https://orcid.org/0000-0001-5552-4301
Figures
Abstract
Otoliths (ear stones) of the inner ears of teleost fishes, which develop independently from the skeleton and are functionally associated with hearing and the sense of equilibrium, have significantly contributed to contemporary understanding of teleost fish systematics and evolutionary diversity. The sagittal otolith is of particular interest, since it often possesses distinctive morphological features that differ significantly among species, and have been shown to be species- and genus-specific, making it an informative taxonomic tool for ichthyologists. The otolith morphology of the Caspian Sea gobiids has not been thoroughly studied yet, with data available for only a few species. The aim of the present paper is to examine the qualitative and quantitative taxonomic and phylogenetic information in the sagittal otoliths of these species. A total of 118 otoliths representing 30 gobiid species (including 53.5% of the Caspian gobiofauna) in three gobiid lineages (i.e., Gobius, Pomatoschistus, and Acanthogobius) and 11 genera (i.e., all Ponto-Caspian gobiid genera except Babka) were analysed at taxonomic levels using an integrated descriptive and morphometric approach. The results indicated high taxonomic efficiency of otolith morphology and morphometry at taxonomic levels for the Ponto-Caspian gobiids. Our qualitative and quantitative otolith data also (i) support the monophyly of neogobiin gobies, (ii) along with other morphological and ecological data, offer a new perspective on the systematics of Neogobius bathybius, (iii) suggest the reassignment of Hyrcanogobius bergi to the genus Knipowitschia, and (iv) question the phylogenetic integrity of the four phenotypic groups previously defined in the tadpole-goby genus Benthophilus; however, more studies are needed to complete these evaluations and confirm our otolith study findings.
Citation: Zarei F, Esmaeili HR, Stepien CA, Kovačić M, Abbasi K (2023) Otoliths of Caspian gobies (Teleostei: Gobiidae): Morphological diversity and phylogenetic implications. PLoS ONE 18(5): e0285857. https://doi.org/10.1371/journal.pone.0285857
Editor: Michael Schubert, Laboratoire de Biologie du Développement de Villefranche-sur-Mer, FRANCE
Received: November 15, 2022; Accepted: May 2, 2023; Published: May 15, 2023
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the manuscript and its Supporting information files.
Funding: Data collection (fish sampling) in Iran was supported by Shiraz University (SU-9630190). Shiraz University had no role in study design, data analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The otoliths of the inner ears in teleost fishes are arranged in three pairs termed the saccular (sagittae, the largest otoliths in most teleosts), utricular (lapilli), and lagenar (asterisci) otoliths. They are aragonitic mineralizations that develop independently from the skeleton, which are functionally associated with the senses of hearing and equilibrium [1]. Analysis of the shapes of otoliths has significantly contributed to knowledge of teleost systematics and biodiversity [2–7], as well as historic diversity, phylogeny, zoogeography, and climatology [8–12], ancient and modern fisheries [13], life history and habitat [14], and population structure [6, 7, 15]. The sagittal otolith (referred to as otolith hereforth) is of particular interest, since it often possesses distinctive morphological features that vary among populations, species, and genera [10, 16–18]. Some studies have concluded that otolith shape is primarily defined by genetic factors [19–21], however, others suggest that while genetics constrain the overall otolith’s shape, environmental conditions often alter its somatic growth rates, which can affect the otolith’s shape and result in intraspecific variation across the species’ distribution [22–26]. The extent to which otolith shape variations are genetically or environmentally induced are controversial and may differ among taxa.
The Ponto-Caspian basin, comprising the Black Sea, Sea of Azov, and Caspian Sea basins, has been the evolutionary stage for two main lineages of Gobiidae sensu Gill & Mooi [27]: (i) the endemic Ponto-Caspian benthophiline gobiids belonging to the Gobius lineage [28], classified in two main groups, nine genera, and three tribes: (1) the neogobiin group divided in the tribes Neogobiini (containing the genus Neogobius Iljin, 1927) and Ponticolini (four genera: Ponticola Iljin, 1927, Mesogobius Bleeker, 1874, Proterorhinus Smitt, 1899, and Babka Iljin, 1927), and (2) the benthophilin or tadpole gobies group comprising the tribe Benthophilini (four genera: Anatirostrum Iljin, 1930, Benthophilus Eichwald, 1831, Benthophiloides Beling & Iljin, 1927, and Caspiosoma Iljin, 1927. These tribes follow recommendations by Neilson and Stepien [29] and Agorreta et al. [28]. (ii) A branch of the Pomatoschistus lineage [28] or sand gobies, represented in the Ponto-Caspian basin by several species of Knipowitschia Iljin, 1927 and the monotypic genus Hyrcanogobius Iljin, 1928 [30].
Gobiid fauna of the Caspian Sea basin now include 43 species in 12 genera. Of these, 35 species are endemic to the Caspian Sea basin, seven are native to the Ponto-Caspian overall, and one species is exotic. The North, Middle and South Caspian Sea sub-basins harbor 21, 31, and 38 species, respectively, of which zero, three, and 10 are endemic to each [31]. Due to their low economic importance, small sizes, and distributions in difficult-to-sample deeper-water habitats [32–34], little is known about their biological characteristics and ecology. However, Gobiidae comprise 1/4 of species native to the Caspian Sea [31] and insights into their diversification and adaptations may help elucidate ecological and biogeographic understanding about the patterns and processes of species flock formation in the Ponto-Caspian realm.
Since the mid-2000s, Caspian Sea gobiid studies have examined the molecular phylogeny and biogeography of a limited number of species [7, 29, 30, 35], with several new species described in recent years [5, 36–39]. On the other hand, detailed discussions of their morphological characters have become rare, and these molecular-based studies suggest phylogenetic relationships that have not been considered with morphological information and therewith inspire new comparative studies to test those relationships. Hitherto, otolith morphology has been studied for just a few species [5–7, 38, 40, 45]. This study, using a taxonomically broad sample, aims to compare the otolith morphology of extant Caspian gobiids at different taxonomic levels based on detailed descriptions and morphometric analysis to reveal the taxonomic and phylogenetic information inscribed in their otoliths. The results not only provide another step towards the total-evidence phylogeny of the extant Pono-Caspian gobiids, but will also be beneficial for all future studies dealing with the taxonomic and phylogenetic assignment of fossil otolith records and otolith-based gobiid taxa described from the Ponto-Caspian basin and thus will help to improve our understanding of the pattern and process of gobiid diversification in the region.
Materials and methods
Sampling, otolith dissection, and SEM imaging
Necessary permits for sampling and observational field studies in the Southern Caspian Sea sub-basin have been obtained by the authors from the Iranian Department of Environment and the Iranian Fisheries Science Research Institute. New gobiid specimens were collected from the Southern Caspian Sea sub-basin using different methods, euthanized with an overdose of quinaldine sulfate and fixed in 70% ethanol, labeled, and deposited in the Zoological Museum of Shiraz University (ZM-CBSU). Fish sampling was approved by the Ethic Committee of Biology Department, Shiraz University (SU-9630190). Taxonomic identifications were accomplished using available keys, primary taxonomic literature, and DNA sequence analysis [5, 7, 30, 31, 33–35, 39, 41]. The skull of each specimen was opened dorsally with a scalpel and the left sagittal otolith was extracted using fine forceps, cleaned in 5% KOH solution (~5 min), washed in distilled water (~5 min), and dried at room temperature. The otoliths were coated with gold [42], and SEM images were taken using a VEGA3 TESCAN at Shiraz University.
The original Caspian Sea material comprised otoliths isolated from 92 specimens belonging to 17 Ponto-Caspian endemics and one introduced species from the genus Rhinogobius Gill, 1859 (Table 1). The original Black Sea basin material comprised eight otoliths isolated from two species, Proterorhinus semilunaris (Heckel, 1837) and Mesogobius batrachocephalus (Pallas, 1814) (Table 1). In addition to the original material, otolith photographs of 14 extant species from the Ponto-Caspian basin were recycled from previous publications [12, 38, 43–45; Table 1]. Thus, a total of 118 otoliths representing three lineages, 11 genera (i.e., all Ponto-Caspian endemic gobiid genera except Babka), and 30 species were compiled. The 30 species included 17 species that are Caspian Sea endemics, six species that are Ponto-Caspian natives, two species that are Azov Sea endemics, three Black Sea/Azov Sea endemics, and one exotic species. Accordingly, this study includes otoliths from 23 species having distributions in the Caspian Sea basin, totaling 53.5% of the known Caspian Sea gobiofauna [31].
Status: presence status in the Ponto-Caspian basin, i.e., CE, Caspian Sea endemic; PCN, Ponto-Caspian native; AZE, Azov Sea endemic; BAN, Black Sea/Azov Sea native; BE, Black Sea endemic; EX, exotic. Quant.: taxonomic level of quantitative analysis, i.e., L, lineage; T, tribe; G, genus; S, species. N, number of specimens; SL, standard length in mm. Institutional abbreviations: CMNFI, Canadian Museum of Nature in Ottawa; NMNH NASU, Zoological Museum of Ukraine in Kiev; RBINS, Royal Belgian Institute of Natural Sciences; SMF, Senckenberg Museum, Frankfurt/Main; ZM-CBSU, Zoological Museum of Shiraz University, Collection of Biology Department, Shiraz, Iran; ZMMU/ZMMSU, Zoological Mueum, Moscow University in Moscow; ZMUC, Zoological Museum, University of Copenhagen; ZSM, SNSB-Bavarian State Collection of Zoology, Munich.
Inclusivity in global research
Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (S1 Checklist).
Otolith terminology, measurements, and variables
Terminology for the otolith morphology (Fig 1) followed Gierl et al. [46] and Schwarzhans et al. [45]. Using the SEM images, 12 measurements were taken for each otolith in ImageJ 1.52a: OL, maximal otolith length; OL2, minimal otolith length; OH, maximal otolith height; CL, colliculum length; SuL, sulcus length; SuH, sulcus height; OP, otolith perimeter (in mm); OA, otolith area (in mm2); SuP, sulcus perimeter (in mm); SuA, sulcus area (in mm2); SuTipV and SuEndV, distance from sulcus tip and end to the ventral margin, respectively. These measurements were used to calculate 24 otolith variables [45, 46]: OL/OH (= aspect ratio or ASr), OP/OL, OP/OH, SuA/OA, SuP/OP, SuP/SuTipV, SuP/SuEndV, SuL/OL, SuL/OH, SuL/SuH, SuL/SuTipV, SuL/SuEndV, SuL/OP, SuL/SuP, SuH/OL, SuH/OH, SuH/SuTipV, SuH/SuEndV, SuH/OP, SuH/SuP, SuTipV/OP, SuTipV/SuEndV, SuEndV/OP, and OL2/CL. Four inclination angles were measured [45]: α, inclination angle of ostium; β, inclination angle of anterior rim; γ, inclination angle of posterior rim; δ, inclination of line connecting preventral angle with tip of posterodorsal projection. In addition to ASr [47], another three shape indices were calculated [47]: (i) roundness (ROx = 4OA/πOL2), which is larger when the shape is more circular; (ii) rectangularity [REx = OA/(OL×OH)], with 1 being a perfect square and <1 being a nonsquare; and (iii) ellipticity [ELx = (OL–OH)/(OL+OH)], ranging from zero (a perfect round shape) to close to 1.0 (a spindle shape).
OL, maximal otolith length; OL2, minimal otolith length; OH, maximal otolith height; OA, otolith area; OP, otolith perimeter; CL, colliculum length; SuL, sulcus length; SuH, sulcus height; SuP, sulcus perimeter; SuA, sulcus area; SuTipV, distance from sulcus tip to the ventral margin; SuEndV, distance from sulcus end to the ventral margin; α, inclination angle of ostium; β, inclination angle of anterior rim; γ, inclination angle of posterior rim; δ, angle of preventral projection to posterodorsal projection traverse.
Otolith data analysis
The otoliths were analysed based on two different approaches:
- (i). A descriptive method for all 30 species.
- (ii). A quantitative method based on otolith morphometric variables and shape descriptors for which comparisons were made at four levels: lineage, tribe, genus, and species (Table 1). In the genus- and species-level analyses, genera and species with less than five otoliths were not considered. All otolith variables including morphometric ratios, inclination angles, and shape indices were analysed using IBM SPSS Statistics 26.0 [48]. The normal distributions of otolith variables for each species, genus, tribe, and lineage were compared using Shapiro-Wilk tests (p > 0.05). Mann-Whitney tests (p < 0.05) were used to evaluate the significance of non-normally distributed otolith variables; Univariate Analysis of Variance (ANOVA) with Tukey’s HSD (p < 0.05) and Dunnett’s T3 (p < 0.05) post-hoc tests (depending on homogeneity of variances, Levene’s test, p > 0.05) were used for taxon comparisons of normally distributed otolith variables. A Bonferroni correction (0.05/number of tests) was used at each taxonomic level to correct for multiple tests. Discriminant function analysis (DFA) was conducted to determine the proportion of otoliths that could be correctly assigned to their corresponding species, genera, tribes, and lineages. The classification success was tested by leave-one-out cross validation. A dendrogram was constructed based on Euclidean distance as a measure of dissimilarity to show the phenotypic relationships among the species. The between groups linkage method was used as the clustering algorithm.
Molecular phylogeny
There have been a few efforts to assess molecular phylogenetic relationships among benthophilines [29, 30, 61]. Neilson & Stepien [29] were the first to estimate relationships within this group; they used a combined dataset of two mitochondrial (cyt b and COI) and two nuclear loci (RAG1 and S7) with maximum parsimony (MP), maximum likelihood (ML), and Bayesian inference (BI) approaches. They introduced a revised classification and nomenclature that does not, however, aptly fit with the present understanding of gobioid family and subfamily phylogenetics (e.g., presented in Agorreta et al. [28] and McCraney et al. [60]), however, their primary phylogenetic inferences and tribe designations remain helpful as they are not in disagreement with any familial and subfamilial systematics. However, the most comprehensive phylogenetic analysis of the benthophiline gobies in terms of specimens and species numbers and also geographic coverage was conducted by Zarei et al. [30] based on mitochondrial COI sequences using ML and BI approaches, on which the current classification was based. That mitochondrial phylogeny is similar to the combined mito-nuclear phylogeny of Neilson and Stepien [29], except for (i) slightly different placement of Babka; and (ii) with regard to the sister group relationship of Ponticolini with Neogobiini rather than with Benthophilini; that grouping was, however, not robustly supported in Neilson and Stepien [29].
Results
Otolith descriptions
Gobius lineage.
Benthophilini. Anatirostrum. Anatirostrum profundorum. Right trapezoid (Fig 2M, from Schwarzhans et al. [45]); OL/OH 1.7; dorsal rim longer than ventral rim, with a shallow broad concavity in the middle, entire; predorsal angle orthogonal; posterodorsal projection long, broad, and rounded at tip; anterior rim vertical, slightly incised at the level of ostium, β 90°; posterior rim oblique, γ 111.3°, with a tiny incision slightly below the level of cauda; δ 22.1°; ventral rim horizontal, broadly undulate; preventral projection absent, orthogonal angle; posteroventral angle obtuse, nearly entire; sulcus centrally positioned, sole-shaped, horizontal, shallow, ostial lobes weakly developed; OL2/CL 2.7; subcaudal iugum indistinct; ventral furrow runs with a moderate distance to ventral rim; dorsal depression indistinct.
Benthophilus leobergius: A, ZM-CBSU 50; B, ZM-CBSU 51; C, ZM-CBSU 53; D, ZM-CBSU 54; E, ZMMU P.22625. Benthophilus pinchuki: F, ZM-CBSU 119; G, ZMMU P.16127. Benthophilus macrocephalus: H, ZMMU P.15889. Benthophilus durrelli: I, ZMMU P.21611. Benthophilus baeri: J, ZMMU P.16141. Benthophilus abdurahmanovi: K, ZMMU P.15890. Benthophilus stellatus: L, ZMMU P.11023. Anatirostrum profundorum: M, CMNFI 1999–0023.1. Benthophiloides brauneri: N, NMNH NASU 5136. Caspiosoma caspium: O, ZMMU P.13965. Scale bars = 0.5 mm.
Benthophilus. Benthophilus leobergius. Long elliptical (Fig 2A–2E); OL/OH 1.6–1.7; dorsal rim regularly curved and broad, crenate or crenulate, highest behind mid-length above CA apex, anteriorly depressed; predorsal angle variable in size, pointed or rounded at tip; posterodorsal projection broad and long, pointed or rounded at tip, moderately bent outwards; anterior rim oblique, β 67.8–79.6°, variably incised at or above the level of ostium; posterior rim almost parallel to anterior rim, γ 109.4–116.0°, without or with a narrow deep incision below the level of CA; δ 10.8–13.2°; ventral rim horizontal, curved upwards anteriorly and posteriorly, entire or slightly crenulate; preventral projection short or long, rounded or pointed at tip; posteroventral angle usually acute and projected, sometimes almost orthogonal; sulcus centrally positioned, sole-shaped, horizontal, shallow, dorsal ostial lobe weakly developed, ventral ostial lobe developed; OL2/CL 1.8–2.1; subcaudal iugum indistinct; ventral furrow runs close to ventral rim; dorsal depression distinct and narrow.
Benthophilus pinchuki. Long elliptical (Fig 2F and 2G); OL/OH 1.4–1.4; dorsal rim broad and curved, highest behind mid-length above caudal apex, irregularly sinuate or crenate, anteriorly strongly depressed; predorsal angle indistinct; posterodorsal projection broad and long, rounded at tip, moderately bent outwards; anterior rim in continuation of dorsal rim, oblique, β 60.5–71.6°, slightly incised at the level of ostium; posterior rim less oblique comparing to the anterior rim, γ 104.4–108.2, with a small incision below the level of cauda; δ 16.9–17.0°; ventral rim horizontal, entire; preventral projection short, rounded or pointed at tip; posteroventral angle projected or orthogonal, entire; sulcus centrally positioned, sole-shaped, horizontal, shallow, ostial lobes weakly developed; OL2/CL 2.7–2.9; subcaudal iugum absent; ventral furrow runs with a moderate distance to ventral rim; dorsal depression indistinct or distinct and relatively wide.
Benthophilus stellatus. Long elliptical (Fig 2L, from Schwarzhans et al. [45]); OL/OH 1.6; dorsal rim broad and curved, highest behind mid-length above caudal apex, sinuate or crenate, anteriorly strongly depressed; predorsal angle and anterior rim in continuation of dorsal rim; posterodorsal projection broad and long, tapering at tip; anterior rim in continuation of dorsal rim, β 62.3°; posterior rim less oblique comparing to the anterior rim, γ 109.9, incision indistinct; δ 18.0°; ventral rim horizontal, entire, preventral projection moderately long, rounded at tip; posteroventral angle obtuse, slightly undulate; sulcus centrally positioned, sole-shaped, α 5.2°, shallow, ostial lobes weakly developed; OL2/CL 3.0; subcaudal iugum absent; ventral furrow runs with a moderate distance to ventral rim; dorsal depression distinct and relatively wide.
Benthophilus baeri. Long elliptical (Fig 2J, from Schwarzhans et al. [45]); OL/OH 1.3; dorsal rim broad, regularly curved, sinuate and crenate, highest behind mid-length behind caudal apex, anteriorly strongly depressed; predorsal angle and anterior rim in continuation of dorsal rim; posterodorsal projection broad and long, tapering at tip; anterior rim in continuation of dorsal rim, β 78.7°, incised at or slightly above the level of ostium; posterior rim oblique as anterior rim, γ 121.0°, incised slightly below the level of cauda; δ 23.6°; ventral rim almost horizontal, undulate; preventral projection short and broad; posteroventral angle undulate; sulcus small, centrally positioned, sole-shaped, α 14.0°, shallow, ostial lobes weakly developed; OL2/CL 2.9; subcaudal iugum absent; ventral furrow runs with a moderate distance to ventral rim; dorsal depression distinct and relatively wide.
Benthophilus abdurahmanovi. Long elliptical (Fig 2K, from Schwarzhans et al. [45]); OL/OH 1.3; dorsal rim broadly crenate, highest behind mid-length above cauda, anteriorly strongly depressed; predorsal angle well-developed and rounded; posterodorsal projection very short, blunt; anterior rim almost vertical or little oblique, β 83.5°, incised above the level of ostium; posterior rim almost vertical, γ 95.9°, with an incision above the level of cauda; δ 20.7°; ventral rim curved, entire or broadly crenate; preventral projection short but broad, rounded; posteroventral angle markedly broad, obtuse, slightly undulate; sulcus centrally positioned, sole-shaped, α 13.0°, shallow, ostial lobes weakly developed; OL2/CL 2.8; subcaudal iugum absent; ventral furrow runs with a close distance to ventral rim; dorsal depression distinct and relatively narrow.
Benthophilus durrelli. Long elliptical (Fig 2I, from Schwarzhans et al. [45]); OL/OH 1.4; dorsal rim gently curved, mostly entire with a shallow incision in mid-length, highest behind incision above cauda, anteriorly slightly depressed; predorsal angle well-developed, broad and round; posterodorsal projection broad and long, rounded at tip; anterior rim little oblique, β 82.9°, incised slightly above the level of ostium; posterior rim more oblique comparing to the anterior rim, γ 108.1°, with a broad incision at the level of cauda; δ 19.1; ventral rim nearly horizontal or gently curving, entire; preventral projection short and broad, rounded at tip; posteroventral angle slightly projected, entire; sulcus centrally positioned, sole-shaped, horizontal, shallow, ostial lobes weakly developed; OL2/CL 2.4; subcaudal iugum absent; ventral furrow runs with a moderate distance to ventral rim; dorsal depression distinct and relatively narrow.
Benthophilus macrocephalus. Long elliptical (Fig 2H, from Schwarzhans et al. [45]); OL/OH 1.4; dorsal rim curved, highest behind caudal apex, irregularly crenate and sinuate, anteriorly strongly depressed; predorsal angle and anterior rim in continuation of dorsal rim; posterodorsal projection short and broad, rounded but strongly crenulate; anterior rim crenulate, incision not recognizable, oblique, β 63.4°; posterior rim crenulate, less oblique comparing to the anterior rim, γ 104.3°, incision not recognizable; δ 13.1°; ventral rim nearly horizontal or gently curving, crenulate; preventral projection short and rounded at tip; posteroventral angle crenulate; sulcus centrally positioned, sole-shaped, α 13.7°, shallow, dorsal ostial lobe indistinct, ventral ostial lobe weakly developed; OL2/CL 2.4; subcaudal iugum absent; ventral furrow runs with a close to moderate distance to ventral rim; dorsal depression distinct and relatively narrow.
Benthophiloides. Benthophiloides brauneri. Long rectangle (Fig 2N, from Schwarzhans et al. [45]); OL/OH 1.4; dorsal rim entire, posterior part horizontal to convex and regularly curved, highest above cauda; predorsal angle nearly orthogonal; posterodorsal projection very short and rounded; anterior rim vertical, β 90°, not incised; posterior rim almost vertical, γ ~90°, with a shallow broad concavity at the level of cauda; δ 22.0°; ventral rim horizontal, entire; preventral projection absent, almost orthogonal angle; posteroventral angle orthogonal, entire; sulcus centrally positioned, sole-shaped, horizontal, shallow, ostial lobes weakly developed; OL2/CL 1.8; subcaudal iugum absent; ventral furrow runs with a moderate distance to ventral rim; dorsal depression distinct and narrow.
Caspiosoma. Caspiosoma caspium. Pentagonal (Fig 2O, from Schwarzhans et al. [12]); OL/OH 1.1; dorsal rim angular, highest above cauda, entire; predorsal angle obtuse; posterodorsal projection very short or indistinct, obtuse angle; anterior rim incised slightly below the level of ostium, vertical, β ~90°; posterior rim not incised, γ 97.4°; δ 24.3°; ventral rim horizontal, entire; preventral projection very short, tapering at tip; posteroventral angle almost orthogonal or slightly obtuse, entire; sulcus centrally positioned, sole-shaped, α 7.5°, relatively deep, dorsal ostial lobe indistinct, ventral ostial lobe developed; OL2/CL 2.0; subcaudal iugum absent; ventral furrow runs with a close distance to ventral rim; dorsal depression distinct.
Neogobiini. Neogobius. Neogobius bathybius. Two-humped long rectangle (Fig 3P–3T); OL/OH 1.1–1.2; dorsal rim two-humped with a deep broad V-shaped concavity at mid-length (i.e., M-shaped dorsal rim), humps angular, posterior hump markedly broader and higher, outline entire; posterodorsal projection bulky and very broad, not bending outwards; anterior rim usually little oblique, β 77.3–84.7°, with or without incision, entire; posterior rim almost parallel to the anterior rim or slightly more oblique, γ 96.7–105.9°, usually with a shallow broad concavity; δ 13.5–22.1°; ventral rim almost horizontal or gently curving, entire; preventral projection usually absent or a short round projection present; posteroventral angle almost orthogonal, entire; sulcus supramedian, sole- or dumbbell-shaped, cauda almost as wide as ostium, α 2.1–10.6°, very deep, ostial lobes usually well-developed; OL2/CL 1.5–1.6; subcaudal iugum usually absent, its length 1/2 cauda length and very slender if present; ventral furrow runs with a large distance to ventral rim; dorsal field absent.
Neogobius pallasi: A, ZM-CBSU 20; B, ZM-CBSU 22; C, ZM-CBSU 23; D, ZM-CBSU 27. Neogobius fluviatilis: E, ZMMU, P.22433. Neogobius caspius: F, ZM-CBSU 2; G, ZM-CBSU 6; H, ZM-CBSU 12; I, ZM-CBSU 13; J, ZM-CBSU 14. Neogobius melanostomus: K, ZM-CBSU 15; L, ZM-CBSU 16; M, ZM-CBSU 18; N, ZM-CBSU 19; O, unnumbered RBINS specimen [43]. Neogobius bathybius: P, ZM-CBSU 101; Q, ZM-CBSU 106; R, ZM-CBSU 108; S, ZM-CBSU 111; T, ZM-CBSU 110. Scale bars = 0.5 mm.
Neogobius caspius. Square-rhomboid (Fig 3F–3J); OL/OH 1.1–1.3; dorsal rim curved, highest at mid-length or in front of it above ostium, posteriorly strongly depressed, dentate, denticulate or crenulate; predorsal angle almost orthogonal to slightly obtuse; posterodorsal projection very long, narrow, tapering and pointed, strongly bent outwards and dorsally as well; anterior rim vertical or little oblique, β 77.0–85.7°, not incised, dentate, denticulate or entire, higher than posterior rim; posterior rim oblique as anterior rim, γ 102.4–111.6°, with or without a shallow concavity below the projection; δ 26.6–33.4°; ventral rim nearly horizontal, sinuate or crenulate; preventral projection short and usually pointed; posteroventral angle orthogonal, entire or slightly denticulate/crenulate; sulcus centrally positioned, sole-shaped, α 14.8–17.9°, relatively deep, ostial lobes developed; OL2/CL 1.4–1.8; subcaudal iugum present, 1/2 cauda length, very slender; ventral furrow runs with a moderate distance to ventral rim; dorsal depression distinct and wide.
Neogobius fluviatilis. Discoid-rhomboid (Fig 3E, from Bratishko et al. [44]); OL/OH 1.1; dorsal rim convex, regularly curved, slightly highest behind mid-length above cauda, entire to slightly crenate, anteriorly depressed; predorsal angle obtuse; posterodorsal projection long and broad, slightly tapering or blunt at tip; anterior rim almost vertical or little oblique, β 85.6°, incision not recognizable from crenation; posterior rim more oblique comparing to the anterior rim, γ 104.5°, a broad concavity presents below the projection; δ 24.7°; ventral rim nearly horizontal, entire to slightly sinuate; preventral projection not recognizable from crenation; posteroventral angle nearly orthogonal, entire; sulcus centrally positioned, sole-shaped, α 14.3°, shallow, ostial lobes weakly developed; OL2/CL 1.6; subcaudal iugum present, 1/2 cauda length, relatively narrow; ventral furrow runs with a moderate distance to ventral rim; dorsal depression distinct and wide.
Neogobius melanostomus. Square-rhomboid (Fig 3K–3O); OL/OH 1.1–1.2; dorsal rim convex, regularly curved, crenate, coarsely crenate or entire, with a deep U-shaped indentation at mid-length; predorsal angle slightly obtuse to orthogonal; posterodorsal projection long and tapering, usually blunt at tip, strongly bent outwards; anterior rim oblique, β 77.3–84.6°, incision absent or not recognizable from dentation; posterior rim oblique as anterior rim, γ 102.3–109.4°, usually without or rarely with a shallow concavity; δ 24.7–27.2°; ventral rim nearly horizontal or gently curved, entire or undulate; preventral projection long, and blunt or pointed at tip; posteroventral angle nearly orthogonal, usually entire, sometimes dentate; sulcus centrally positioned, sole-shaped, α 17.0–19.8°, relatively deep, ostial lobes well-developed; OL2/CL 1.4–1.7; subcaudal iugum 1/2 cauda length, thick; ventral furrow runs with a moderate distance to ventral rim; dorsal depression distinct and relatively wide.
Neogobius pallasi. Discoid-rhomboid (Fig 3A–3D); OL/OH 1.0–1.1; dorsal rim convex, regularly curved, highest behind mid-length above cauda, entire to denticulate, anteriorly depressed; predorsal angle obtuse or in continuation of dorsal rim; posterodorsal projection short and broad, rounded or blunt at tip, slightly bent outwards; anterior rim nearly vertical or slightly oblique, β 83.9–88.8°, incision not recognizable from crenation; posterior rim little more oblique comparing to the anterior rim, γ 95.7–112.2°, usually with a shallow concavity; δ 24.3–31.0°; ventral rim nearly horizontal to gently curved, entire to slightly undulate; preventral projection absent; posteroventral angle nearly orthogonal, entire to slightly undulate; sulcus centrally positioned, sole-shaped, α 16.0–22.5°, relatively deep, ostial lobes well-developed; OL2/CL 1.4–1.6; subcaudal iugum 1/2 cauda length, thick to relatively slender; ventral furrow runs with a moderate to close distance to ventral rim; dorsal depression distinct and wide.
Ponticolini. Mesogobius. Mesogobius batrachocephalus. Long parallelogram with a deep irregular dorsal concavity (Fig 4A–4C); OL/OH 1.5–1.6; dorsal rim with a deep irregular concavity at mid-length, highest behind mid-length; posterodorsal projection very long and pointed, slightly bending outwards; anterior rim oblique, β 60.6–67.1°, slightly incised at the level of ostium, slightly dentate to sinuate; posterior rim parallel to the anterior rim, γ 112.5–113.1°, not incised; δ 15.1–21.8°; ventral rim almost horizontal, gently curving, entire to slightly undulate; preventral projection very long and pointed; posteroventral angle orthogonal to obtuse, entire to slightly undulate; sulcus centrally positioned, sole-shaped, α 6.8–13.4°, relatively deep, ostial lobes developed; OL2/CL 1.4–1.6; subcaudal iugum present, its length 1/2 cauda length and slender; ventral furrow runs with a close to intermediate distance to ventral rim; dorsal depression distinct and wide.
Mesogobius batrachocephalus: A–B, ZMUC P2395071-72; C, drawing from Nolf [49]. Mesogobius nonultimus: D, ZM-CBSU S036-1. Scale bars = 1 mm.
Mesogobius nunultimus. Two-humped long rectangle (Fig 4D); OL/OH 1.3; dorsal rim two-humped with a deep broad V-shaped concavity at mid-length (i.e., M-shaped dorsal rim), humps angular, posterior hump markedly broader and higher, outline entire to sinuate; posterodorsal projection bulky and very broad, not bending outwards; anterior rim oblique, β 73.7°, with a small incision above the level of ostial apex, entire; posterior rim almost parallel to the anterior rim or slightly less oblique, γ 101.6°, with a shallow small incision below the level of cauda; δ 27.2°; ventral rim almost horizontal, entire; preventral projection very long and pointed; posteroventral angle rounded, entire; sulcus centrally positioned, sole-shaped, α 17.5°, relatively deep, ostial lobes developed; OL2/CL 1.7; subcaudal iugum present, its length 1/2 cauda length and slender; ventral furrow runs with a moderate to large distance to ventral rim; dorsal depression indistinct.
Ponticola. Ponticola gorlap. Long parallelogram (Fig 5N–5R); OL/OH 1.3–1.6; dorsal rim horizontal, irregularly sinuate and dentate, anteriorly slightly depressed; predorsal angle obtuse; posterodorsal projection moderately long and tapering, pointed or blunt at tip, strongly bent outwards; anterior rim oblique, β 66.1–81.3°, not incised or slightly incised at the level of ostium or above it; posterior rim almost parallel to the anterior rim or a little less oblique, γ 92.5–104.6°, with a marked concavity at or slightly above the level of cauda; δ 24–28°; ventral rim horizontal, entire or slightly undulate, slightly projecting downwards near the posterior end; preventral projection long, pointed or blunt at tip; posteroventral angle usually broadly rounded, undulate, crenate or denticulate; sulcus centrally positioned, sole-shaped, α 9.4–18.8°, moderately deep, ostial lobes well-developed, long, relatively wide, OL2/CL 1.4–1.5; subcaudal iugum present, 1/3 cauda length, slender; ventral furrow runs with a moderate to close distance to ventral rim; dorsal depression distinct, narrow to moderately wide.
Ponticola iranicus: A, ZM-CBSU 37; B, ZM-CBSU 38; C, ZM-CBSU 39; D, ZM-CBSU 40; E, ZM-CBSU 141; F, ZM-CBSU 144. Ponticola patimari: G, ZM-CBSU P1; H, ZM-CBSU P2; I, ZM-CBSU P3; J, ZM-CBSU P4; K, ZM-CBSU P6. Ponticola syrman: L, ZM-CBSU 142; M, ZM-CBSU 139. Ponticola gorlap: N, ZM-CBSU 41; O, ZM-CBSU 44; P, ZM-CBSU 45; Q, ZM-CBSU 46; R, ZM-CBSU 48. Ponticola iljini: S–T, ZMMSU P-23516; U–V, Ponticola kessleri: non-catalogued specimens from the RBINS collection, Danube, near Vienna in Austria [43]. Ponticola hircaniaensis: W, ZM-CBSU K18; X, ZM-CBSU K28 [5]. Scale bars = 0.5 mm.
Ponticola iljini. Long parallelogram (Fig 5S abd 5T, from Vasil’eva et al. [38]); OL/OH 1.5–1.8; dorsal rim nearly horizontal, anteriorly strongly depressed, entire to slightly undulate; predorsal angle obtuse; posterodorsal projection long and broad, tapering or blunt at tip, strongly bent outwards; anterior rim oblique, β 61.5–72.0°, incised at the level of ostium; posterior rim less oblique comparing to the anterior rim, γ 95.7–112.2°, incised at the level of cauda; δ 16.8–27.8°; ventral rim horizontal, undulate or sinuate, slightly projecting downwards near the posterior end; preventral projection long and tapering; posteroventral angle broadly rounded, crenate; sulcus centrally positioned, sole-shaped, α 13.7–15.4°, moderately deep, ostial lobes developed; OL2/CL 1.6–1.7; subcaudal iugum present, 1/3 cauda length, slender; ventral furrow runs with a moderate to close distance to ventral rim; dorsal depression distinct, narrow to moderately wide.
Ponticola hircaniaensis. Long parallelogram (Fig 5W and 5X); OL/OH 1.2–1.4; dorsal rim horizontal with a broad U-shaped concavity in the middle, entire; predorsal angle obtuse; posterodorsal projection highly positioned, long and broad, pointed or blunt at tip, slightly bents outwards; anterior rim oblique, β 72.3–84.3°, broadly incised at or slightly above the level of ostium; posterior rim almost parallel to the anterior rim, γ 101.3–104.9°, incised at or slightly below the level of cauda; δ 24.4–34.1°; ventral rim horizontal and entire; preventral projection usually long, sometimes short, pointed or blunt at tip; posteroventral angle usually orthogonal, entire or undulate; sulcus centrally positioned, sole-shaped, α 10.2–19.1°, very deep, with developed ostial lobes; OL2/CL 1.6–1.7; subcaudal iugum present, usually 1/3 cauda length, slender; ventral furrow runs with a moderate distance to ventral rim; dorsal depression indistinct.
Ponticola iranicus. Long parallelogram (Fig 5A–5F); OL/OH 1.4–1.6; dorsal rim nearly horizontal, sometimes with a shallow broad concavity in the middle, usually entire, sometimes crenate; predorsal angle slightly obtuse; posterodorsal projection very long and broad, pointed or blunt at tip, highly positioned, strongly bent outwards, and facing upwards dorsally; anterior rim oblique, β 66.1–78.9°, with a broad concavity or incision not recognizable from crenation; posterior rim almost parallel to the anterior rim, γ 103.7–110.9°, without a recognizable incision or notch; δ 20.9–28.6°; ventral rim horizontal, undulate or sinuate, usually slightly projecting downwards near the posterior end; preventral projection long, pointed or blunt at tip; posteroventral angle usually obtuse, sometimes orthogonal, entire, undulate or sinuate; sulcus centrally positioned, sole-shaped, α 11.0–16.4°, very deep, ostial lobes usually well-developed; OL2/CL 1.4–1.7; subcaudal iugum present or indistinct, 1/2 cauda length if present, slender; ventral furrow runs with a close to large distance to ventral rim; dorsal depression absent or distinct, narrow to moderately wide if distinct.
Ponticola kessleri. Long parallelogram (Fig 5U and 5V, from Vasil’eva et al. [38], and Jacobs and Hoedemakers [43]); OL/OH 1.3–1.6; dorsal rim nearly horizontal, anteriorly depressed, crenate, sinuate or undulate; predorsal angle obtuse; posterodorsal projection long and broad, blunt at tip, strongly bent outwards; anterior rim oblique, β 60.2–79.1°, slightly incised at the level of ostium; posterior rim less oblique comparing to the anterior rim, inclined at 95.4–108.0°, without or with a distinct incision; δ 24.1–28.3°; ventral rim horizontal, entire to slightly undulate, slightly projecting downwards near the posterior end; preventral projection long, and pointed or blunt at tip; posteroventral angle almost orthogonal or obtuse, crenate or sinuate; sulcus centrally positioned, sole-shaped, α 10.7–18.4°, moderately deep, ostial lobes developed; OL2/CL 1.4–1.5; subcaudal iugum present, 1/2 cauda length, slender; ventral furrow runs with a moderate to close distance to ventral rim; dorsal depression distinct or indistinct, narrow to moderately wide if distinct.
Ponticola patimari. Long parallelogram (Fig 5G–5K); OL/OH 1.4–1.6; dorsal rim horizontal, entire, crenate or sinuate; predorsal angle orthogonal to slightly obtuse; posterodorsal projection long, usually pointed or blunt at tip, strongly bent outwards and usually slightly facing upwards dorsally; anterior rim oblique, β 68.1–77.8°, with or without incision at the level of ostium; posterior rim parallel to the anterior rim, γ 98.5–109.5°, without or with a shallow concavity; δ 22.2–27.6°; ventral rim horizontal, usually entire to slightly undulate; preventral projection long, pointed or blunt at tip; posteroventral angle orthogonal, usually entire to slightly undulate; sulcus centrally positioned, sole-shaped, α 7.8–13.8°, very deep, ostial lobes developed; OL2/CL 1.5–1.8; subcaudal iugum usually absent or small, 1/2–1/3 cauda length, slender; ventral furrow runs with a moderate distance to ventral rim; dorsal depression usually distinct and moderately wide.
Ponticola syrman. Long parallelogram (Fig 5L and 5M); OL/OH 1.4; dorsal rim horizontal with a broad concavity in the middle, entire to slightly sinuate; predorsal angle well-developed, projecting forwards and upwards dorsally; posterodorsal projection very long and broad, highly positioned, tapering or pointed at tip, strongly bent outwards; anterior rim nearly vertical or little oblique, β 78.5–82.6°, with a broad concavity in the middle; posterior rim oblique, γ 105.4–108.0°, not incised; δ 23.5–27.3°; ventral rim nearly horizontal or gently curving, entire to slightly undulate; preventral projection short, blunt at tip; posteroventral angle nearly orthogonal or slightly obtuse, sometimes undulate or sinuate; sulcus centrally positioned, sole-shaped, α 4.7–13.3°, very deep, dorsal ostial lobe weakly developed or indistinct, ventral ostial lobe developed; OL2/CL 1.7–1.9; subcaudal iugum indistinct or very small, 1/4 cauda length; ventral furrow runs with a large distance to ventral rim; dorsal depression absent or indistinct.
Proterorhinus. Proterorhinus nasalis. Square-trapezoid (Fig 6A–6F); OL/OH 0.8–1.0; dorsal rim variable in shape, rounded, horizontal, and sometimes angular in middle, entire or crenulate, usually shorter than ventral rim; predorsal angle obtuse to orthogonal; posterodorsal projection very short; anterior rim nearly vertical to little oblique, β 81.0–88.4°, usually not incised or sometimes incised above the level of ostium; posterior rim oblique to vertical, γ 80.0–90.0°, usually with a shallow broad incision above the level of cauda; δ 34.2–41.4°; ventral rim horizontal, longer than dorsal rim, usually broadly undulate; preventral projection absent to short, rounded, pointed or blunt; posteroventral angle projected, broad and rounded, sometimes nearly orthogonal; sulcus centrally positioned, sole-shaped, α 0.0–14.8°, moderately deep, ostial lobes variably developed; OL2/CL 1.5–1.8; subcaudal iugum usually indistinct to very small, 1/3 cauda length; ventral furrow runs with a large distance to ventral rim; dorsal depression indistinct or distinct and wide.
Proterorhinus nasalis: A, ZM-CBSU 56; B, ZM-CBSU 57; C, ZM-CBSU 59; D, ZM-CBSU 60; E, ZM-CBSU 61; F, ZM-CBSU 62. Proterorhinus semilunaris: G, ZM-CBSU 122; H, ZM-CBSU 126; I, ZM-CBSU 125. Scale bar = 0.2 mm.
Proterorhinus semilunaris. Square (Fig 6G–6I); OL/OH 0.9–1.0; dorsal rim variable in shape, horizontal or depressed anteriorly, undulate to crenate; predorsal angle obtuse to orthogonal; posterodorsal projection broad, rounded to blunt; anterior rim nearly vertical, β 82.1–108.2°, with or without a small incision; posterior rim nearly parallel to anterior rim, γ 79.1–90.0°, with a shallow broad incision at the level of cauda; δ 33.3–39.1°; ventral rim horizontal, broadly undulate; preventral projection absent, short or long, blunt or pointed; posteroventral angle nearly orthogonal or little obtuse, broadly undulated; sulcus centrally positioned, sole-shaped, α 11.1–20.2°, moderately deep to shallow, ostial lobes variably developed; OL2/CL 1.6–2.1; subcaudal iugum and ventral furrow indistinct because of bad preservation; dorsal depression indistinct or distinct and wide.
Pomatoschistus lineage. Knipowitschia. Knipowitschia caucasica. Pentagonal (Fig 7A–7F); OL/OH 0.9–1.0; dorsal rim angular, highest medially, thereafter inclined with a steep slope anteriorly and posteriorly, usually entire; predorsal angle and posterodorsal projection usually distinct and somewhat symmetrical; otolith sometimes widest across predorsal angle to posterodorsal projection; anterior rim vertical to little oblique, β 90.0–97.6°, with a broad shallow incision at the level of ostium or slightly above it; posterior rim vertical to oblique, γ 90.0–110.1°, usually with a broad shallow incision at the level of cauda; δ 26.3–40.2°; ventral rim usually horizontal or little curved, entire; preventral projection absent; posteroventral angle nearly orthogonal, little obtuse or projected, entire; sulcus centrally positioned, sole-shaped, α 12.7–22.7°, moderately deep or shallow, ostial lobes variably developed; OL2/CL 1.8–2.2; subcaudal iugum indistinct (because of bad preservation); ventral furrow indistinct (because of bad preservation) or runs with a moderate distance to ventral rim; dorsal depression indistinct or distinct and wide.
Knipowitschia caucasica: A, ZM-CBSU 63; B, ZM-CBSU 64; C, ZM-CBSU 80; D, ZM-CBSU 82; E, ZM-CBSU 128; F, ZM-CBSU 129. Knipowitschia longecaudata: G, ZM-CBSU 65; H, ZM-CBSU 66; I, ZM-CBSU 67; J, ZM-CBSU 68; K, ZM-CBSU 69; L, ZM-CBSU 92; M, ZM-CBSU 93; N, ZM-CBSU 94; O, ZM-CBSU 95; P, ZM-CBSU 96; Q, ZM-CBSU 97. Hyrcanogobius bergi: R, ZM-CBSU 84; S, ZM-CBSU 85; T, ZM-CBSU 86; U, ZM-CBSU 87; V, ZM-CBSU 90; W, ZM-CBSU 91. Rhinogobius sp.: X, ZM-CBSU 70; Y, ZM-CBSU 71; Z, ZM-CBSU 72; A’, ZM-CBSU 73; B’, ZM-CBSU 74; C’, ZM-CBSU 76. Scale bars = 0.2 mm.
Knipowitschia longecaudata. Pentagonal (Fig 7G–7Q); OL/OH 0.9–1.0; dorsal rim angular, highest infront of mid-length, thereafter inclined with a steep slope anteriorly and posteriorly, entire; predorsal angle and posterodorsal projection obtuse; anterior rim almost vertical, β 81.3–90.0, usually without or with a tiny shallow incision above the level of ostium; posterior rim nearly vertical, γ 87.4–104.7, with a broad shallow incision at the level of cauda; δ 24.9–30.4; ventral rim horizontal, entire; sulcus centrally positioned, sole-shaped, α 15.5–26.1, moderately deep, ostial lobes variably developed; OL2/CL 1.7–2.1; subcaudal iugum thick and long, as long as cauda; ventral furrow runs with a large to moderate distance to ventral rim; dorsal depression indistinct or distinct and wide.
Hyrcanogobius. Hyrcanogobius bergi. Pentagonal (Fig 7R–7W); OL/OH 0.9–1.0; dorsal rim usually angular to rounded, highest medially, thereafter inclined with a steep slope anteriorly and posteriorely, entire; predorsal angle and posterodorsal projection obtuse; anterior rim nearly vertical, β 81.3–90.0°, usually without incision, sometimes a small shallow incision above level of ostium present; posterior rim nearly vertical, γ 87.4–104.7°, with a broad shallow incision at the level of cauda; δ 24.9–30.4°; ventral rim horizontal, entire, preventral projection absent; sulcus centrally positioned, sole-shaped, α 15.5–26.1°; moderately deep, ostial lobes variably developed; OL2/CL 1.7–2.1; subcaudal iugum large and long, as long as cauda or slightly shorter; ventral furrow runs with a moderate distance to ventral rim; dorsal depression indistinct or distinct and wide.
Acanthogobius lineage.
Rhinogobius. Rhinogobius sp. Almost pentagonal or with a rounded dorsal margin (Fig 7X–7C’); OL/OH 0.9–1.0; dorsal rim angular or rounded, highest medially, entire to sinuate; predorsal angle and posterodorsal projection obtuse; anterior rim nearly horizontal, β ~90°, without or with a small shallow incision; posterior rim nearly vertical, γ ~90°, with or without a shallow broad incision at the level of cauda; δ 32.6–42.5°; ventral rim usually horizontal, sometimes rounded, entire to little undulate; preventral projection usually indistinct to very small; sulcus centrally positioned, sole-shaped, α 12.9–24.1°, moderately deep, ostial lobes variably developed; OL2/CL 1.6–2.2; subcaudal iugum wide and long, as long as cauda or slightly shorter; ventral furrow runs with a moderate distance to ventral rim; dorsal depression indistinct.
Otolith morphometric variables and classical shape descriptors
Variation among the lineages.
Twenty-eight otolith variables significantly differed between two or more of the three lineages (Table 2, S1 Fig, S1 Table). Fourteen variables were normally distributed: three involved the SuL, three the SuH, two the SuP, and the remainder were OL/OH, OP/OH, SuA/OA, SuTipV/SuEndV, SuEndV/OP, and REx. Fourteen variables were non-normally distributed: four inclination angles (α, δ, γ, and β), three the SuL, two shaple indices (ROx and ELx), and the remainder were OP/OL, SuP/OP, SuH/SuP, SuTipV/OP, and OL2/CL.
These characters indicated clear separation of the Gobius lineage from the Pomatoschistus (28 variables) and Acanthogobius (23 variables) lineages (Table 2, S1 Fig, S1 Table). The Pomatoschistus and Acanthogobius lineages however, were separated by just two variables (δ and γ). Therefore, δ and γ were the variables that discriminated among all three lineages, whereas 22 variables distinguished the Gobius lineage from both the Pomatoschistus and Acanthogobius lineages.
The 28 variables were subjected to DFA, yielding two discriminant functions for the variables, DF 1 accounted for 94.1% (eigenvalue: 5.608; λ = 0.112, p < 0.0001) and DF 2 accounted for 5.9% (eigenvalue: 0.353; λ = 0.739, p = 0.13) of the among-group variability. In order of importance, the most significant variables loadings on DF 1 and DF 2 were OL/OH, ROx, OP/OL, OP/OH, SuP/OP, SuP/SuEndV and OP/OH, OL/OH, SuP/SuTipV, SuH/SuTipV, SuH/OH, OP/OL, respectively. The DFA biplot showed clear separation between the Gobius lineage and the other two lineages (Fig 8). The proportions of individuals correctly classified into their original lineages were 91.3%, 77.8%, and 57.1% for the Gobius, Pomatoschistus, and Acanthogobius lineages, respectively (Table 3).
(a) Lineage, (b) tribe, (c) genus, and (d–f) species in different genera.
Correctly classified samples are shown in bold.
Variation among the benthophiline tribes.
Twenty-nine otolith variables revealed significant differences between at least two of the three benthophiline tribes (Table 4, S2 Fig; S2 Table). Ten variables were normally distributed: three involved the SuL, four the SuH, and the remainder were SuA/OA, SuEndV/OP, and β. Nineteen variables were non-normally distributed: three inclination angles (α, δ, and γ), four involved the SuL, three the SuP, two the OP, two shape indices (ROx and ELx), and the remainder were OL/OH, SuH/SuEndV, SuTipV/OP, SuTipV/SuEndV, and OL2/CL.
Benthophilini was separated from Neogobiini and Ponticolini by 25 and 21 variables, respectively (Table 4, S2 Fig; S2 Table). Neogobiini and Ponticolini were separated by 17 variables. SuL/SuP, SuH/OL, SuEndV/OP, α, SuP/SuTipV, SuP/SuEndV, and SuL/SuTipV discriminated among the three tribes.
The 29 variables were subjected to DFA, which produced two discriminant functions, DF 1 accounted for 60.7% (eigenvalue: 7.836; λ = 0.019, p < 0.0001) and DF 2 for 39.3% (eigenvalue: 5.064; λ = 0.165, p < 0.0001) of the among-group variability. In order of importance, the most significant variables loadings on DF 1 and DF 2 were SuL/OH, SuH/OL, SuH/OH, SuP/SuTipV, SuH/OP, SuH/SuTipV and SuH/OP, SuP/SuTipV, SuH/OL, SuH/OH, SuL/OH, SuL/OH, respectively. The DFA biplot revealed clear separation among the three tribes (Fig 8), for which classification successes were estimated (Table 5). Proportions of individuals that were correctly assigned to their original lineages were 93.3%, 100%, and 90.3% for Benthophilini, Neogobiini, and Ponticolini, respectively.
Correctly classified samples are shown in bold.
Variation among the genera.
Thirty-one otolith variables revealed significant differences between at least two of the seven genera (Table 6, S3 Fig, S3 Table). Thirteen variables were normally distributed: four variables involved the SuL, four the SuH, and the remainder were γ, OL/OH, OP/OH, SuEndV/OP, and SuH/SuP. Eighteen variables were non-normally distributed: three inclination angles (α, δ, and β), three involved the SuL, three the SuP, two the SuH, two shape indices (ROx and ELx), and the remainder were OP/OL, SuA/OA, SuTipV/OP, SuTipV/SuEndV, and OL2/CL.
Benthophilus clearly diverged from the other three benthophiline genera, i.e., Neogobius (22 variables), Ponticola (19), and Proterorhinus (21) (Table 6, S3 Fig, S3 Table). It was distinguished from Hyrcanogobius, Knipowitschia, and Rhinogobius, each by 23 variables. The strongest parameters in this respect were SuL/OL, SuH/OL, SuH/OP, SuP/OP, and OL2/CL, which separated Benthophilus from all other genera.
Neogobius was clearly distinctive from the other three benthophiline genera, i.e., Benthophilus (22 variables), Ponticola (20), and Proterorhinus (20). It also diverged from Hyrcanogobius, Knipowitschia, and Rhinogobius, by 25, 27, and 22 variables, respectively. The strongest contributors were OL/OH, OP/OH, SuEndV/OP, ROx, SuP/SuTipV, SuL/SuTipV, and SuH/SuEndV in separating Neogobius from all other genera.
Ponticola differed from the other three benthophiline genera, i.e., Benthophilus (19 variables), Neogobius (20), and Proterorhinus (25). It also diverged from Hyrcanogobius, Knipowitschia, and Rhinogobius, by 28, 28, and 27 variables, respectively. The characters SuL/OH, SuL/SuTipV, SuL/SuEndV, SuH/SuTipV, and SuH/SuEndV best supported the distinction of Ponticola from all other genera.
Proterorhinus diverged from the other three benthophiline genera, i.e., Benthophilus (21 variables), Neogobius (20), and Ponticola (25). It differed in γ, OL/OH, OP/OH, SuEndV/OP, ROx, δ, OP/OL, SuP/OP, SuL/OP, SuTipV/OP, REx, and ELx. Proterorhinus differed from Hyrcanogobius, Knipowitschia, and Rhinogobius, by 9, 10, and 6 variables, respectively. It was distinct from them in α, β, OP/OL, SuTipV/SuEndV, SuTipV/OP, and OL2/CL. The greatest contributing characters were OP/OL and SuTipV/OP, supporting the separation of Proterorhinus from all other genera.
Hyrcanogobius varied from Knipowitschia and Rhinogobius by small differences in four (SuL/SuP, α, δ, REx) and two (γ, δ) variables, respectively. Knipowitschia and Rhinogobius varied in only γ and δ. Therefore, just a single otolith variable (i.e., δ) supported the differentiation of these three genera.
Therefore, of the 31 otolith variables, none distinguished all seven genera. OL/OH, OP/OH, SuEndV/OP, ROx, γ, δ, SuP/OP, REx, and ELx discriminated between two groups, one comprising Benthophilus + Neogobius + Ponticola, and the second group containing Proterorhinus + Hyrcanogobius + Knipowitschia + Rhinogobius.
The 31 variables were subjected to DFA; DF 1 accounted for 50.8% (eigenvalue: 17.698; λ = 0.001, p < 0.0001) and DF 2 for 23.8% (eigenvalue: 8.291; λ = 0.002, p < 0.0001) of among-group variability. In order of importance, the most significant variables loadings on DF 1 and DF 2 were SuH/OL, SuH/OP, SuH/OH, SuH/SuEndV, SuP/SuTipV, SuH/SuP and SuL/SuTipV, SuP/SuTipV, SuH/OH, SuH/OP, SuL/OH, OP/OH, respectively. The DFA biplot showed clear separation of Benthophilus, Neogobius, and Ponticola, whereas the other four genera, Proterorhinus, Hyrcanogobius, Knipowitschia, and Rhinogobius overlapped (Fig 8). Classification success rates were estimated for the seven genera (Table 7), correctly classifying 100%, 100%, and 100% of Benthophilus, Neogobius, and Ponticola, respectively. For Proterorhinus, Hyrcanogobius, Knipowitschia, and Rhinogobius, the proportion of individuals correctly classified into their original genera were 85.7%, 77.8%, 77.8%, and 71.4%, respectively.
Correctly assigned samples are shown in bold.
Variation among the Neogobius and Ponticola species.
Twenty-seven otolith variables supported significant differences between at least two of the four Neogobius species (Table 8, S4 Fig, S4 Table). Twenty-one variables were normally distributed: five variables involved the SuL, four the SuH, three inclination angles (α, γ, and β), three shape indices (ROx, REx, and ELx), and the remainder were SuTipV/OP, OL/OH, SuA/OA, SuP/OP, OP/OL, and SuTipV/SuEndV. Six variables were non-normally distributed: δ, OP/OH, SuH/SuP, SuEndV/OP, SuP/SuTipV, and SuL/SuTipV.
Neogobius caspius diverged from N. melanostomus, N. pallasi and N. bathybius by 9, 14 and 15 variables, respectively (Table 8, S4 Fig, S4 Table). Neogobius melanostomus differed from N. pallasi and N. bathybius by 13 and 12 variables, respectively. Neogobius pallasi and N. bathybius were separated by 13 variables. Of the 27 variables, none separated all four species. SuL/OL, SuL/OP, SuTipV/OP, OP/OH, and SuEndV/OP discriminate N. caspius from the other three species. SuL/OL, SuL/SuH, and SuL/SuP showed that N. melanostomus differed from the other three species. The three shape indices ROx, REx, ELx, and OP/OH discriminated N. pallasi from the other three species. The variables α, δ, SuA/OA, SuL/OH, and SuTipV/OP differed N. bathybius from the other three species. None of the variables separated all four species.
The 27 variables were subjected to DFA; DF 1 accounted for 63.8% (eigenvalue: 43.557; λ = 0.0001, p < 0.0001) and DF 2 accounted for 21.4% (eigenvalue: 14.617; λ = 0.006, p < 0.001) of among-group variability. In order of importance, the most significant variables loadings on DF 1 and DF 2 were ROx, SuH/OL, SuL/SuH, OL/OH, SuH/OP, SuL/OH and SuH/OP, SuL/SuH, SuH/SuTipV, ROx, SuTipV/OP, OL/OH, respectively. The DFA biplot showed clear separation among three groups (Fig 8), one comprising N. caspius, the second one N. melanostomus and N. pallasi, and the third N. bathybius alone. The proportion of individuals correctly classified into their original species was 83.3%, 66.7%, 66.7%, and 100% in N. caspius, N. melanostomus, N. pallasi, and N. bathybius, respectively (Table 9).
Correctly classified samples are shown in bold.
In comparisons among Ponticola spp. and Neogobius bathybius, 23 otolith variables differed between at least two species (Table 10, S5 Fig, S4 Table). Twenty variables were normally distributed: five variables involved the SuL, two the SuH, two the SuP, two the OP, three shape indices (ROx, REx, and REx), and the remainder were α, γ, OL/OH, SuTipV/OP, SuEndV/OP, and OL2/CL. Three variables were non-normally distributed: δ, SuA/OA, and SuP/SuEndV.
The studied species of Ponticola diverged from one another by one (P. iranicus vs. P. patimari), four (P. gorlap vs. P. patimari), or five (P. gorlap vs. P. iranicus) variables (Table 10, S5 Fig, S4 Table). Neogobius bathybius significantly differed from P. gorlap, P. iranicus, and P. patimari in 17–18 otolith variables. The variables OL/OH, OP/OH, SuL/SuTipV, SuL/SuP, SuH/OL, SuH/OP, SuTipV/OP, SuEndV/OP, ROx, ELx, δ, SuA/OA, and SuP/SuEndV supported separation of N. bathybius from the three Ponticola species. None of the variables distinguished all species.
Species of Neogobius and Ponticola together were subjected to DFA; DF 1 accounted for 43.3% (eigenvalue: 59.995; λ = 0.0001, p < 0.0001) and DF 2 for 35.6% (eigenvalue: 49.303; λ = 0.0001, p < 0.0001) of among group variability. In order of importance, the most significant variables loadings on DF 1 and DF 2 were SuH/OH, SuTipV/SuEndV, SuTipV/OP, SuH/SuEndV, SuL/OH, OL/OH and SuH/OH, SuP/SuEndV, ROx, SuH/SuEndV, SuH/SuTipV, SuH/OL, respectively. The DFA biplot showed clear separation among three groups (Fig 8), one comprising Neogobius spp., the second one Ponticola spp., and the third one N. bathybius alone. Classification success rates were estimated for three Ponticola species (P. gorlap, P. iranicus, and P. patimari) (Table 9). The proportion of individuals correctly assigned to their original species was 60%, 50%, and 40%, respectively.
Variation in the Pomatoschistus and Acanthogobius lineages.
Twenty-seven otolith variables differed between at least two of the four studied species from the Pomatoschistus and Acanthogobius lineages (Table 11, S6 Fig, S4 Table). Twenty variables were normally distributed: five variables involved the SuL, four the SuH, two the SuP, three inclination angles (α, γ, and δ), and the remainder were SuTipV/OP, REx, OP/OH, SuA/OA, OL2/CL, and SuEndV/OP. Seven variables were non-normally distributed: two involved the SuL, and the remainder were SuP/OP, SuH/SuTipV, SuTipV/SuEndV, OL/OH, and ELx.
Hyrcanogobius bergi was separated from Knipowitschia caucasica, K. longecaudata, and Rhinogobius sp. by 14, 8, and 3 variables, respectively (Table 11, S6 Fig, S4 Table). Knipowitschia caucasica differed from K. longecaudata and Rhinogobius sp. by 19 and 8 variables, respectively. Knipowitschia longecaudata and Rhinogobius sp. were separated by 3 variables. None discriminated H. bergi from the other three species. The variables SuH/OH, SuH/SuEndV, SuH/OP, SuTipV/OP, SuH/SuTipV, and SuTipV/SuEndV supported K. caucasica differing from the other three species. δ and SuL/SuEndV discriminated K. longecaudata from the other three species. None of the variables supported differentiation of Rhinogobius sp. from the other three species. Therefore, none of the variables discriminated among the four species.
The 27 variables were subjected to DFA; DF 1 accounted for 67.1% (eigenvalue: 9.017; λ = 0.01, p < 0.0001) and DF 2 for 19.2% (eigenvalue: 2.574; λ = 0.098, p < 0.001) of among-group variability. In order of importance, the most significant variables loadings on DF 1 and DF 2 were SuP/SuEndV, SuH/SuEndV, SuL/SuTipV, SuH/OP, SuH/OH, SuH/SuTipV and SuH/OP, SuH/OL, SuL/SuH, SuH/OH, SuP/SuEndV, respectively. The DFA biplot is shown in Fig 8. The proportion of correct classifications to their original species was 77.8%, 42.9%, 90.9%, and 42.9% in H. bergi, K. caucasica, K. longecaudata, and Rhinogobius sp., respectively (Table 12).
Correctly classified samples are shown in bold.
Phenotypic relationships based on otolith variables.
A linkage dendrogram based on average Euclidean distances was calculated for 31 otolith variables (Fig 9), which separated the studied fish species into three groups. Group I contained the species in the Pomatoschistus and Acanthogobius lineages as well as Proterorhinus nasalis of the Gobius lineage. Group I itself included two sub-groups, with one containing species in the Pomatoschistus lineage and Acanthogobius lineage, and the second with Proterorhinus nasalis. Within the first sub-group, species of the Pomatoschistus lineage were monophyletic, sister to Rhinogobius sp. The dendrogram shows a closer phenotypic relationship of Knipowitschia longecaudata with Hyrcanogobius bergi, rather than with K. caucasica.
Yellow, Benthophilini; green, Neogobiini; pink, Ponticolini.
Group II included all benthophiline gobies except for Proterorhinus nasalis. Four sub-groups were resolved within Group II. The basal sub-group was Benthophilus leobergius and B. pinchuki. Species in the genus Neogobius were arranged in two sub-groups: one with N. pallasi, N. caspius, and N. melanostomus, and the other of N. bathybius alone. The fourth sub-group included all species in the genus Ponticola with two sub-clusters: (i) the Ponticola syrman group comprising P. syrman, and two freshwater endemic species, P. iranicus and P. patimari, and (ii) the Ponticola kessleri group comprising P. iljini, P. kessleri and P. gorlap. Neogobius bathybius occupied an intermediate position between the Neogobius and Ponticola sub-groups.
Discussion
Taxonomic significance of otolith morphology in the Caspian gobies
This study aimed to compare the otoliths of extant Caspian gobiids at different taxonomic levels based on detailed descriptions and morphometric method to reveal the taxonomic and phylogenetic information inscribed in their otoliths. The results showed that the otoliths belonging to the Gobius lineage are significantly different from those of the Pomatoschistus and Acanthogobius lineages based on general shape, characteristics of anterior and posterior rims, and otolith variables (overall, 22 otolith variables distinguished the Gobius lineage from the other two lineages). Based on morphometric otolith variables, otoliths of the Pomatoschistus and Acanthogobius lineages from the Caspian Sea basin were largely indistinguishable from each other, separated by small differences in inclination of line connecting preventral angle with tip of posterodorsal projection and inclination angle of posterior rim (see Table 2, S1 Fig, S1 Table). Otolith descriptions however, showed that the otoliths of the two lineages also differed in the shape of their dorsal rim, i.e., being more angular in the Pomatoschistus lineage. These results appear to be consistent with the phylogenetic relationships among the three lineages [28]. The overall rate of classification success at this taxonomic level was 85.4%, but 91.3% for the Gobius lineage. Within the Gobius lineage, the otoliths of the three benthophiline tribes were clearly different based on both methods, with an overall classification success of 94.2%.
The results indicated that the overall shape of the otolith is genus-specific for all Ponto-Caspian gobiid genera, except for Hyrcanogobius and Knipowitschia, making it the most efficient otolith characteristic for gobiid genus identification in the Caspian Sea basin. The monotypic genus Hyrcanogobius and Knipowitschia were separated by slight differences in just four variables, i.e., SuL/SuP, α, δ, and REx. Hyrcanogobius bergi was originally described by Iljin [50] from the river mouths of the north Caspian Sea as a separate genus Hyrcanogobius, because of the reduced condition of the head lateral-line canal system, but later suggested to be congeneric with Knipowitschia by Economidis and Miller [51]. However, according to Miller [52], the possession of transverse interorbital papillae rows and a much greater extent of anterior transverse oculoscapular row tra (extending downwards to or near the longitudinal suborbital row b vs. noticeably short of row b in Knipowitschia) appear to warrant recognition as a separate genus in any classification based on the head lateral-line system. In addition, paleontological data (otolith) by Bratishko et al. [53] suggest that Hyrcanogobius is recognizable as a separate lineage in the fossil record since 11 million years ago. Molecular data of H. bergi are still missing, however, otolith data support the reassignment of H. bergi to the genus Knipowitschia Iljin, 1927, hereby highlighting the necessity of an integrative molecular, morphological, and paleontological analysis on these species to evaluate their taxonomy.
Neogobius presently comprises five valid species: N. caspius, N. pallasi, and N. bathybius are Caspian endemics, N. fluviatilis in the Black Sea is a sister species of the Caspian N. pallasi, and N. melanostomus is native to the Ponto-Caspian basins. The four Caspian species which were well-represented in this study, can be easily distinguished from one other based on otolith morphology results from both descriptive and morphometric methods. The otolith of N. fluviatilis however, which is poorly represented in this study, is very similar to those of the Caspian N. pallasi, but a preliminary description of the N. fluviatilis otolith suggests that they differ with regard to the shape of posterodorsal projection (long and tapering in N. fluviatilis vs. short and rounded in N. pallasi) and sulcus (α 14.3° in N. fluviatilis vs. 16.0–22.5° in N. pallasi; shallow and ostial lobes weakly developed in N. fluviatilis vs. relatively deep and ostial lobes well-developed in N. pallasi). Obviously, more data on the otolith of N. fluviatilis is needed before drawing any conclusion about its otolith morphology. Gobius fluviatilis Pallas, 1814 was originally described in part from near the mouths of rivers falling into the Black Sea and similarly the Caspian Sea. Neogobius fluviatilis pallasi (Berg, 1916) was the subspecies described in the Caspian Sea basin. Kottelat and Freyhof [54] recognized N. pallasi as the Caspian Sea species and restricted N. fluviatilis to the Black Sea basin. This taxonomic decision was later confirmed by molecular data [55].
Ponticola presently comprises eight recognized species in the Caspian Sea basin, six of which were included in this study. The overall shape of otolith in these species is invariably long parallelogram, however, the otoliths of P. syrman and P. hircaniaensis are easily distinguishable. The otoliths of P. syrman and P. hircaniaensis show concavity in their dorsal rim, however, they are different in their OL/OH (1.40–1.44 vs. 1.22–1.39), OL2/CL (1.65–1.85 vs. 1.59–1.66), SuA/OA (0.09–0.11 vs. 0.13–0.16), SuL/SuH (2.25–3.07 vs. 1.47–1.82), SuH/OL (0.15–0.21 vs. 0.27–0.33), SuH/OH (0.21–0.31 vs. 0.34–0.40), SuH/SuTipV (0.64–0.81 vs. 0.9–1.08), SuH/SuEndV (0.58–0.64 vs. 0.64–0.73), SuTipV/SuEndV (0.79–0.91 vs. 0.65–0.77), ROx (0.65–0.66 vs. 0.69–0.78), and ELx (0.17–0.18 vs. 0.1–0.16) [5]. The otoliths of P. gorlap differ from those of two south Caspian freshwater endemic species, P. iranicus and P. patimari in five (γ, SuP/SuTipV, SuL/SuTipV, REx, and SuA/OA) and four (SuP/SuTipV, SuL/SuTipV, SuTipV/OP, and OL2/CL) otolith variables, respectively. The otoliths of P. gorlap also differ from those of both species with regard to the incision below the posterodorsal projection (markedly incised vs. slightly incised or not incised), the shape of the predorsal angle (broadly rounded vs. obtuse or orthogonal) and dorsal rim (anteriorly slightly depressed vs. anteriorly not depressed). The otoliths of P. gorlap are most similar to those of P. iljini, which may reflect their close phylogenetic relationships and morphological similarities. Ponticola gorlap was first identified in the Caspian Sea as Gobius kessleri Günther by Kessler [56], who found some morphological differences between the Caspian and typical Black Sea forms. Based on several morphological differences, Iljin [57] suggested that the Caspian gobies from the Mangyshlak region (western Kazakhstan) should be erected as a distinct species, which he described as Gobius gorlap. Later, karyological, cranial, head scale, and morphometric data of samples from the Dnieper, Dniester and Volga rivers provided the data supporting specific status of the Mangyshlak samples [58], and subsequently, it was described as a separate species by Vasil’eva and Vasil’ev [59] who regarded the species name gorlap as invalid and proposed the new name iljini, which later was synonymized with P. gorlap (Iljin) in a modern phylogenetic systematic study [29]. Vasil’eva et al. [38] reestablished the validity of P. iljini based on karyological data, but restricted its distribution to the coast of the Mangyshlak Peninsula, western Kazakhstan. As presently understood, P. kessleri, P. iljini, and P. gorlap are closely related and form independent phyletic lineages within a clade of Ponticola [29, 38].
The Ponticola syrman group comprises two freshwater endemic and cryptic species distinguished from each other mainly based on molecular characters and geographic distributions [7], i.e., P. iranicus endemic to the upper Sefidroud sub-basin, and P. patimari endemic to the western freshwater habitats of the south Caspian sub-basin. PCA and DFA plots for the meristic and morphometric data also showed a clear separation of the two species [7]. Our otolith morphometric variables, inclination angles, and classical shape descriptors show that the otoliths of P. iranicus and P. patimari are only slightly different in their REx. Also, the otolith shape analysis of 213 specimens representing six sub-basin samples of these species presented a high level of shape variation which did not show congruence with their taxonomy and phylogeographic structure [7]. Studies suggest that while genetics constrain the overall shape of the otolith itself, environmental conditions may eventually alter the rates of somatic and otolith growth, which in turn may affect otolith shape.
Tadpole-gobies of the genus Benthophilus are a group of 21 poorly known species from the fresh and brackish waters of the Caspian and Black Sea basins, including the Sea of Azov [37]. Boldyrev and Bogutskaya [41] recognized 20 species and assigned them to four phenotypic groups (i.e., I, II, III, and IV), based on differences in size, arrangement and counts of dermal ossifications, fin ray counts, and body shape. However, the phylogenetic integrity of these groups has never been tested, since the phylogeny of the genus is poorly known and genetic data are available for only a few species [29, 30]. The phylogenetic integrity of the four phenotypic groups established by Boldyrev & Bogutskaya [41] was questioned by Kovačić et al. [37], since a newly described species B. persicus Kovačić, Esmaeili, Zarei, Abbasi & Schliewen, 2021 featured a mix of characters of phenotypic groups II and III. Interestingly though, the phylogenetic inferences of Neilson and Stepien [29] and Zarei et al. [30] estimated a closer relationship between a group II member (a specimen identified as B. abdurahmanovi) and the group I member (a specimen identified as B. granulosus Kessler, 1877), than to three other group II species used in their analyses (identified as B. mahmudbejovi Ragimov, 1976, B. stellatus and B. leobergius). Our otolith data also question the phylogenetic integrity of the four phenotypic groups defined by Boldyrev & Bogutskaya [41]: the otoliths of the species examined here comprised four phenotypic groups based on their overall shapes, which did not show congruence to their hypothesis: group 1: B. leobergius (II); group 2: B. pinchuki (III), B. microcephalus (II), B. stellatus (II), B. baeri (IV); group 3: B. durreli (II); and group 4: B. abdurahmanovi (II). However, we consider our results preliminary and await substantially increased taxon sampling and a thorough phylogenetic analysis of the species.
Otolith data suggest the monophyly of neogobiin gobies
Molecular phylogenies have supported a monophyletic clade comprising the neogobiin and benthophilin gobies [29, 60]. This monophyletic clade contains three distinctive sub-clades designated as tribes Benthophilini, Neogobiini, and Ponticolini. The phylogenetic placement of Benthophilini has been inconsistent among analyses. The two combined mito-nuclear analyses resolve Benthophilini as the sister clade to Ponticolini [29, 60], however, this relationship had mixed support from different analysis methods, represented a short internal branch, having low internode certainty and gene support frequency. In the cyt b analysis of Neilson and Stepien [29], Benthophilini was the sister clade to Neogobiini, yet in their COI analysis and also the cyt b analysis of Medvedev et al. [61], Benthophilini again grouped with Ponticolini. The COI analysis of Zarei et al. [30] resolved Benthophilini as the sister clade to Ponticolini + Neogobiini (Fig 10). The latter phylogenetic hypothesis was also supported by our otolith data: Benthophilini differed significantly from Neogobiini and Ponticolini by 25 and 21 variables, respectively, whereas Neogobiini and Ponticolini were separated by 17 variables. The average linkage dendrogram based on the Euclidean distance for the otolith variables also clustered Benthophilini as the sister clade of Neogobiini + Ponticola. This phylogenetic hypothesis also agrees with Schwarzhans et al. [12] based on articulated skeletons and otoliths of fossil gobiids: (i) both the neogobiin and benthophilin subgroups were represented by ‘‘primitive” extinct genera (i.e., †Proneogobius and †Protobenthophilus) considered to be the sister group to all extant members of their respective subgroups, and (ii) the origin and separation of the two subgroups likely links to the segregation of the Eastern Paratethys during the Langhian stage (16–13.6 Ma) of the middle Miocene. Nevertheless, additional genomic, morphological, and taxonomic sampling are needed to further resolve relationships among the major Ponto-Caspian endemic gobiid lineages.
Each tribe is represented by a different color: yellow, Benthophilini; green, Neogobiini; pink, Ponticolini. BI posterior probability/ML bootstrap support values are indicated beside the nodes. Double bar at the root indicates that branch has been reduced in length and is not proportional to the scale. Alphabetical letters inside parentheses after the species names (A–Q: 17 species) refer to their otolith drawings on the right. Otolith drawings on the upper left corner belong to Ponticola iljini (R–S), Mesogobius nonultimus (T), Benthophilus pinchuki (U), B. macrocephalus (V), B. durrelli (W), B. baeri (X), Anatirostrum profundorum (Y), and Benthophiloides brauneri (Z), which have not been phylogenetically analysed.
Systematics of Neogobius bathybius
The taxonomic status of Neogobius bathybius has been controversial. Gobius bathybius Kessler (1877) originally was described from the Svinoi Island, Caspian Sea. The name Chasar appeared in print for the first time in Berg [62] as a subgenus of Neogobius to accommodate bathybius. Berg [62] provided a brief description of the species as Neogobius (Chasar) bathybius, but did not define the genus-group category. Although Vasil’eva [63] stated that Iljin [64, 65] used this subgeneric name for bathybius, a search of the latter publications by Miller [66] found this species mentioned only as being incertae sedis, but without reference to any previous use of the name Chasar or to a definition by Iljin. Both Iljin [57] and Ragimov [67] used the name at a subgeneric level, but again provided no diagnosis. Pinchuk and Ragimov [68], in their redescription of bathybius, placed this species in Neogobius without a subgenus. The first subgeneric diagnosis of Chasar, with indication of the type and only species, thus appears to be by Vasil’eva [63]. Detailed osteological comparison with other gobiid taxa by Vasil’eva [63] suggested that bathybius occupied a distinct subgeneric position. The monotypic genus Chasar was recognized as a valid taxon by Miller [66], on the basis of the head sensory papillae patterns noted by Pinchuk and Ragimov [68] and Vasil’eva [63]. The resulting paraphyly of Neogobius sensu lato [62] however was changed in Neilson and Stepien’s [29] revised classification, by elevating two of Iljin’s [64] subgenera to genus rank, i.e., Babka and Ponticola for the remainder of the ‘neogobiin’ species. Neilson and Stepien [29] included bathybius in Ponticola in their study without further justification, since they lacked bathybius specimens to sequence and did not hypothesize a nominal genus.
Recent phylogenetic analyses by Zarei et al. [30] and Tajbakhsh et al. [69] provided support for the reclassification of Gobius bathybius from Ponticola to Neogobius sensu stricto [29]. Neogobius bathybius however, differs from the other four Neogobius species in its cheek sensory papillae pattern by possessing one additional transverse row before row b (i.e., five transverse suborbital rows before row b; Fig 11) [68]. The presence of eight transverse suborbital rows and five before row b might be interpreted as a synapomorphy with Mesogobius, but N. bathybius possesses two transverse rows below row b, a plesiomorphic feature shared with Neogobius, Ponticola (except for P. syrman that has three rows), and Proterorhinus; this differs from Mesogobius, which has three rows. The suborbital lateral line system pattern of Mesogobius and N. bathybius do not completely match and might be attributed to parallel evolution or to two step development, where the first step was the synapomorphy of an additional transverse suborbital row in front of suborbital row b, and the second step was the addition of one more transverse suborbital row below row b. In terms of otolith morphology, the otoliths of all Neogobius species except for bathybius are characterized by having a square-rhomboid (N. caspius and N. melanostomus) to a discoid-rhomboid (N. fluviatilis and N. pallasi) shape, a convex and regularly curved dorsal rim, a posterodorsal projection that is usually long and narrow, tapering or rounded, and strongly bent outwards, δ 24.3–33.4°, α 14.7–22.5°, and presence of subcaudal iugum and dorsal depression. On the other hand, the N. bathybius otolith is characterized by having a two-humped long rectangle shape, a highly positioned posterior hump, a dorsal rim with a deep broad V-shaped concavity, a posterodorsal projection that is bulky and very broad and does not bend outwards, δ 13.5–22.1°, α 2.1–10.6°, and absence of subcaudal iugum and dorsal depression. Among the otoliths of all Ponto-Caspian endemic gobiids studied here, a dorsal rim with two angular humps (with the posterior hump being highly positioned) and a deep broad V-shaped concavity, and a bulky and very broad posterodorsal projection are found in only one other species, Mesogobius nunultimus. Zarei et al. [30] suggest that as an alternative to the mitochondrial phylogenetic hypothesis [30, 69], an ancient hybridization scenario between Neogobius melanostomus and Mesogobius nunultimus might have led to the same phylogenetic pattern as the hypothesized sister group relationship between bathybius and N. melanostomus. In addition to the suborbital papillae pattern and the otolith shape, N. bathybius shows several other intermediate morphological and ecological characteristics (Table 13). However, its intermediate characters (best exemplified by the intermediate number of vertebrae, second dorsal and anal fin elements, longitudinal scale rows, head, body and caudal peduncle depths, longevity, and migration) and those completely different from the characters of the putative parental species (e.g., first dorsal elements, growth, habitat, depth, and spawning period; Table 13) could be the result of hybridization and later isolating evolution, as well as the outcome of other scenarios. Therefore, we suggest a chromosome analysis and genomic approach to resolve the generic classification of Neogobini and Ponticolini.
For comparison, the head lateral line system and otolith drawings of M. nonultimus, N. bathybius, and N. melanostomus are shown in the bottom row.
Phylogenetic placement of Anatirostrum and Benthophiloides
The morphologically well-defined benthophilin group is characterized by several apomorphies: (i) suborbital row 5i longer than row 6i, extending below rear termination of row d, with the row 6i separated from row d, (ii) suborbital row 4 not ascending above the level of row b, (iii) interorbital papillae present, (iv) loss of all head canals, (v) a tubular anterior nostril without a rim process, and (vii) absence or, at least, reduction in scale cover, with no squamation on head. Hitherto, the two phylogenetically studied benthophilin genera, Benthophilus and Caspiosoma, formed a monophyletic clade referred to by Neilson and Stepien [29] as Benthophilini. The otoliths of the benthophilin group are characterized by a reduced, shallow sulcus that lacks well-developed ostial lobes and has no subcaudal iugum. Each of the four benthophilin genera possess a distinct otolith shape, reflecting their generic classification and phylogenetic affinities: long elliptical in Benthophilus, right trapezoid in Anatirostrum, long rectangle in Benthophiloides, and pentagonal in Caspiosoma.
Caspiosoma caspium, having the most distinctive otolith shape among the four benthophilin genera, has a basal phylogenetic placement within Benthophilini [29]. Absence of scales or scale-like derivatives in Caspiosoma is a reductive feature shared with Benthophiloides; however, the otoliths of the two genera differ chiefly in the shape of dorsal rim above the cauda (angular vs. convex and gently curved, highest above cauda) and the inclination of ostium (inclined at 7.5° vs. not inclined). Also, a significant number of morphological features link Benthophilus with Anatirostrum [34], but Anatirostrum lacks the chin barbel and dermal fold at the jaw angle of Benthophilus as well as the large tubercles seen in Benthophilus. Within the benthophilin group, Anatirostrum shows a number of autapomorphies including an additional suborbital row below level of suborbital longitudinal row b, an elongated and duck-bill shaped snout with posterior nostrils displaced well anterior to the orbit, and a complete sequence of pterygiophores between the first and second dorsal fins. Furthermore, the otoliths of Anatirostrum and Benthophilus differ mainly in the dorsal rim’s curvature (horizontal to slightly concave vs. convex and gently curving), and the shape of the predorsal angle (orthogonal and slightly raised vs. obtuse and markedly depressed). Molecular data of Benthophiloides and Anatirostrum are still missing, however, these mosaic patterns of morphological features indicate relatively divergent but sister group relationships between Caspiosoma and Benthophibides, and between Benthophilus and Anatirostrum.
Conclusion
The results indicated high taxonomic efficiency of otolith morphology combined with morphometry at different taxonomic levels for the Ponto-Caspian gobiids. Separation of otoliths at different taxonomic levels requires consideration of different morphological characters and otolith variables [46]. This was also the case in our study. In addition, it also appears that these qualitative and quantitative otolith data contain important phylogenetic signals, however, more studies are needed to complete these evaluations and confirm our otolith study findings.
Supporting information
S1 Checklist. Ethical, cultural, and scientific considerations specific to inclusivity in global research.
https://doi.org/10.1371/journal.pone.0285857.s001
(DOCX)
S1 Fig. Box plots of 28 otolith variables that were useful in the separation of the three lineages.
G.L., Gobius lineage; P.L., Pomatoschistus lineage; A.L., Acanthogobius lineage.
https://doi.org/10.1371/journal.pone.0285857.s002
(TIF)
S2 Fig. Box plots of 29 otolith variables that were useful in the separation of the three benthopheline tribes.
Green, normally distributed variables; blue, non-normally distributed variables.
https://doi.org/10.1371/journal.pone.0285857.s003
(TIF)
S3 Fig. Box plots of 31 otolith variables that were useful in the separation of the seven genera.
Be., Benthophilus; Ne., Neogobius; Po., Ponticola; Pr., Proterorhinus; Hy., Hyrcanogobius; Kn., Knipowitschia; Rh., Rhinogobius. Green, normally distributed variables; blue, non-normally distributed variables.
https://doi.org/10.1371/journal.pone.0285857.s004
(TIF)
S4 Fig. Box plots of 27 otolith variables that were useful in the separation of the studied species in the genus Negobius.
N. c., Neogobius caspius; N. m., Neogobius melanostomus; N. p., Neogobius pallasi; N. b., Neogobius bathybius. Green, normally distributed variables; blue, non-normally distributed variables.
https://doi.org/10.1371/journal.pone.0285857.s005
(TIF)
S5 Fig. Box plots of 23 otolith variables that were useful in the separation of the studied species of Ponticola and Negobius bathybius (N. b.).
P. g., Ponticola gorlap; P. i., Ponticola iranicus; P. p., Ponticola patimari. Green, normally distributed variables; blue, non-normally distributed variables.
https://doi.org/10.1371/journal.pone.0285857.s006
(TIF)
S6 Fig. Box plots of 27 otolith variables that were useful in the separation of the studied species in the Pomatoschistus and Acanthogobius lineages.
H. b., Hyrcanogobius bergi; K. c., Knipowitschia caucasica; K. l., Knipowitschia longecaudata; R. sp., Rhinogobius sp. Green, normally distributed variables; blue, non-normally distributed variables.
https://doi.org/10.1371/journal.pone.0285857.s007
(TIF)
S1 Table. Calculated otolith variables for the three gobiid lineages.
https://doi.org/10.1371/journal.pone.0285857.s008
(XLSX)
S2 Table. Calculated otolith variables for the three benthopheline tribes.
https://doi.org/10.1371/journal.pone.0285857.s009
(XLSX)
S3 Table. Calculated otolith variables for the seven gobiid genera.
https://doi.org/10.1371/journal.pone.0285857.s010
(XLSX)
S4 Table. Calculated otolith variables for the studied species.
https://doi.org/10.1371/journal.pone.0285857.s011
(XLSX)
Acknowledgments
Samples of Mesogobius batrachocephalus and Proterorhinus semilunaris from the Black Sea basin were made available by Werner Schwarzhans (Natural History Museum of Denmark) and Ulrich Schliewen (SNSB-Zoologische Staatssammlung München), respectively. Our thanks are attributed to Werner Schwarzhans, Ekaterina Vasil’eva (Lomonosov Moscow State University) and Kristiaan Hoedemakers (Royal Belgian Institute of Natural Sciences), who graciously allowed us to recycle and use the otolith photographs of several species (see Table 1) from their previous publications [12, 38, 43–45]. We thank Werner Schwarzhans and an anonymous reviewer for their constructive comments and suggestions on this work.
References
- 1. Schulz-Mirbach T, Ladich F, Plath M, Heß M. Enigmatic ear stones: what we know about the functional role and evolution of fish otoliths. Biological Reviews. 2019; 94(2):457–482. pmid:30239135
- 2. Teimori A, Esmaeili HR, Hamidan N, Reichenbacher B. Systematics and historical biogeography of the Aphanius dispar species group (Teleostei: Aphaniidae) and description of a new species from Southern Iran. Journal of Zoological Systematics and Evolutionary Research. 2018; 56(4):579–598.
- 3. Teimori A, Khajooei A, Motamedi M, Hesni MA. Characteristics of sagittae morphology in sixteen marine fish species collected from the Persian Gulf: demonstration of the phylogenetic influence on otolith shape. Regional Studies in Marine Science. 2019; 29:100661.
- 4. Tuset VM, Farré M, Otero-Ferrer JL, Vilar A, Morales-Nin B, Lombarte A. Testing otolith morphology for measuring marine fish biodiversity. Marine and Freshwater Research. 2016; 67(7):1037–1048.
- 5. Zarei F, Esmaeili HR, Kovai M, Schliewen UK, Abbasi K. Ponticola hircaniaensis sp. nov., a new and critically endangered gobiid species (Teleostei: Gobiidae) from the southern Caspian Sea basin. Zootaxa. 2022; 5154(4):401–430. pmid:36095612
- 6. Zarei F, Esmaeili HR, Sadeghi R, Reichenbacher B, Schliewen UK, Abbasi K, et al. Phylogeography and population structure of Ponticola gorlap (Teleostei: Gobiidae) in an evolutionary distinctive and ecologically threatened Caspian Sea sub‑basin. Aquatic Sciences. 2023; 85.
- 7. Zarei F, Esmaeili HR, Sadeghi R, Schliewen UK, Kovačić M, Abbasi K, et al. An integrative insight into the diversity, distribution, and biogeography of the freshwater endemic clade of the Ponticola syrman group (Teleostei: Gobiidae) in the Caucasus biodiversity hotspot. Ecology and Evolution. 2022; 12(9):e9300. pmid:36177146
- 8. Andrus CFT, Crowe DE, Romanek CS. Oxygen isotope record of the 1997–1998 El Nino in Peruvian sea catfish (Galeichthys peruvianus) otoliths. Paleoceanography and Paleoclimatology. 2002; 17(4):1053.
- 9. Bratishko A, Schwarzhans W, Reichenbacher B, Vernyhorova Y, Ćorić S. Fish otoliths from the Konkian (Miocene, early Serravallian) of Mangyshlak (Kazakhstan): testimony to an early endemic evolution in the Eastern Paratethys. Paläontologische Zeitschrift. 2015; 89(4):839–889.
- 10.
Nolf D. Handbook of paleoichthyology. Vol. 10, otolithi piscium. München: Verlag Dr. Friedrich Pfeil; 1985.
- 11. Nolf D. Studies on fossil otoliths—the state of the art. Recent Developments in Fish Otolith Research. 1995; 19:513–544.
- 12. Schwarzhans W, Ahnelt H, Carnevale G, Japundžić S, Bradić K, Bratishko A. Otoliths in situ from Sarmatian (Middle Miocene) fishes of the Paratethys. Part III: tales from the cradle of the Ponto-Caspian gobies. Swiss Journal of Palaeontology. 2017; 136(1):45–92.
- 13. Van Neer W, Ervynck A, Bolle LJ, Millner RS, Rijnsdorp AD. Fish otoliths and their relevance to archaeology: an analysis of medieval, post-medieval, and recent material of plaice, cod and haddock from the North Sea. Environmental Archaeology. 2002; 7(1):61–76.
- 14. Avigliano E, Domanico A, Sánchez S, Volpedo AV. Otolith elemental fingerprint and scale and otolith morphometry in Prochilodus lineatus provide identification of natal nurseries. Fisheries Research. 2017; 186:1–10.
- 15. Sadeghi R, Esmaeili HR, Zarei F, Reichenbacher B. Population structure of the ornate goby, Istigobius ornatus (Teleostei: Gobiidae), in the Persian Gulf and Oman Sea as determined by otolith shape variation using ShapeR. Environmental Biology of Fishes. 2020; 103(10):1217–1230.
- 16. Mehraban H, Esmaeili HR, Gholamhosseini A, Seifali M. Hidden taxonomic characters in otoliths of blenniid fishes (Teleostei: Blenniidae) from the Iranian coasts of the Persian Gulf and Oman Sea ecoregions: A scanning electron microscopy approach. Acta Zoologica. 2022.
- 17.
Smale MJ, Watson G, Hecht T. Otolith atlas of southern African marine fishes: JLB Smith Institute of Ichthyology; 1995.
- 18. Tuset VM, Lombarte A, Assis CA. Otolith atlas for the western Mediterranean, north and central eastern Atlantic. Scientia Marina. 2008; 72(S1):7–198.
- 19. Berg F, Almeland OW, Skadal J, Slotte A, Andersson L, Folkvord A. Genetic factors have a major effect on growth, number of vertebrae and otolith shape in Atlantic herring (Clupea harengus). PLoS One. 2018; 13(1):e0190995. pmid:29324892
- 20. Cardinale M, Doering-Arjes P, Kastowsky M, Mosegaard H. Effects of sex, stock, and environment on the shape of known-age Atlantic cod (Gadus morhua) otoliths. Canadian Journal of Fisheries and Aquatic Sciences. 2004; 61(2):158–167.
- 21. Söllner C, Burghammer M, Busch-Nentwich E, Berger J, Schwarz H, Riekel C, et al. Control of crystal size and lattice formation by starmaker in otolith biomineralization. Science. 2003; 302(5643):282–286. pmid:14551434
- 22. Campana SE, Thorrold SR. Otoliths, increments, and elements: keys to a comprehensive understanding of fish populations? Canadian Journal of Fisheries and Aquatic Sciences. 2001; 58(1):30–38.
- 23. Di Franco A, Calò A, Sdiri K, Cattano C, Milazzo M, Guidetti P. Ocean acidification affects somatic and otolith growth relationship in fish: evidence from an in situ study. Biology Letters. 2019; 15(2):20180662. pmid:30958127
- 24. Hong P, Jiang S, Katayama S. Experimental study on the influence of water temperature on the otolith formation of the marbled flounder (Pseudopleuronectes yokohamae). International Aquatic Research. 2021; 13(2):119–128.
- 25. Oozeki Y, Watanabe Y. Comparison of somatic growth and otolith increment growth in laboratory-reared larvae of Pacific saury, Cololabis saira, under different temperature conditions. Marine Biology. 2000; 136(2):349–359.
- 26. Vignon M. Short-term stress for long-lasting otolith morphology—brief embryological stress disturbance can reorient otolith ontogenetic trajectory. Canadian Journal of Fisheries and Aquatic Sciences. 2018; 75(10):1713–1722.
- 27. Gill AC, Mooi RD. Thalasseleotrididae, new family of marine gobioid fishes from New Zealand and temperate Australia, with a revised definition of its sister taxon, the Gobiidae (Teleostei: Acanthomorpha). Zootaxa. 2012; 3266(1):41–52.
- 28. Agorreta A, San Mauro D, Schliewen U, Van Tassell JL, Kovačić M, Zardoya R, et al. Molecular phylogenetics of Gobioidei and phylogenetic placement of European gobies. Molecular Phylogenetics and Evolution. 2013; 69(3):619–633. pmid:23911892
- 29. Neilson ME, Stepien CA. Escape from the Ponto-Caspian: Evolution and biogeography of an endemic goby species flock (Benthophilinae: Gobiidae: Teleostei). Molecular Phylogenetics and Evolution. 2009; 52(1):84–102. pmid:19475709
- 30. Zarei F, Esmaeili HR, Schliewen UK, Abbasi K, Sayyadzadeh G. Mitochondrial phylogeny, diversity, and ichthyogeography of gobies (Teleostei: Gobiidae) from the oldest and deepest Caspian sub-basin and tracing source and spread pattern of an introduced Rhinogobius species at the tricontinental crossroad. Hydrobiologia. 2021; 848(6):1267–1293.
- 31. Zarei F, Esmaeili HR, Abbasi K, Kovačić M, Schliewen UK, Stepien CA. Gobies (Teleostei: Gobiidae) of the oldest and deepest Caspian Sea sub-basin: an evidence-based annotated checklist and a key for species identification. Zootaxa. 2022; 5190(2):151–193. pmid:37045174
- 32. Boldyrev VS, Bogutskaya NG. Description of two new species of tadpole-gobies (Teleostei: Gobiidae: Benthophilus). Zoosystematica Rossica. 2004; 13(1):129–135.
- 33.
Miller PJ. The freshwater fishes of Europe, Vol. 8/I Mugilidae, Atherinidae, Atherinopsidae, Blenniidae, Odontobutidae, Gobiidae 1. Berlin: AULA-Verlag GmbH Wiebelsheim, Verlag fur Wissenschaft und Forschung; 2003.
- 34.
Miller PJ. The freshwater fishes of Europe Vol. 8/II Gobiidae 2. Berlin: AULA-Verlag GmbH Wiebelsheim, Verlag fur Wissenschaft und Forschung; 2004.
- 35. Zarei F, Esmaeili HR, Schliewen UK, Abbasi K. Taxonomic diversity and distribution of the genus Proterorhinus (Teleostei: Gobiidae) in the Caucasus biodiversity hotspot with conservation implications. Aquatic Conservation: Marine and Freshwater Ecosystems. 2022; 32(1):129–138.
- 36. Eagderi S, Nikmehr N, Poorbagher H. Ponticola patimari sp. nov.(Gobiiformes: Gobiidae) from the southern Caspian Sea basin, Iran. FishTaxa. 2020; 17:22–31.
- 37. Kovačić M, Esmaeili HR, Zarei F, Abbasi K, Schliewen UK. A new species of tadpole-goby, Benthophilus persicus sp. nov. (Teleostei: Gobiidae) from the southern Caspian Sea. Zootaxa. 2021; 4980(1):45–63. pmid:34186992
- 38. Vasil’eva ED, Schwarzhans W, Medvedev DA, Vasil’ev VP. Cryptic species of Ponto-Caspian bighead goby of the genus Ponticola (Gobiidae). Journal of Ichthyology. 2016; 56(1):1–18.
- 39. Vasil’eva ED, Mousavi-Sabet H, Vasil’ev VP. Ponticola iranicus sp. nov. (Actinopterygii: Perciformes: Gobiidae) from the Caspian Sea basin. Acta Ichthyologica et Piscatoria. 2015; 45(2):189–197.
- 40. Davoodi P. Characteristics of sagittae morphology in seven goby species from the Caspian Sea. Iranian Journal of Ichthyology. 2021; 8(3):170–180.
- 41. Boldyrev VS, Bogutskaya NG. Revision of the tadpole-gobies of the genus Benthophilus (Teleostei: Gobiidae). Ichthyological Exploration of Freshwaters. 2007; 18(1):55–120.
- 42. Esmaeili HR, Zarei F, Vahed NS, Masoudi M. Scale morphology and phylogenetic character mapping of scale-surface microstructures in sixteen Aphanius species (Teleostei: Aphaniidae). Micron. 2019; 119:39–53. pmid:30660090
- 43. Jacobs P, Hoedemakers K. The round goby Neogobius melanostomus (Pallas, 1814)(Perciformes: Gobiidae), an invasive species in the Albert Canal (Belgium). Belgian Journal of Zoology. 2013; 143(2):148–153.
- 44. Bratishko A, Kovalchuk O, Scwarzhans W. Bessarabian (Tortonian, Late Miocene) fish otoliths from a transitional freshwater-brackish environment of Mykhailivka, Southern Ukraine. Palaeontologia Electronica. 2017; 20(3):1–13.
- 45. Schwarzhans W, Agiadi K, Carnevale G. Late Miocene–Early Pliocene evolution of Mediterranean gobies and their environmental and biogeographic significance. Rivista Italiana di Paleontologia e Stratigrafia. 2020; 126(3):657–724.
- 46. Gierl C, Liebl D, Sanda R, Vukic J, Esmaeili HR, Reichenbacher B. What can goby otolith morphology tell us? Cybium. 2018; 42(4):349–363.
- 47. Ponton D. Is geometric morphometrics efficient for comparing otolith shape of different fish species? Journal of Morphology. 2006; 267(6):750–757. pmid:16526058
- 48.
IBM Corp. R. IBM SPSS statistics for windows, version 26.0. NY: IBM Corp.; 2019.
- 49.
Nolf D. The diversity of fish otoliths, past and present. Brussels: Koninklijk Belgisch Instituut voor Natuurwetenschappen; 2013.
- 50. Iljin BS. Two new genera and a new species of Gobiidae from the Caspian sea. Reports of the Astrakhan Scientific Fishery Station. 1928; 6(3):1–14.
- 51. Economidis PS, Miller PJ. Systematics of freshwater gobies from Greece (Teleostei: Gobiidae). Journal of Zoology. 1990; 221(1):125–170.
- 52.
Miller PJ. Knipowitschia Iljin, 1927. In: Miller PJ, editor. The freshwater fishes of Europe Vol 8/II Gobiidae 2. Berlin: AULA-Verlag GmbH Wiebelsheim, Verlag fur Wissenschaft und Forschung; 2004. p. 331–337.
- 53. Bratishko A, Schwarzhans W, Vernyhorova Y. The endemic marine fish fauna from the eastern Paratethys reconstructed from otoliths from the Miocene (Middle Sarmatian S.L.; Bessarabian) of Jurkine (Kerch Peninsula, Crimea). Rivista, Italiana di Paleontologiia e Stratigrafia. 2023; 129(1):111–183.
- 54.
Kottelat M, Freyhof J. Handbook of European freshwater fishes. Cornol, Switzerland, and Berlin, Germany: Kottelat and Freyhof; 2007.
- 55. Neilson ME, Stepien CA. Historic speciation and recent colonization of Eurasian monkey gobies (Neogobius fluviatilis and N. pallasi) revealed by DNA sequences, microsatellites, and morphology. Diversity and Distributions. 2011; 17(4):688–702.
- 56. Kessler KF. A description of fishes belonging to the families’ common to both the Black and the Caspian seas. Trudy St Petersburg Obsc Estestvoisp. 1874; 5:191–324.
- 57.
Iljin BS. Remarks and corrections to suborder Gobioidei in the book of LS Berg “Freshwater fishes of the USSR and neighbouring countries”, edit. 4, 1948–1949. Voprosy Ichtiologii. 1956; 7:85–192.
- 58. Vasil’ev VP, Vasil’eva ED. Karyological evidence of the separate species status of Neogobius kessleri (Günther) and Neogobius gorlap Iljin (Pisces, Gobiidae). Proceedings of the USSR Academy of Sciences. 1992; 324:898–900.
- 59. Vasil’eva ED, Vasil’ev VP. The description of Neogobius iljini sp. nov. within former N. kessleri (Gobiidae, Pisces). Acta-Universitatis Carolinae Biologica. 1995; 39:261–270.
- 60. McCraney WT, Thacker CE, Alfaro ME. Supermatrix phylogeny resolves goby lineages and reveals unstable root of Gobiaria. Molecular Phylogenetics and Evolution. 2020; 151:106862. pmid:32473335
- 61. Medvedev DA, Sorokin PA, Vasil’ev VP, Chernova NV, Vasil’eva ED. Reconstruction of phylogenetic relations of Ponto-Caspian gobies (Gobiidae, Perciformes) based on mitochondrial genome variation and some problems of their taxonomy. Journal of Ichthyology. 2013; 53(9):702–712.
- 62.
Berg LS. Freshwater fishes of the USSR and adjacent countries. Jerusalem: Israel Program for Scientific Translations; 1949.
- 63. Vasil’eva ED. Skull morphology of the deep-sea goby Gobius bathybius Kessler in connection with its systematic position within the genus Gobius sensu lato (Gobiidae). Voprosy Ikhtyologii. 1996; 36:448–453.
- 64. Iljin BS. A guide to the gobies (family Gobiidae) of the Azov and Black seas. Trudy Azovo-Chernomorsky Nauchno Ekspeditsii. 1927; 2:128–143.
- 65.
Iljin BS. Le système des Gobiidés. Instituto Español de Oceanografía: Ministerio de Fomento; 1930.
- 66.
Miller PJ. Chasar Vasil’eva, 1996. In: Miller PJ, editor. The freshwater fishes of Europe Vol 8/II Gobiidae 2. Berlin: AULA-Verlag GmbH Wiebelsheim, Verlag fur Wissenschaft und Forschung; 2004. p. 94–96.
- 67.
Ragimov DB. On the systematics of fishes belonging to the genus Gobius in the Caspian Sea. In: Khalilov RA, editor. Biological productivity of the Kurinsk-Caspian fishing region: Trudy Zoologicheskogo Instituta Akademii Nauk SSSR; 1967. p. 252–277.
- 68. Pinchuk VI, Ragimov DB. The lateral line system in two endemic species of gobies of the Caspian Sea. Zoologicheskii Zhurnal. 1985; 64:562–567.
- 69. Tajbakhsh F, Rajabi-Maham H, Abdoli A, Stepien CA, Kiabi BH. DNA sequence support for reclassification of the endemic southern Caspian Sea deepwater goby as Neogobius bathybius (Formerly Ponticola; Perciformes: Gobiidae) and recent population expansion of a continuous population. Ichthyology & Herpetology. 2022; 110(1):13–21.
- 70.
Pinchuk VI, Miller PJ. Mesogobius nonultimus (Iljin, 1936). In: Miller PJ, editor. The freshwater fishes of Europe Vol 8/II Gobiidae 2. Berlin: AULA-Verlag GmbH Wiebelsheim, Verlag fur Wissenschaft und Forschung; 2004. p. 132–137.
- 71.
Pinchuk VI, Vasil’eva ED, Miller PJ. Chasar bathybius (Kessler, 1877). In: Miller PJ, editor. The freshwater fishes of Europe Vol 8/II Gobiidae 2. Berlin: AULA-Verlag GmbH Wiebelsheim, Verlag fur Wissenschaft und Forschung; 2004. p. 97–105.
- 72.
Pinchuk VI, Vasil’eva ED, Vasil’ev VP, Miller PJ. Neogobius melanostomus (Pallas, 1814). In: Mille PJ, editor. The freshwater fishes of Europe Vol 8/II Gobiidae 2. Berlin: AULA-Verlag GmbH Wiebelsheim, Verlag fur Wissenschaft und Forschung; 2003. p. 293–345.