Revision of the Garra species of the Hajar Mountains in Oman and the United Arab Emirates with the description of two new species (Teleostei: Cyprinidae) Author Kirchner, Sandra Author Kruckenhauser, Luise Author Pichler, Arthur Author Borkenhagen, Kai Author Freyhof, Jörg text Zootaxa 2020 2020-03-18 4751 3 521 545 journal article 10.11646/zootaxa.4751.3.6 0fd5ecf7-48de-4920-8ef5-93b6537f735e 1175-5326 3714753 74A7039B-1B5B-4E67-8CF2-E4D57E44C06B Garra longipinnis Banister & Clarke, 1977 ( Fig. 5–6 ) Garra longipinnis Banister & Clarke, 1977:137 ( type locality: Saiq, Jabal al Akhdar, 23°02’N , 57°28’E , Oman ); Garra barreimiae Central Clade ( Pichler et al. 2018 , Kirchner et al . submitted). Material examined. BMNH 1968.10 .11.1, holotype , 52 mm SL , BMNH 1968.10.11.2-8, 7, paratypes , 37–50 mm SL; Oman : Saiq , Jebel Akhdar .— FSJF 4078 , 16 , 43–58 mm SL; Oman : pool in front of Ghubrat Tanuf Cave , 23.071°N , 57.368°E .— FSJF 4080 , 22 , 42–59 mm SL; Oman : spring pool in old town of Saiq , 23.073°N , 57.663°E . Material used in molecular genetic analysis. NMW-100034_1; NMW-100034_2; Oman : Wadi Dhum, 23.233°N , 57.074°E (GenBank accession numbers: MN 830846 , MN 830847 ).— NMW-100047_1, NMW-100047_ 2; Oman : Wadi Falahi ; 23.080°N , 57.350°E ( GenBank accession numbers: MN830848 , MN830849 ) .— NMW- 100032; Oman : Al Hoota cave , 23.084°N , 57.357°E ( GenBank accession numbers: MN 830414 ) .— NMW-100039; Oman : Hoti Pit , 23.101°N , 57.372°E ( GenBank accession numbers: MN 830415 ) .— NMW-100038; Oman : Lizq , 22.715°N , 58.172°E ( GenBank accession numbers: MN 830413 ) .— NMW-100031; Oman : Wadi Nakhar , 23.174°N , 57.200°E ( GenBank accession numbers: MN 830412 ).— FSJF DNA-3238 ; Oman : spring pool in old town of Saiq , 23.073°N , 57.663°E . ( GenBank accession numbers: MN830851 , MN830852 , MN830853 ).— FSJF DNA-3236 ; Oman : pool in front of Ghubrat Tanuf Cave , 23.071°N , 57.368°E . ( GenBank accession number: MN 830850 ) . Diagnosis. Garra longipinnis is distinguished from G. barreimiae by having a greenish-brown or grey flank without or with slight mottling (vs. strongly mottled), a faint lateral stripe present in some individuals (vs. absent), orange midlateral scales on the flank absent (vs. present), orange spot at the upper opercle absent (vs. present), dorsal-fin tip not white (vs. white), and 12–14 gill rakers on the lower limb of the first gill arch (vs. 15–18). Garra longipinnis is distinguished from G. gallagheri by possession of bold, dark-brown or grey scale pockets on the dorsal surface and dark-brown or grey scale margins on the flank which are clearly darker than the centre of the scales (vs. scale pockets and scale margins same colour as or only slightly darker than centre of scales). Distribution . Garra longipinnis is endemic to Oman where it occurs in wadis and falaj irrigation systems draining southwards from Jebel Akhdar and the Samail Gap, from Jebel Kawr and Jebel Shams east to about Lizq and the tributaries of Wadi ‘Andam ( Fig. 2 ). A subterranean population exists in the Al Hoota cave system. Remarks. Garra longipinnis and G. gallagheri are very similar morphologically and are sister species according to our phylogenetic tree reconstruction (Fig. 1). However, their distributions are clearly disjunct, and they exhibit average genetic distances (p-distance) of 2.1% in the analysed COI sequence (min: 1.8%, max: 2.4%). FIGURE 5. Garra longipinnis , from top: holotype, BMNH 1968.10.11.1, 52 mm SL; paratype, BMNH 1968.10.11, 50 mm SL; Saiq, Jebel Akhdar, Oman. FIGURE 6. Garra longipinnis , from top: FSJF 4080, 57 mm SL; 55 mm SL; Saiq, Oman; not preserved, ~60 mm SL; Wadi Falahi, Oman; not preserved, ~55 mm SL; Al-Hoota cave, Oman. Banister & Clarke (1977) based their description of G. longipinnis on eight individuals from Saiq in Oman , collected in 1968 by the British Major Elliott-Legg. Freyhof et al . (2015) previously stated that only fishes identified as G. barreimiae have been found in the areas adjacent to the type locality since. Individuals conforming to the long-finned types of G. longipinnis have never been re-discovered (Gary Feulner, pers. comm.). We collected in the surrounding area of Saiq, but there is very little permanent water today, and the streams were found to be very cold or dry. We found a Garra population inhabiting a spring at the top of a small wadi situated on the outskirts of the old village of Saiq ( Fig. 7 ), a place very likely to have been a point of attraction during the times of Major Elliott-Legg. Banister & Clarke (1977) list their materials as “from Saiq” (p. 135) but give the distribution as “from a wadi near the village of Saiq” (p. 136). The label of the type series in the BMNH reads “Saiq, Jebel Akhdar”, suggesting that the fish were collected directly from Saiq. We suspect that the type series material from Elliott-Legg was collected from the old village of Saiq, at the same site we found. The subterranean water feeding the spring could mitigate low water temperatures in the area and this permanent water body is not only easy to find but was inhabited by fish in the 1960s according to the locals we spoke to. We found no wadis containing fish “near” Saiq, but it cannot be unreservedly assumed that this was the case in the 1960s. The material we collected in Saiq agrees well with the description by Banister & Clarke (1977) and the types of G. longipinnis examined by JF at BMNH. Banister & Clarke (1977) described G. longipinnis as being much less mottled than G. barreimiae , having a midlateral stripe, and that the posterior edge of the scales was darker than the centre. In addition to these character states the meristic characters closely match our findings. Nonetheless, there are almost no differences in meristics between all the Hajar Mountain Garra species (see Pichler et al . 2018 for details). The major difference between the types of G. longipinnis and the fresh material we collected is the length of the fins. In the types of G. longipinnis all fins are elongated (pectoral-fin length 22–32% SL; pelvic-fin length 19–26% SL). While there is some overlap, no individuals with such elongated fins have ever been found again (pectoral-fin length 18–25% SL; pelvic-fin length 15–19% SL according to Pichler et al . 2018 ; 20–26% SL, 17–22% SL in FSJF material). We cannot fully exclude that the long-finned type material of G. longipinnis represented a separate species, which is now likely extinct. DNA extraction and PCR amplification from the type material was unsuccessful, probably due to the method of preservation and age of the material. FIGURE 7. Spring pool close to the old town of Saiq, Oman; probable type locality Garra longipinnis . All the same, the scenario of a species becoming extinct and being replaced by a very similar congener seems unlikely. We suspect that Major Elliott-Legg selected long-finned individuals from a large series of Garra he might have seen in Saiq. In all the Hajar Mountain Garra species the males have considerably longer pectoral and pelvic fins than the females. Fish with hypertrophied fins are rare in nature, but long-finned mutations are very common in domesticated fish. It is possible that long-fin mutations accumulated over time in the small, pond-like spring in Saiq, which today harbours only a few hundred individuals. We consider it more likely that the long-finned individuals represent aberrant specimens of a species that is widespread in Oman outside the Saiq plateau. A blind and unpigmented population of G. longipinnis was described by Banister (1984) from the Al Hoota cave system in northern Oman , and this was studied in detail by Kruckenhauser et al . (2011) and Kirchner et al. (2017) . The latter study includes sound data that the subterranean population is well isolated from its surface counterpart, and there seems to be no introgression of surface individuals into the subterranean population, while subterranean individuals are flushed out from time to time and hybridise with the local surface population.