Aphanius kruppi, Freyhof & Weissenbacher & Geiger, 2017

Freyhof, Jörg, Weissenbacher, Anton & Geiger, Matthias, 2017, Aphanius kruppi, a new killifish from Oman with comments on the A. dispar species group (Cyprinodontiformes: Aphaniidae), Zootaxa 4338 (3), pp. 557-573 : 561-568

Aphanius kruppi , new species

( Figs. 2–5 )

Holotype. ZFMK-ICH 103668 , 46 mm SL; Oman: spring in Al Mudayrib , 22°36'46"N 58°40'31"E; captive-bred . Paratype. FSJF 3771 , 69 , 21–52 mm SL; same data as holotype .

Additional material. FSJF 3537 , 2 , 26–32 mm SL; Oman: Wadi Bani Khalid at Sayh al Hayl, 22°35'43"N 59°5'13"E.

Material used in the molecular genetic analysis. FSJF DNA-2577 , Oman: Wadi Bani Khalid at Sayh al Hayl 22°35'43"N 59°5'12"E; (GenBank accession numbers: MF918597 , MF918598 ) . — FSJF DNA-2588 , Oman: spring in Al Mudayrib , 22°36'46"N 58°40'31"E (GenBank accession numbers: MF918593 , MF918594 , MF918595 , MF918596 ) .

Diagnosis. Aphanius kruppi is superficially similar to A. sirhani by colour pattern. Female A. kruppi are distinguished from female A. sirhani by having a silvery flank with a series of 7–12 short bars or narrow, verticallyelongated blotches along the lateral midline and a series of 3–7 vertically-elongated blotches below the lateral midline in some individuals, or a plain flank without pattern (vs. one midlateral row of 1–8 dark-brown, roundish spots; spots sometimes also at other places on the flank). Male A. kruppi are distinguished from male A. sirhani by lacking a black dorsal-fin margin (vs. wide black margin), having a dorsal-fin pattern of 5–10 irregularly set and shaped bands of very small spots or blotches, often restricted to blotches on the rays (vs. 1–2 black bars or a marbled pattern) and the anal fin being yellow anteriorly with 3–8 narrow black bars on the posterior part (vs. anal fin completely yellow with or without a narrow black margin and 0–2 black bars). Male A. kruppi have three black bars on the caudal fin. The first bar is situated shortly behind the caudal-fin base and is often faded or grey. In male A. sirhani , there are two black bars on the caudal fin; the bar on the caudal-fin base is absent.

Aphanius kruppi is distinguished from A. dispar , A. richardsoni and A. stoliczkanus by having black or darkgrey bars on the flank, the first being situated under the pectoral fin and the last on the caudal-fin base in males (vs. flank bars in male absent or restricted to the caudal peduncle) and a diamond-shaped or slightly vertically-elongate black or dark-brown blotch at the caudal-fin base in females (vs. a long, narrow bar in A. dispar , A. richardsoni and A. stoliczkanus , bar often absent in A. richardsoni ). Aphanius kruppi is also distinguished from A. dispar , A. richardsoni and A. stoliczkanus by having 2–3 scale rows on the caudal-fin base reaching to the anterior margin of the second black caudal-fin bar in males (vs. 4–7 scale rows, reaching to the middle or the posterior margin of the second black caudal-fin bar). The dorsal and anal fins are short in males of A. kruppi , not reaching or reaching to about middle of caudal peduncle (vs. reaching to beyond middle of caudal peduncle in A. dispar and A. stoliczkanus ). But we are not sure if this might be related to the conditions in which the fish were kept. Other species of the A. dispar group kept under similar conditions develop long fins.

Aphanius kruppi is distinguished from A. ginaonis by having 6½–7½ branched dorsal-fin rays (vs. 5) and 2–3 scale rows on caudal-fin base (vs. 4–5). Several individuals of A. kruppi have 7 dorsal-fin rays, lacking the ½ ray and all examined A. ginaonis lack the ½ ray at the end of the dorsal fin. Male A. kruppi are also distinguished from male A. ginaonis by having 9–14 brown or grey bars on the flank (vs. 15–19). The flank bars are as wide as or wider than the interspaces in A. kruppi (vs. as wide as, or narrower than the interspaces). Female A. kruppi have a diamond-shaped or slightly vertically elongate black or dark brown blotch at the caudal-fin base (vs. a narrow and long bar).

Aphanius kruppi is distinguished from A. furcatus by having a truncate caudal fin (vs. emarginated) and the body completely covered by scales (vs. naked). Aphanius kruppi is distinguished from A. stiassnyae by the lower jaw gently upturned, oriented about 45° to body axis (vs. sharply upturned, its tip almost above, about 90° to body axis).

Description. See Figures 2–5 for general appearance and Table 3 for morphometric data of holotype and 20 paratypes. Dorsal head profile straight or concave. Dorsal profile straight or slightly convex from nape to dorsal-fin origin. Ventral profile convex. Body deeper than wide, compressed. Body deepest at about dorsal-fin origin. Greatest body width at pectoral-fin base or at middle of belly in female. Lower jaw gently upturned, oriented about 45° to body axis. Caudal peduncle compressed laterally, 1.5 in holotype, 1.4–1.6 in males and females, times longer than deep. Pectoral fin rounded, reaching half or full of distance between pectoral- and pelvic-fin bases. Pelvic-fin origin below vertical of 2nd or 3rd branched dorsal-fin ray. Pelvic fin not reaching to anus. One large scale between pelvic-fin bases. Anus situated slightly in front of anal-fin origin, tissue around genital papillae swollen and papillae deeply hidden between scales in nuptial females. Dorsal and anal fins with convex posterior margin. Dorsal and anal fins roundish in females, tip of longest dorsal-fin ray reaching to a vertical of middle of last analfin ray, elongated in males, reaching to tip of anal fin or to a point slightly before. Caudal fin rounded to truncate. Largest individual examined 52 mm SL.

Dorsal fin with 6½ (5), 7(5), or 7½ (11) branched rays. Anal fin with 7½ (7)–8½ (13) branched rays. Caudal fin with 8+7 (12)–8+8 (8) branched rays. Pectoral fin with 14–16 and pelvic fin with 6–7 rays. Trunk and head entirely scaled. Scales large and cycloid in females, with small ctenae in males. Scale above pectoral-fin origin enlarged. One scale row on upper part of opercle. Flank with 24(1), 25(5), 26 (25), 27 (4), 28 (1) scales along lateral series. Two or three additional rows of small scales on anterior caudal-fin base. Between dorsal- and pelvic-fin origins 9, rarely 10, scale rows. Caudal peduncle with 16, rarely 14, circumpeduncular scales. Lateral line incomplete, with 9–15 pores, scales pored mostly behind vertical of pelvic-fin origin. Teeth tricuspid, median tip is longer than laterals.

Coloration. See Figures 2–5 for general appearance. Living and preserved males: All yellow, orange, silvery and blue colours faded in preserved fish. Head and flank silvery, whitish with brown or dark-grey pattern, head and back brown or dark-grey. Lower cheek, breast and belly whitish or pale yellow. Lateral head with a blueish hue in life in some individuals. A roundish or horizontally elongated orange blotch on opercle in life. Flank between pectoral-fin base and a vertical of pelvic-fin origin with a brown or yellowish-brown net-shaped pattern forming roundish silvery or pale-blueish blotches, often just few silvery blotches on grey or brown background. Flank with 9–14 bars, confluent with brown or dark-grey back. Scales on back with silvery or yellowish margin in life. Bars brown in life, black in preserved individuals. Bars below lateral midline, in front of dorsal-fin origin, usually yellowish-brown in life. Interspaces silvery, as wide as or narrower than bars. First bar directly behind pectoral-fin base, last bar shortly before caudal-fin origin. Top of head pale yellowish-brown in life, black in preserved individuals. Pectoral fin hyaline or greyish-blue in life, blackish in preserved individuals. Pelvic fin hyaline or white with 1–3 short, narrow black bars on posterior rays. Anal fin hyaline in life, whitish in preserved individuals, greyish-blue or yellow anteriorly with 2–3 short, narrow black bars on posterior part. Dorsal fin with 5–10 irregularly set and shaped bands, often restricted to blotches on rays, dorsal fin without black margin or black bars. Caudal fin hyaline with 3 wide, black bars. First caudal-fin bar situated shortly behind caudal-fin base, often faded, pale grey, only upper and lower tips bold black.

Living and preserved females: Top of head and back pale-brown, cheek, ventral side of head, belly and flank silvery-grey or pale-brown. Scales on back with silvery-brown margin. Flank plain, without colour pattern in some individuals larger than 40 mm SL, with a faint stripe in some preserved and live individuals. One series of 7–12 narrow, vertically elongated blotches or short bars along lateral midline and a series of 3–7 vertically elongated blotches below lateral midline in some individuals. A diamond-shaped or slightly vertically elongate, bold, black blotch of about one or ½+1+½ scale size at middle of caudal-fin base. All fins hyaline in life, grey with a grey hue in preserved individuals.

Distribution and conservation. Aphanius kruppi was found in a spring in Al Mudayrib and in Wadi Bani Khalid, both in Oman. Both places are situated in Wadi al Batha, which drains to the Arabian Sea north of Al Ashkharah. Aphanius kruppi might be a threatened species, as we know only two very small populations.

Etymology. Named for Fareed Krupp (Doha, Qatar) for his many value contributions to the explorations of the freshwater fishes of the Middle East. A noun in genitive, indeclinable.

Remarks. Besides A. kruppi from its type locality, we are aware of two Omani populations showing bars on the flank anterior to the pelvic-fin origin. These are fishes from the Wadi Bani Khalid (FSJF 3537, Fig. 5 ) and from Al-Hoota (FSJF DNA-2590, Fig. 6 ). As the type locality of A. kruppi, Wadi Bani Khalid also belongs to the Wadi al Batha drainage, but Al-Hoota is a headwater to Wadi Sayfam. Wadi Sayfam is a very large drainage system bringing water into the desert and it might have reached the Arabian Sea between Al Khaluf in the north and Ras Madrakah during more humid climatic periods. Maizels (1987) reported various such humid periods in the area, the last most likely at the end of the last glaciation. The other Omani populations studied by us live in coastal drainages flowing to the Oman Sea. The fish from Wadi Bani Khalid do not only agree in their colour patterns with A. kruppi , both populations also have very similar DNA barcode sequences. Therefore, we identify the Wadi Bani Khalid population as A. kruppi . Fish from Al-Hoota ( Fig. 6 ) have the same colour pattern as those from Wadi Bani Khalid, but share the DNA barcode haplotype with populations of A. stoliczkanus from the Omani coast. The Al-Hoota population is identified here as A. cf. kruppi as they show slightly intermediate character states between A. kruppi and A. stoliczkanus . Fish from the Al-Hoota population have black or dark-grey bars on the flank, the first being situated under the pectoral fin (vs. flank bars in male absent or restricted to the caudal peduncle in A. stoliczkanus ), a vertically elongate black or dark-brown blotch at the caudal-fin base in females (vs. a long, narrow bar), 4 scale rows on the caudal-fin base (vs. 2–3 in A. kruppi , 4–7 in A. stoliczkanus ) and the dorsal and anal fins in males are long (vs. short in A. kruppi , long in A. stoliczkanus ). More field work is needed to clarify the distribution of A. kruppi and further molecular studies are needed to test whether the population from Al-Hoota (and maybe others), might be the victims of introgressive hybridization of A. stoliczkanus into A. kruppi . Hybridisation between the two species might be the result of translocations of Aphanius for mosquito control as indicated by Krupp (1988). However, the biogeographic situation in the area is not well-understood. For example, widespread cyprinids of the genus Garra are represented by two different species in Wadi Bani Khalid ( G. gallagheri ) and Wadi Sayfam ( G. cf. longipinnis ) ( Lyon et al. 2016).

Aphanius dispar was described by Rüppell (1829:66) from the Red Sea. No details are given in the original description and Rüppell (1929:67) only mentioned that this fish lives all over in the Red Sea and is also found in the thermal springs of Hadjer Elme at Tor (Sinai). Joachim Scholz (SMF) studied the unpublished diaries of Rüppell for us without finding details, where Rüppell might have collected the type series of A. dispar . Already Villwock (1983) wrote that there is no detailed information on the type locality of A. dispar available. Only the local name of that species in Suez is mentioned. Indeed, Rüppell collected fishes in the Gulf of Suez, in Aqaba, in Mohila (= Al Muwaylih) in the northern Red Sea as well as close to Jeddah in the middle Red Sea, in Massaua in the southern Red Sea in Eritrea and maybe at other places. Rüppell (1829) mentioned these places as he received some marine fish just from one or few of these places. But no details are given for the more widespread species he collected as A. dispar and several others. It cannot be excluded that the type series of A. dispar was collected at another place in the Red Sea.

We were able to study molecular characters of three populations of A. dispar from the Red Sea basin: one from Saudi-Arabia, one from Eritrea and one from the northernmost part of the Gulf of Suez in Egypt. These populations might represent three different species. Having been in the Gulf of Suez, Jeddah ( Saudi Arabia) and Massaua ( Eritrea), Rüppell might have collected his A. dispar from at least two out of the three molecular groups found by us. It is beyond the aim of this study to determine which of the molecular lineages found corresponds to the types of A. dispar . Also, we do not exclude that more distinct molecular groups, actually identified as A. dispar , occur in the Red Sea and in East Africa; and Rüppell's A. dispar is not included in the materials examined in this study. More undescribed species might occur and several of the nominal species from Somalia placed in the synonymy of A. dispar by Villwock et al. (1983) might indeed represent valid species.

Our molecular analysis suggests that the Saudi Arabian and Eritrean populations are considerably isolated from those found in the northernmost Gulf of Suez. Aphanius dispar from the Gulf of Suez share the same COI haplotypes as those from Atlit in the Israeli Mediterranean Sea basin. Ben-Tuvia (1985) and also Galil (2007) speculate that " A. dispar " is an early Lessepsian immigrant from the Red Sea to the Mediterranean. However, already Villwock et al. (1983) and Villwock (1985) discussed how the distribution of A. dispar corresponds to several geological events and concluded that A. dispar is native to the Mediterranean. We find the populations from the Mediterranean and the Gulf of Suez to be very closely related. The excellent materials at ZMH from the lower Nile, Siwas and Faiyum oasis, Bitter and Timsach lakes and Suez represent one species, which is identical to fish from Atlit in Israel. Therefore, we speculate that there might be an endemic lineage of A. dispar in the Mediterranean, which is also native to the Gulf of Suez and the northern Red Sea pointing to a very recent biogeographic disconnection of the present-day area of the Suez Canal, at least for fish species tolerant of extreme water chemistries, such as A. dispar .

Aphanius richardsoni has been considered a subspecies of A. dispar ( Goren 1974, Villwock et al. 1983). We treat A. richardsoni as a valid species as it is well defined by its unique colour pattern (silvery vermiculations on a grey or brown background on the flank in males) and molecular characters (2.8% minimum K2P sequence divergence to A. dispar from the adjacent Mediterranean Sea).

Our molecular data do not allow us to identify the populations usually considered to represent A. dispar from the Arabian/Persian Gulf, the Oman Sea and the northern Arabian Sea as this species. The only names from this area available for this Aphanius species are A. ginaonis and A. stoliczkanus . Both are well-distinguished from A. dispar from the Red Sea by molecular characters (4.1 % minimum K2P sequence divergence to populations from Suez and Saudi Arabia, 4.6 % minimum K2P sequence divergence to populations from Eritrea). Aphanius stoliczkanus was described from Kachh, ( Day 1872) which is the coastal region in today's state of Gujarat in India. The PTP model-based species delimitation approach detected the Indian and the populations from the Arabian/ Persian Gulf and the Oman Sea as two entities. Despite a 1.7 % minimum K2P distance between the Indian and the Arabian populations, we identify all these as A. stoliczkanus , based on the absence of clear differences in morphology or colour pattern. The molecular distance between Indian and Arabian populations of A. stoliczkanus might be due "isolation by distance" as the geographic distance between both groups (between Kajou River in Iran and Gujarat in India, Fig. 7 ) is about 1000 km shoreline. However, we studied also the morphology of fish from the Salt Pans of Karachi in Pakistan, which is only about 600 km coastline east of the Kajou River and found these fully identical with those fish examined from Gujarat in India. More studies on material from the Pakistani coast are necessary to find out, if there is a clear border of distribution of both molecular groups or if populations with “intermediate”, i.e. connecting haplotypes exist in that area.

The molecular difference between Indian and Arabian populations of A. stoliczkanus also questions the hypothesis of a continuous geneflow between populations of this euryhyaline species. This result supports the view of Teimori et al. (2012), who find a remarkable molecular diversity in different populations in the Arabian/Persian Gulf and reject the hypothesis of a continuous geneflow in that area. As discussed above, we have to reject the hypothesis of the existence of one widespread, coastal species not only in the Red Sea basin, but also in the Arabian/Persian Gulf and the Arabian Sea.

Aphanius kruppi is immediately distinguished from A. stoliczkanus ( Fig. 7–8 ) by its colour pattern. We are aware that Teimori et al. (2016) identified fish from Iran as A. dispar , which have the flank almost completely covered by bars. We examined such fish from Iraq ( ZMH 4373, ZMH 4380). These populations are identified as A. stoliczkanus . Aphanius kruppi is distinguished from these populations by having a diamond-shaped or slightly vertically-elongate black or dark-brown blotch at the caudal-fin base in females (vs. a long, narrow bar), 2–3 scale rows on the caudal-fin base reaching to the anterior margin of the second black caudal-fin bar in males (vs. 5–6 scale rows) and short dorsal and anal fin in males not reaching or reaching to about middle of caudal peduncle (vs. anal fin reaching, dorsal fin almost reaching to caudal-fin base) in males larger than 35 mm SL. It has to be noted, that the PTP model-based species delimitation approach detected another entity, represented by one individual from the Khor Kalba nature reserve ( UAE). Two individuals were analysed from that site, one clustering inside and one clustering outside A. stoliczkanus . We suspect that Khor Kalba is situated at the border area to another molecular group identified as A. dispar . However, fish from Bahrain, on the southern shore of the Arabian/Persian Gulf, are nested within A. stoliczkanus .

Aphanius ginaonis is well-distinguished from A. stoliczkanus by having fewer total dorsal-fin rays (5 vs. 6½– 7½) and male having 15–19 brown bars on flank, as wide or narrower than interspaces (vs. roundish or ovoid silvery spots or blotches, sometimes with short vermiculation). In our molecular dataset, A. ginaonis form a distinct group. Reichenbacher et al. (2009) speculated that " A. dispar " from downstream of Giano spring might have been introduced to the spring between 2001 and 2009 and that both species have hybridised. Our studied material was collected in 2009 and A. ginaonis is still well separated from A. stoliczkanus by the number of dorsal-fin rays and their DNA barcodes (minimum K2P 2.3%). However, it has to be noted that no populations of A. stoliczkanus adjacent to Giano spring were studied. Also, the relationships between the different species and species groups within our molecular dataset as shown in Fig. 1 are only poorly supported and no strong phylogenetic conclusions should be derived based on this single marker alone.

Figures

FIGURE 1. Maximum likelihood estimation of the phylogenetic relationships based on the mitochondrial COI barcode region (Kimura 2 - parameter model, discrete Gamma distribution for rate differences with 5 categories + G parameter = 0.3241. Nucleotide positions with less than 95 % site coverage were eliminated resulting in 581 analysed positions. Numbers of major nodes indicate bootstrap values from 1000 pseudoreplicates from the ML, NJ and MP method. Except to outgroup, the tree is drawn to scale with branch lengths depicting number of substitutions per site. Solid black bar shows the delimitation result using the PTP approach (eleven groups), dashed line the result based on the mPTP model (two groups).

FIGURE 2. Aphanius kruppi, ZFMK-ICH 103668, holotype, 46 mm SL; Oman: spring in Al Mudayrib.

FIGURE 3. Aphanius kruppi, FSJF 3771, paratypes, left side from top: males 28, 37 and 44 mm SL; right side from top: females, 23, 35 and 49 mm SL; Oman: spring in Al Mudayrib.

FIGURE 4. Aphanius kruppi, FSJF DNA- 2588, female, 45 mm SL, male, 34 mm SL; Oman: spring in Al Mudayrib.

FIGURE 5. Aphanius kruppi, FSJF DNA- 2577, female, 27 mm SL; male, 24 mm SL; Oman: Wadi Bani Khalid.

FIGURE 6. Aphanius cf. kruppi, not preserved, female, about 40 mm SL, male, about 45 mm SL; Oman: Al-Hoota. Photo by Barbara Nicca (female), Kai Borkenhagen (male).

FIGURE 7. Aphanius stoliczkanus, FSJF 3670, female 28 mm SL, male, 30 mm SL; India: Narara-Salt Pans.