According to Perea et al. (2010), there are 14 distinct clades within Leuciscinae of inner waters across the Mediterranean Region, 9 of which being represented in Anatolia. Among these, phylogenies of the genera Petroleuciscus and Pseudophoxinus are most debated; indeed even taxonomic status of some Pseudophoxinus species is uncertain. Two well-defined monophyletic clades represent Anatolian Pseudophoxinus species: first of these corresponds to Central Anatolian species complex including Pseudophoxinus alii, Pseudophoxinus anatolicus, Pseudophoxinus antalyae, Pseudophoxinus battalgilae, Pseudophoxinus crassus, Pseudophoxinus elizavetae, Pseudophoxinus evliyae, Pseudophoxinus fahrettini, Pseudophoxinus ninae, and a probable undescribed species; while the other includes Levantine taxa, namely Pseudophoxinus firati, Pseudophoxinus kervillei, Pseudophoxinus zeregi and Pseudophoxinus zekayi (Perea et al. 2010). Speciation in Pseudophoxinus is heterogenous, as mentioned by Hrbek et al. (2004) and Bogutskaya et al. (2007), and discrepancies between morphological and molecular relationships can be seen (Perea et al. 2010). Separation of the Pseudophoxinus taxa in Anatolia and western Asia into two groups by Bogutskaya et al. (2007), according to a comparison of morphological parameters (sensory pores, scales and their arrangement on the body, vertebral counts, and supraethmoid bone) with that of the type species Pseudophoxinus zeregi is another example of the high degree of variability in the genus.

Morphological and phylogenetic distinctness of Pseudophoxinus egridiri (Hrbek et al. 2004: 305), and its closeness to the Pelasgus-Delminichthys lineage (Perea et al. 2010) brings some doubts about monophyly of the Anatolian Pseudophoxinus taxa.

As stated by Hrbek et al. (2002, 2004), Anatolia is an important diversification center for the genus Pseudophoxinus which shows allopatric speciation especially in basins of Bey Dağları, Büyük Menderes, Tuz Lake and Lakes District. However, the complex taxonomy of the genus in these basins is still unresolved. Although including several lakes and springs of Lakes Region (Lake Salda, Karapınar Spring near Yeşilova district, Düğer Spring, Lake Bahçeözü, Sazak Spring and Kırkpınar Springs) in the distribution area of Pseudophoxinus maeandri, Bogutskaya (1992) mentioned that Pseudophoxinus maeandri populations from Upper Büyük Menderes basin (Lake Işıklı and Düden Spring near Dinar) morphologically differed from remaining populations in having larger scales and shorter lateral line, fewer lateral series scales and gill rakers on the first branchial arch, as well as fewer vertebrae.

According to molecular data presented by Hrbek et al. (2004) Kırkpınar (Lake Söğüt source, Korkuteli) and Avlan source (Elmalı) populations were clearly separated from those in the basins of the lakes Salda and Burdur. Freyhof and Özuluğ (2009) identified populations of Kırkpınar as a new species, Pseudophoxinus evliyae, and populations of Lake Salda and Lake Burdur basin as Pseudophoxinus ninae. Perea et al. (2010) did not examine any material from either the type locality (Onaç Stream-Bucak) or other known localities of Pseudophoxinus ninae (Kestel Swamp-Bucak, Düğer Spring, Lake Karataş and Sazak Spring) and used only mitochondrial and nuclear DNA markers of Lake Salda specimens as representatives of Pseudophoxinus ninae. Thus, it was not possible to explain diversification of Pseudophoxinus in southwestern Anatolia (Fig. 1) in its entirety. Furthermore, since its description by Ladiges (1960), sufficient information on distribution and taxonomy of Pseudophoxinus maeandricus has not been given up to date. With regard to the above mentioned morphological and molecular data, it became apparent that the taxonomic position of Pseudophoxinus populations in Burdur and Salda lake basins needed clarification and this led to the present study.

Fish specimens were caught by pulsed DC electrofishing equipment and killed by over anaesthetization, preserved in 5% formalin. Material is deposited in: IFC-ESUF, Inland Fishes Collection, Eğirdir Fisheries Faculty of Süleyman Demirel University. Counts and measurements follow Kottelat and Freyhof (2007), all measurements being point to point made with a digital calliper (0.01 mm sensitive). Standard length (SL) was measured from the tip of the upper lip to the end of the hypural complex. The length of the caudal peduncle was measured from behind the base of the last anal fin ray to the end of the hypural complex, at mid-height of the caudal fin base. Lateral line scales are counted from the anteriormost scale (the first one to touch the shoulder girdle) to the posteriormost one. Scales in lateral series are counted along the midlateral line from the first one to touch the shoulder girdle to the last scale at the end of the hypural complex. Scales on the caudal fin itself are indicated by “+” (Freyhof and Özuluğ 2009). The last two branched dorsal and anal fin rays articulating on a single pterygiophore were counted as 1½. Vertebral counts were obtained from radiographs and counted as total, predorsal, abdominal and caudal vertebrae following Naseka (1996). Abdominal vertebrae were counted from the first Weberian vertebra to the one just anterior the first caudal vertebra. The first caudal vertebra is that with its haemal spine fully developed. The count of total and caudal vertebrae includes the last complex vertebra bearing hypurals. Osteological characters were examined in cleared and stained with alizarin Red-S specimens and from radiographs (Bogutskaya 1996).

The morphometric characters of the two species of Pseudophoxinus from Turkey were compared by Principal Component Analysis (PCA) using a covariance matrix on log–transformed measurements and counts with the software package PAST version 1.8 (Hammer et al. 2001).

Abbreviations.CSO

supraorbital canal

Hrbek et al. (2004) stated that monophly of Anatolian Pseudophoxinus taxa was not well supported and these represented 6 distinct clades. Of these, Lakes Region populations (Lakes Salda and Karataş, Düğer Spring, Karapınar) identified as Pseudophoxinus maeandri formed Clade IV, while the Avlan and Kırkpınar populations identified as Pseudophoxinus fahirae (now Pseudophoxinus evliyae) formed a separate clade (Clade VI). Perea et al. (2010) mentioned of two well defined monophyletic groups within Anatolian Pseudophoxinus taxa, though not discussion interrelationships of these two. A phylogenetic tree based on the cytb gene sequence, Hrbek et al. (2004: 299) showed presence of two different clade; (I) Lake Avlan and Kırkpınar (Lake Söğüt source) populations (Pseudophoxinus evliyae) and (II) populations from Burdur and Salda basins.

In this study, we also morphologically compared the new species with Pseudophoxinus ninae (the Onaç Stream and Kestel Swamp), Pseudophoxinus evliyae (dried Lake Söğüt [Kırkpınar Village, Korkuteli] and source of Lake Avlan [Elmalı]), Pseudophoxinus maeandri [Lake Işıklı], Pseudophoxinus maeandricus (Karadirek Stream [Upper Büyük Menderes basin]), Pseudophoxinus alii (Köprüçay and Ilıca Stream), Pseudophoxinus battalgilae (Manavgat River basin, Lake Akgöl [Ereğli], Lake Çavuşcu [Ilgın] and Lake Suğla [Seydişehir]), Pseudophoxinus fahrettini (Köprüçay River basin [Bağıllı Village and Değirmenözü Stream]), Pseudophoxinus antalyae (Kırkgöz Spring, Karamanlı Stream and tributaries of Düden canal in Antalya), and Pseudophoxinus elizavetae (Sultansazlığı [Kayseri]).

Pseudophoxinus burduricus is most similar to Pseudophoxinus ninae (Fig. 4). Itis distinguished from Pseudophoxinus ninae by having fewer branched pelvic fin rays (7, vs. 8 or 9), slightly fewer gill rakers in outer side of the first gill arch (7–8, rarely 9, vs. 8–9) and a longer caudal peduncle (caudal peduncle length 1.4–1.9 times caudal peduncle, vs. 1.3–1.4). It further differs from Pseudophoxinus ninae by having a shorter predorsal distance (52–57% SL, mean 55.0, vs. 57–61, mean 58.0), a shorter preanal distance (69–74% SL, mean 72.0, vs. 73–77, mean 74.5), a somewhat shorter head (head length 26–28 % SL, mean 26.8, vs. 27–31, mean 28.6), and narrower and relatively deeper hyomandibular bone (vs. wide and shallow) (see Figs 5a, b, c). Also Pseudophoxinus burduricus and Pseudophoxinus ninae were compared by Principal Component Analysis (PCA). The PCA was performed in using 18 morphometric characters of the two Pseudophoxinus species. The PCA clearly separated Pseudophoxinus burduricus from Pseudophoxinus ninae (Fig. 6). Variables loading on the first metric PC I–II are given in Table 3.

Pseudophoxinus burduricus is distinguished from Pseudophoxinus evliyae by having fewer scales in lateral series (47–57+1-2, vs. 54–64+1-2), fewer branched pelvic fin rays (7, vs. 8), fewer gill rakers in outer side of first gill arch (7-8, rarely 9, vs. 8-9 to 10 in some specimens), fewer scales between lateral line and dorsal fin origin (10½-12½, vs. 13½-15½), and a faint epidermal black or violet stripe along lateral midline from eye to caudal fin base (vs. black prominent stripe). Pseudophoxinus burduricus is distinguished from Pseudophoxinus maeandri by having more lateral line and lateral series scales (21–39, vs. 19–27 and 47–57+1-2, vs. 40–44+1-2 respectively), more total vertebrae, 36-39 (vs. 35–36) (see Table 2). Pseudophoxinus burduricus differs from Pseudophoxinus maeandricus by having an incomplete lateral line (vs. complete), a shorter pelvic-fin (reaching to anus, vs. not reaching), wider and deeper head (head width at nape 55-62 % HL, vs. 46–48; head depth at nape 77–85 % HL, vs. 67–73). Pseudophoxinus burduricus isdistinguished from Pseudophoxinus alii by having more scales in the lateral series (47–57+1-2, vs. 38-43+1-2), fewer perforated scales (21–39, vs. 38–41) and smaller eyes (eye diameter 22–26 % HL, vs. 26–32). Pseudophoxinus burduricus is distinguished Pseudophoxinus antalyae by presence of a faint black or violet lateral stripe (vs. plain golden or orange stripe when alive), rounded snout (vs. pointed) and shorter pharyngeal teeth. Pseudophoxinus burduricus is distinguished from Pseudophoxinus battalgilae by an incomplete lateral line (vs. complete), fewer gill rakers in outer side of first gill arch (7–8, rarely 9, vs. 13–16), fewer branched anal fin rays (6–7, vs. 8, respectively), and the absence of a keel between the pelvic fin base and the anus (vs. presence). Pseudophoxinus burduricus is distinguished Pseudophoxinus fahrettini by an incomplete lateral line (vs. complete), fewer perforated scales lateral line (21–39, vs. 73–88) and fewer gill rakers in outer side of first gill arch (7–8, rarely 9, vs. 11–13). Pseudophoxinus burduricus is distinguished Pseudophoxinus elizavetae by having fewer scales in the lateral series (47–57+1-2, vs. 56-62+2-3), fewer gill rakers in outer side of first gill arch (7–8, rarely 9, vs. 11–13).