consisted of three genetically distinct lineages, and corresponding primary species hypotheses are proposed. The primary species hypothesis of one lineage, Otolithes sp. WA group, is supported by all lines of evidence from conventional comparison of meristic and morphological traits, landmark analysis and otolith shape analysis. Therefore, the name Otolithes arabicus sp. nov. is proposed for this new species. However, the evidences for the other two primary species hypotheses reach only partial congruence, leading to inconclusive identification results. Therefore, these two species hypotheses regarding cannot be fully supported at present and more information are required. Acknowledgments: This work and the new species name have been registered with ZooBank under urn:lsid:zoobank.org:pub:5C22431A-306F-4798-989C1A08D5D2064B. The authors are grateful for the team of Marine Studies Section, Center for Environment and Waters, King Fahd University, for proving logistic and technical help, Dr. Todd Ryan Clardy and Brian Jessop for reviewing the manuscript, and the fishermen at the Jubail Fish Market, Eastern Province, Saudi Arabia, for kindly providing information about the geological origin of the fish. Authors’ contributions: Yu-Jia Lin: Collection of specimens from the field, laboratory work, data analysis, and composing manuscript; Mohammad A. Qurban: Composing manuscript; Kang Ning Shen: DNA laboratory work, composing manuscript; Ning Labbish Chao: Composing manuscript and reviewing taxonomical description. All authors participated in revising the manuscript. All authors read and approved the final manuscript. Competing interests: YJL, MAQ, KNS, and NLC declare that they have no conflict of interest. Availability of data and materials: The summary of the meristic and morphological measurement data has already provided in table 2 and 3. The CO1 sequences of examined specimens will be uploaded to GenBank. The whole specimens preserved in formalin has been stored separately in the museum of Marine Studies Section, Center for Environment and Water, King Fahd University of Petroleum and Minerals, Saudi Arabia, as well as in National Museum of Marine Biology and Aquarium, Taiwan. Consent for publication: As the corresponding author, I state that all authors agree to its submission and the Corresponding author has been authorized by the co-authors. Ethics approval consent to participate: This study is done using dead specimens collected from the fish landing centers. No human subjects or laboratory animals are involved. REFERENCES Adams DC, Otárola-Castillo E. 2013. geomorph: an R package for the collection and analysis of geometric morphometric shape data. page 16 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan Methods Ecol Evol 4:393–399. doi:10.1111/2041-210X.12035. Anderson MJ, Willis TJ. 2003. Canonical analysis of principal coordinates: a useful method of constrained ordination for ecology. Ecology 84:511–525. doi:10.1890/0012-9658 (2003)084[0511:CAOPCA]2.0.CO;2. Berner D. 2011. Size correction in biology: how reliable are approaches based on (common) principal component analysis? Oecologia 166:961–971. doi:10.1007/s00442-011-1934-z. Christensen HT, Rigét F, Backe MB, Saha A, Johansen T, Hedeholm RB. 2018. Comparison of three methods for identification of redfish (Sebastes mentella and S. norvegicus) from the Greenland east coast. Fish Res 201:11–17. doi:10.1016/ j.fishres.2018.01.003. Claude J. 2008. Morphometrics with R. Springer Science & Business Media. D’Anatro A, Lessa EP. 2011. Phenotypic and genetic variation in the white croaker Micropogonias furnieri Desmarest 1823 (Perciformes: Sciaenidae): testing the relative roles of genetic drift and natural selection on population divergence. J Zool 285:139–149. doi:10.1111/j.1469-7998.2011.00823.x. Dayrat B. 2005. Towards integrative taxonomy. Biol J Linn Soc Lond 85:407–415. doi:10.1111/j.1095-8312.2005.00503.x. Felsenstein J. 1981. Evolutionary trees from DNA sequences: a maximum likelihood approach. J Mol Evol 17:368–376. Felsenstein J. 1985. Confidence limits on phylogenies. An approach using the bootstrap. Evolution 39:783–791. doi:10.1111/j.15585646.1985.tb00420.x. Guo CC, Liu M, Lin JJ, Dai FQ. 2017. Complete mitochondrial genome and the phylogenetic position of the tigertooth croaker Otolithes ruber (Perciformes: Sciaenidae). Mitochondrl DNA Part B 2:132–133. doi:10.1080/23802359.2016.1247676. Hubert N, Hanner R. 2015. DNA barcoding, species delineation and taxonomy: a historical perspective. DNA barcodes 3:44–58. doi:10.1515/dna-2015-0006. Hubert N, Dettai A, Pruvost P, Cruaud C, Kulbicki M, Myers RF, Borsa P. 2017. Geography and life history traits account for the accumulation of cryptic diversity among Indo-West Pacific coral reef fishes. Mar Ecol Prog Ser 583:179–193. doi:10.3354/ meps12316. Iwatsuki Y, Jawad LA, Al-Mamry JM. 2012. Johnius (Johnius) majan sp. nov., a sciaenid fish (Pisces: Sciaenidae) from Oman, Indian Ocean. Ichthyol Res 59:151–155. doi:10.1007/s10228-0110265-8. Libungan LA, Pálsson S. 2015. ShapeR: an R package to study otolith shape variation among fish populations. PLoS ONE 10:e0121102. doi:10.1371/journal.pone.0121102. Lin YJ, Roa-Ureta RH, Rabaoui L, Grandcourt EM, Maneja RH, AlAbdulkader K, Qurban MA. 2019. Association to vegetated habitats and different vulnerability to habitat degradation for two fish species, Epinephelus areolatus (Serranidae) and Siganus canaliculatus (Siganidae), from the western Arabian Gulf. Mar Poll Bull 141:482–492. doi:10.1016/j.marpolbul.2019.03.011. Lleonart J, Salat J, Torres GJ. 2000. Removing allometric effects of body size in morphological analysis. J Theor Biol 205:85–93. doi:10.1006/jtbi.2000.2043. Lo PC, Liu SH, Nor SAM, Chen WJ. 2017. Molecular exploration of hidden diversity in the Indo-West Pacific sciaenid clade. PLoS ONE 12:p.e0176623. doi:10.1371/journal.pone.0176623. Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, O’Hara RB, Simpson GL, Solymos P, Stevens MHH, Szoecs E, Wagner H. 2018. vegan: Community Ecology Package. R package. version 2.4-6. https://CRAN.R-project.org/ package=vegan. Pante E, Puillandre N, Viricel A, Arnaud-Haond S, Aurelle D, Castelin M, Chenuil A, Destombe C, Forcioli D, Valero M, Viard F, Samadi S. 2015. Species are hypotheses: avoid connectivity assessments based on pillars of sand. Mol Ecol 24:525–544. doi:10.1111/mec.13048. Paradis E, Claude J, Strimmer K. 2004. APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20:289–290. doi:10.1093/bioinformatics/btg412. R Core Team. 2017. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/. Schlick-Steiner BC, Steiner FM, Seifert B, Stauffer C, Christian E, Crozier RH. 2010. Integrative taxonomy: a multisource approach to exploring biodiversity. Annu Rev Entomol 55:421–438. doi:10.1146/annurev-ento-112408-085432. Schliep KP. 2011. Phangorn: phylogenetic analysis in R. Bioinformatics 27:592–593. doi:10.1093/bioinformatics/btq706. Sievers F, Higgins DG. 2014. Clustal Omega. Curr Protoc Bioinformatics 48:1.25.1-1.25.33. doi:10.1002/0471250953. bi0313s48. Song YS, Kim JK, Kang JH, Kim SY. 2017. Two new species of the genus Atractoscion, and resurrection of the species Atractoscion atelodus (Günther 1867) (Perciformes: Sciaenidae). Zootaxa 4306:223–237. doi:10.11646/zootaxa.4306.2.3. Srivathsan A, Meier R. 2012. On the inappropriate use of Kimura-2parameter (K2P) divergences in the DNA-barcoding literature. Cladistics 28:190–194. doi:10.1111/j.1096-0031.2011.00370.x. Talwar PK. 1995. Fauna of India and the adjacent countries: Pisces. Perciformes, Scienidae. Zoological Survey of India. Trewavas E. 1977. The sciaenid fishes (croakers or drums) of the Indo-West-Pacific. Trans Zool Soc Lond 33:253–541. Ward RD, Zemlak TS, Innes BH, Last PR, Hebert PD. 2005. DNA barcoding Australia’s fish species. Philos Trans R Soc Lond B Biol Sci 360:1847–1857. doi:10.1098/rstb.2005.1716. Zemlak TS, Ward RD, Connell AD, Holmes BH, Hebert PD. 2009. DNA barcoding reveals overlooked marine fishes. Mol Ecol Resour 9:237–242. doi:10.1111/j.1755-0998.2009.02649.x. Zhang C, Jiang YQ, Ye ZJ, Li ZG, Dou SZ. 2016. A morphometric investigation of the small yellow croaker (Larimichthys polyactis Bleeker, 1877): evidence for subpopulations on the Chinese coast. J Appl Ichthyol 32:67–74. doi:10.1111/jai.12923. page 17 of 18Zoological Studies 58: 10 (2019) © 2019 Academia Sinica, Taiwan Supplementary Materials Appendix 1. Descriptions and abbreviations of the morphology measurements and counts. (download) Appendix 2. Name of the sequences used in this study, accession ID and sampling site from GenBank. (download) Appendix 3. A picture showing the relative size and position of the teeth on both jaws for Otolithes arabicus sp. nov. (download) page 18 of 18Zoological Studies 58: 10 (2019)