Phoxinus sp.

Bogutskaya, Nina G., Diripasko, Oleg A. & Palandačić, Anja, 2023, Novel data support validity of Phoxinus chrysoprasius (Pallas, 1814) (Actinopterygii, Leuciscidae), European Journal of Taxonomy 861, pp. 1-20 : 6-8

Phoxinus sp. Kuban’, Clade 19

UKRAINE • 6 specs, 42.2–48.7 mm SL; Adagum River at Krymsk, Kuban’ drainage; 44°54′03″ N, 37°58′31″ E; 23 Jul. 2001; N. Bogutskaya, A. Naseka and J. Freyhof leg.; ZFMK 79044–49 .

Genetic analysis

A single haplotype was detected in all five examined Crimean specimens (Supp. file 2), previously recognised as genetic clade 20 from the Salhir drainage ( Palandačić et al. 2017, 2020). From the CO1 mitochondrial gene sequences, this haplotype forms its own haplogroup in the network ( Fig. 2 ) and is 33–97 mutational steps distant from mitochondrial clades analysed in this study, supporting its distinctiveness. Pairwise distances between P. chrysoprasius and the other clades or species range between 4% and 7% (Supp. file 5).

Statistical morphological analysis

Samples were first compared using only non-morphometric characters to exclude the influence of sexual dimorphism and be able to include the P. marsilii samples for which we did not have a complete set of morphometric data because of variable preservation condition of the specimens. The number of specimens per sample with SL> 45 mm that are sub-adult and adult (314 specimens in total), and a list of characters (11 counts and 2 coded characters represented by 97 character states) are reported in Supp. file 4. DFA ( Fig. 3 ) did not reveal a clear differentiation of the groups. We found that only 63.7% of the specimens were classified correctly. However, the Crimean specimens were differentiated at a relatively high level with 26 out of 29 specimens correctly classified. Characters that contributed most to the discrimination between the groups were the relative length of the first uninterrupted section of the lateral line, the total number of vertebrae and the numbers of abdominal and caudal vertebrae.

DFA was then performed using all characters (meristic, coded and morphometric) ( Fig. 4 ) in the samples geographically closest to the Black Sea (Bulgarian P. strandjae, Turkish P. strandjae , P. colchicus and Crimean) for females and males separately, as the species exhibits considerable sexual dimorphism as described below. All specimens were classified successfully (100%), and both males and females of P. chrysoprasius were clearly distant from the respective subsamples of P. colchicus and P. strandjae ( Fig. 4A ). Total number of lateral-line scales contributed most to the discrimination between samples, followed by the caudal-peduncle length (% SL), caudal-peduncle depth (both % of caudal-peduncle length and % SL), prepelvic length (% SL) and total number of scales in the lateral series.

As some of the result might be influenced by differences in the number of females and males (only two males in Turkish P. strandjae and absent in P. colchicus ), we undertook an analysis for females only ( Fig. 4B ). All female specimens were 100% classified. The subsample of Crimean females was almost equidistant from P. colchicus and Bulgarian P. strandjae females (squared Mahalanobis distance 568.445 and 609.09, respectively), but less so (419.83) from the Turkish P. strandjae subsample.

In combination, CO1 and the morphological data of the present study strongly support the validity of the native Crimean Phoxinus under the earliest available name of the species: P. chrysoprasius .

Figures

Fig. 2. A haplotype network based on cytochrome oxidase I (CO1) fragment using 112 previously published GenBank sequences and representing 20 genetic clades numbered as in Palandačić et al. (2017, 2020), of which 12 are considered valid species including P. chrysoprasius (Pallas, 1814) and Kuban’ Phoxinus (an available species name not known). The lines carry the number of mutational steps (shown in red).

Fig. 3. DFA based on 11 counts and 2 coded characters as in Supp. file 4. Clade 5 = 5a, Danubian tributaries in Bulgaria (Nishava, Beli Vit, and Palakaria samples and P. csikii Hankó, 1922 from type locality). Clade 14 = non-Danubian rivers of the Black Sea coast in Bulgaria (14a, Izvorska, Veleka, Karaagach and Kamchiya) and Turkey (14b, Gönen and Sapanca). Clade 9 = P. marsilii Heckel, 1836. Clade 20 = Crimea, Salhir. DFA statistics values: Wilks’ Lambda 0.16694, approx. F (45.1345) = 14.709, p <0.0000 (perfect discrimination).

Fig. 4. DFA based on 11 counts, two coded characters, and 55 relative measurements, for males and females separately (A) and for females only (B). Numbers of samples or clades as in Fig. 3. Abbreviations: f = females; m = males. DFA statistics values: A. Wilks’ Lambda 0.00000, approx. F (156.189) = 11.390, p <0.0000. B. Wilks’ Lambda 0.00002, approx. F (48.48) = 36.279, p <0.0000 (perfect discrimination).