Introduction

Aphanius is the only representative of the Cyprinodontidae (Teleostei, Cyprinodontiformes) in Eurasia. The genus occurs in coastal (brackish) and landlocked (freshwater to saline) water bodies in the Mediterranean and Persian Gulf basins from Iberian Peninsula as far eastwards as Iran and Pakistan (Wildekamp 1993). Aphanius species diversity is highest in the endorheic basins of the mountainous regions of central Anatolia and the Iranian plateau (Coad 2000; Hrbek and Meyer 2003, Hrbek et al. 2006, Esmaeili et al. 2012). Though central Anatolia is believed to represent the center of Aphanius speciation (Wildekamp et al. 1999), a high number of Aphanius species also occurs in Iran. Apart from the widely distributed Aphanius dispar (Rüppell, 1829), seven endemic Aphanius species have been described from Iran to date, namely Aphanius ginaonis (Holly, 1929) from the Genow hot spring near the Persian Gulf; Aphanius isfahanensis Hrbek, Keivany & Coad, 2006 from the endorheic Esfahan basin; Aphanius farsicus Teimori, Esmaeili and Reichenbacher, 2011 from the endorheic Maharlu Lake basin [Aphanius farsicus is a replacement name for the previous Aphanius persicus (Jenkins, 1910) because this name has been recognized as a homonym of the fossil Aphanius persicus (Priem, 1908) (Gaudant 2011, Teimori et al. 2011)]; Aphanius sophiae (Heckel, 1849) from the endorheic Kor River Basin; Aphanius vladykovi Coad, 1988 from the upper reaches of the Karoun basin; Aphanius mesopotamicus Coad, 2009 from the Tigris-Euphrates drainage; and the recently re-established Aphanius pluristriatus (Jenkins, 1910) from the Mond River drainage. In addition to the species listed above, Lebias punctatus and Lebias crystallodon were originally described from the Nemek Deria near Shiraz by Heckel (1846–1849). Berg (1949) and Coad (1996) considered Lebias punctatus to be a synonym of Aphanius sophiae but at that time most of now valid species distributed in Iran were thought to be synonyms of the widely distributed Aphanius sophiae. Coad (1996) strongly suggested that the type locality of Lebias punctatus is not the Lake Maharlu but some other lake nearby as a name Nemek Deria is a very common name in Farsi for a salt lake. However, later, the Kotschy’s itinerary in southern Iran in 1841 and 1842 was studied in detail based on botanical labels and it was clearly shown that collections by Kotschy studied by Heckel indeed came from a lake now called Maharlu (Edmondson and Lack 2006). This aspect is not in the focus of this very paper; we tentatively consider Lebias punctatus to be a synonym of Aphanius sophiae until a proper examination of the extant syntypes of Lebias punctatus is done.

A number of isolated Aphanius populations that might deserve species status have been reported from endorheic drainages in Iran, but have not yet been investigated in detail (Coad and Abdoli 2000; Hrbek et al. 2006; Esmaeili et al. 2010). They were commonly identified as Aphanius sophiae (Heckel, 1849) (Coad and Abdoli 2000; Kamal et al. 2009); however, it was shown that the true Aphanius sophiae is restricted to the endorheic Kor River basin near Shiraz (Fars Province) (Coad 2009; Esmaeili et al. 2012). This study describes a newly discovered Aphanius population from the Namak Lake basin in northern central Iran (Fig. 1). The specimens were collected in 2007 because they appeared to be different from other Iranian Aphanius species by a specific coloration. Here it is shown that the population from the Namak Lake basin in fact represents a new species, Aphanius arakensis. Our study is based on a total-evidence approach including morphometric and meristic characters, otolith morphology, and molecular data.

Material and methods

Institutional acronyms: ZM-CBSU, Zoological Museum of Shiraz University, Collection of Biology Department; ZSM, Zoological State Collection, Munich.

Material for morphological comparison.

Aphanius sophiae: 35 males (19.3–33.3 mm SL) and 35 females (18.6–36.6 SL) from the Ghadamgah spring-stream system (close to type locality) in the Kor River Basin (Iran, Fars Province), 30°15'N, 52°25'E. Males: ZM-CBSU, 8460, 8462, 8462, 8466, 8468, 8470, 8472-73, 8475, 8477, 8479, 8481, 8483, 8485, 8487, 8479, 8489, 8491-97, 8499, 8501, 8503-09, 8511-13; females: ZM-CBSU, 8461, 8463, 8465, 8467, 8469, 8471, 8474, 8476, 8478, 8480, 8482, 8484, 8486, 8488, 8490, 8498, 8500, 8502, 8510, 8514-29.

Aphanius farsicus: 35 males (20.0–26.8 mm SL) and 35 females (20.4–35.2 mm SL) from the Barm-e-Shur spring in the Maharlu Lake Basin (type locality) (Iran, Fars Province), 29°27'N, 52°42'E. Males: ZM-CBSU, 9413, 9415, 9417, 9421, 9441, 9443, 9447, 9449, 9459, 9467, 9481, 9483, 9485, 9487, 9489, 9493, 9497, 9499, 9503, 9511, 9513, 9515, 9517, 9519, 9527, 9529, 9531, 9533, 9537, 9539, 9541, 9555, 9557, 9559, 6375; females: ZM-CBSU, 9410, 9412, 9420, 9422, 9428, 9442, 9444, 9452, 9458, 9472, 9474, 9478, 9482, 9488, 9492, 9494, 9498, 9500, 9502, 9504, 9506, 9510, 9516, 9520, 9530, 9532, 9534, 9536, 9558, 9560, 9562, 9564, 6364, 6359, 6385.

Aphanius isfahanensis: 18 males (17.6–23.8 mm SL) and 25 females (17.7–34.0 mm SL) from the Zayanderh River near Varzaneh, Esfahan Basin (type locality) (Iran, Esfahan Province), 32°25'N, 52°39'E. Males: ZM-CBSU, 6472, 6474, 6476, 6478, 6480, 6482, 6484, 6486, 6488, 6490, 6492, 6494, 6496, 6498, 6500, 8602, 8604, 8613; females: ZM-CBSU, 6471, 6473, 6475, 6477, 6479, 6481, 6483, 6485, 6487, 6489, 6491, 6493, 6495, 6497, 6499 6501, 8603, 8605-8612.

Aphanius vladykovi: 35 males (17.3–29.2 mm SL) and 35 females (16.1–41.4 mm SL) from the Chaghakhor wetland in the upper reaches of the Karoun Basin (Iran, Chahar Mahale Bakhtyari Province), 31°55'N, 50°56'E. Males: ZM-CBSU, 6408-09, 6413-14, 6416, 6418, 6420-21, 6423, 6425-27, 6430, 6433-41, 6443-44, 6446, 6448-49, 6451-57: females: ZM-CBSU, 6401-03, 6405-07, 6410-12, 6415, 6417, 6419, 6422, 6424, 6428-29, 6431-32, 6442, 6445, 6447, 6450, 6458-70.

Material for molecular comparison.

Aphanius sophiae ZM-CBSU, M46, M97, M98, M174-176 (Ghadamgah spring-stream system); Aphanius farsicus ZM-CBSU, M47, M136, M177-178 (Barm-e-Shur spring); Aphanius isfahanensis ZM-CBSU, M211, M213-214 (Zayanderh River near Varzaneh); Aphanius arakensis sp. n. ZM-CBSU, M198-200 (Namak Lake Basin, 34°00'N, 49°50'E); Aphanius vladykovi ZM-CBSU, M60, M139, M209 (Chaghakhor wetland in the upper reaches of the Karoun Basin).

Materials from GenBank.

Aphanius vladykovi: GenBank DQ367526; Aphanius fasciatus:GenBank AF299273; Aphanius iberus:GenBank AF299290. Poeciliopsis gracilis (GenBank AF412155) was used as outgroup.

Morphological analysis

Based on the morphometric schemes introduced in Holcik et al. (1989) and Doadrio et al. (2002), 18 morphometric parameters were measured using a Vernier calliper and recorded to the nearest 0.5 mm. The standard length was measured from the most anterior part of the snout to the base of the caudal fin rays. In total, 21 relative variables were calculated from the measurements (Table 1).

Scales removed from the left side of each fish, from the 3rd or 4th row below the dorsal fin, were mounted between microscope slides, and length and width of scales were measured to the nearest 0.1 mm by using a scale reader (Xerox 320). For each individual, scale length and scale width measurements were averaged to obtain a single length value and a single width value per individual and relative width and length of scales were calculated following Esmaeili (2001).

The meristic characters were counted under a stereomicroscope and consist of the numbers of (i) dorsal (ii) pectoral (iii) pelvic and (vi) anal fin rays, (v) lateral line series scales, (vi) caudal peduncle scales (the numbers of scales along the caudal peduncle, i.e. from the base of the last anal fin ray to the base of the caudal fin rays in a direct line), (vii) gill rakers and (viii) flank bars of males. Two posteriormost rays in dorsal and anal fins were calculated as one ray.

For examination of otolith morphology fish skulls were opened ventrally in order to remove the right and left otoliths. Otoliths were cleaned from tissue remains in 1% potassium hydroxide solution for 3–6 h, washed several times and finally rinsed in distilled water for 12 h. Otolith morphology was analyzed under a stereo microscope. In addition, five or six otoliths from each population were examined by a scanning electron microscope (SEM) with a LEO 1430 VP at ZSM.

Univariate analysis of variance (ANOVA, with Duncan’s post hoc test, p < 0.05) was used to test the significance of phenotypic differences among species and also between sexes. Canonical discriminant analysis (CDA) was used for multivariate analyses in order to document the classification success of the groups. The statistical analyses were carried out using PASW 19.00 (SPSS Inc 2011) and PAST (Hammer et al. 2001: PAlaeontological STatistics, version 1.81).

Laboratory protocols and molecular analyses

Total genomic DNA was extracted according to phenol/chloroform procedures (Sambrook et al. 1989). A 900 base pairs (bp) fragment of the cytochrome b gene was successfully amplified via PCR using the primers (forward: Glu-F, 5’ - AACCACCGTTGTATTCAACTACAA-3’; reverse: ThrR, 5’-CCTCCGATCTTCGGATTACAAGACCG-3’ (Machordom and Doadrio 2001). Amplification was performed in a thermal cycler programmed as follows: initial 94 °C for 3 min, 35 cycles at 94 °C for 50 s, 56 °C for 45s, 72 °C for 1 min, followed by a final extension at 72 °C for 5 min. Sequencing was performed by Macrogen company, South Korea. Cytochrome b nucleotide sequences were edited with BioEdit and aligned through Geneious pro v5.4 (Drummond et al. 2011). Additional Aphanius sequences were obtained from the NCBI GenBank (http://www.ncbi.nlm.nih.gov) and included in the analyses (see above). The achieved cytb sequences for the here studied Aphanius populations were deposited in GenBank under numbers JX154880–JX154898.

Maximum likelihood-based phylogenetic relationships were estimated by using the program SeaView version 4 (Gouy et al. 2010). The best-fit model of nucleotide substitution was obtained using the program JmodelTest 0.1.1 (Posada 2008). Accordingly, the GTR + I + G model (= General Time Reversible model + proportion of Invariable sites + Gamma-shaped distribution of rates across sites) was chosen.

Maximum parsimony based phylogenetic relationships were estimated using the program SeaView version 4 (Gouy et al. 2010) with 100 heuristic searches using random additions of sequences and implementing the Close-Neighbor-Interchange (CNI) on random tree algorithm. To test this phylogeny, bootstrap method using 2000 replication was used. To document the degree of homoplasy and degree to which potential synapomorphy is exhibited on the tree, the Consistency Index (CI) and the Retention Index (RI) were calculated by using the parsimony model within the Mesquite system for phylogenetic computing (Maddison and Maddison 2011).

The Neighbor Joining (NJ) distance-based phylogenetic relationships were estimated by using the computer program Geneious pro v5.4 (Drummond et al. 2011). The HKY85 model (Hasegawa et al. 1985) of molecular evolution was used with gamma distributed among site rate variation. There were a total of 771 positions in the final dataset.

Results Discussion Probable reasons for morphological similarities between endemic Aphanius species

Several endemic Aphanius species are known that are soundly circumscribed by genetic differentiation and specific otolith morphology (see below), whereas they differ only weakly (or only in multivariate space) with regard to morphometry and meristics. Examples are Aphanius isfahanensis from central Iran, Aphanius sophiae and Aphanius farsicus from southern Iran (Hrbek et al. 2006, this study); another example from the Mediterranean area is Aphanius baeticus from Spain (Doadrio et al. 2002). Aphanius arakensis sp. n., from the Namak Lake basin represents yet another example for a species that is difficult to distinguish from its relatives based on external characters (with the exception of the features mentioned above). It is likely that the overall morphological similarity between these taxa are a result of the similar habitats, in which the various endemic Aphanius species are thriving. Thus, common environmental variables may have acted as a stabilizing selection on morphological characters (see also Hrbek et al. 2006). This offers an explanation as to why speciation events in Aphanius have affected genetic characters, rather than morphology, and why rapid genetic diversification can occur with little morphological change in this taxon (see also Adams et al. 2009).

Probable reasons for otolith differences between endemic Aphanius species

Otolith morphology is known to support the distinctive taxonomic state of several Aphanius species (Reichenbacher et al. 2007, 2009a-b). However, otolith morphology has not been used in previous studies on the endemic Iranian inland Aphanius species. Here we have compared the otoliths of Aphanius vladykovi, Aphanius isfahanensis, Aphanius farsicus and Aphanius sophiae (Fig. 3) to show that these species are clearly different with regard to otolith morphology. Also Aphanius arakensis shows clear divergence of its otolith morphology in comparison to the other inland Aphanius species, in particular with regard to the weakly pronounced antirostrum (Fig. 3W–Aa).

Notably, the otoliths of Aphanius vladykovi are most distinctive in comparison to those of the other studied species as they are characterized by a long ventral part, angular overall shape and long rostrum (Fig. 3R–V). This uniqueness of the Aphanius vladykovi otoliths corresponds well to our and previous phylogenetic analyses, which have established Aphanius vladykovi as being sister to all other Iranian inland species that diverged approximately 10 Ma ago (Hrbek et al. 2003). As a result, Aphanius likely has a higher rate of divergence in otolith morphology than in overall morphology. This difference in divergence rate may be related to the function of the otoliths as parts of the inner ear. In general, otoliths provide a mechanism for measuring motion and position of the head relative to gravity (Manley et al. 2004). However, it is quite important for a fish to know from where a sound is coming, so as to be able to distinguish between different sounds and pick out the biologically most relevant sounds (Popper et al. 2005). In addition, differences in otolith morphology are related to the balance and orientation of a fish (Popper et al. 2005). This means that differences in otolith morphology can reflect changes in intraspecific communication and behavior in fishes, that may have acted as evolutionary pressures.

Role of coloration pattern (flank bar numbers) in Aphanius diversification

Coloration and flank bar numbers are significant characters for the identification of Aphanius species, in particular for the identification of male individuals. Among the allopatric Iranian Aphanius species, males of Aphanius arakensis have the largest number of flank bars, and flank bars are non-overlapping, whereas the number of flank bars is lowest in Aphanius sophiae. Also the flank bars of the central Anatolian Aphanius species vary in thickness and number between species (Hrbek et al. 2002). However, the mechanisms underlying male flank bar variation have not been studied. We hypothesize that flank bar patterns play an important role in sexual selection, and thus represent important factors in the evolutionary history and speciation of Aphanius.

Sexual selection has long been believed to promote species divergence among groups of animals (see Kraaijeveld et al. 2010 for a review). Sexual selection may facilitate speciation because it can cause rapid evolutionary diversification of male mating signals and female preferences (Boughman 2001). Divergence in these traits may then contribute to reproductive isolation.

Several studies indicate that fishes can adapt to variation in underwater light environments by changing their colour, most likely as a result of a more effective intraspecific communication (Boughman 2001, 2002; Fuller 2002; Seehausen et al. 2008). Adding support to this interpretation is provided by studies on cichlids from the Victoria Lake (Seehausen et al. 2008) and African elephant fishes (Leal and Losos 2010). These studies indicate that variation in male nuptial coloration due to specific light conditions in different environments can result in ecological, phenotypic, genetic and behavioral differentiation. Additionally, color contrast with the visual background was found to be more important for effective intraspecific communication than color brightness (Fuller 2002). Thus, our conclusion is that the specific male flank bar patterns in different Aphanius species may have evolved as a response to different light regimes prevalent in respective habitats for increasing contrast and optimizing intraspecific communication. It can therefore be suggested that sensory-driven speciation mighthave played a prominent role in Aphanius speciation.

Conclusion

The noticeable features of the present-day diversity of the endemic Aphanius species in Iran include high genetic divergence and clear differences in otolith morphology, but only weak differences in general external morphology, morphometry and meristics. These patterns are probably caused by different rates of evolution in the mentioned characters that may be linked to the similarity of the individual environments, intra-species communication, and vicariance events. It is likely that additional Aphanius species are present in remote areas of Iran, especially in the Zagros and Alburz Mountains.