Marcusenius multisquamatus Kramer & Wink 2013, sp. nov.

Marcusenius multisquamatus sp. nov.

(Figure 4E)

Holotype: SAIAB 78781 (field no. KUNE24), live SL 20.9 cm, fixed SL 20.2 cm, fixed TL 22.9 cm, male, Namibia: Cunene River: Epupa Falls, Hot Springs Campsite, estimated 300 m upstream from the Falls, 17 ◦ 00 ′ 07 ′′ S, 13 ◦ 14 ′ 57 ′′ E, about 600 m altitude, 15 August 2006, coll. E. Swartz, B. Kramer and L. da Costa at ≤ 1.5 m water depth. Paratypes: SAIAB 78780 (2), SAIAB78792; ZSM 38526 (2), ZSM 38527 (2), size range 10.1–20.2 cm SL, Namibia: Cunene River: Epupa Falls, Hot Springs Campsite, estimated 300 m upstream from the Falls, 17 ◦ 00 ′ 07 ′′ S, 13 ◦ 14 ′ 57 ′′ E, about 600 m altitude, coll. E. Swartz, B. Kramer, and L. da Costa at ≤ 1.5 m water depth, Cunene River water at Hot Springs: Saturday, 12 August 2006, 12.50 h: 19.9 ◦ C, 48 µS cm−1, from 11 Augure 2006– 17 August 2006.

Non-types. SAIAB 78785 (2), SAIAB 78789 (2), ZSM 38528, ZSM 38529 (2), size range 11.6–15.4 cm SL (live), Ruacana Falls, Hippo Pool Campsite, just below the Falls, 17 ◦ 24 ′ 24 ′′ S, 14 ◦ 13’01” E, about 800 m altitude, coll. E. Swartz and B. Kramer, at ≤ 1.5 m water depth, Cunene River water at Hippo Pool: Saturday, 19 August 2004, 10.00 h: 21.1 ◦ C, 45.8 µS cm−1; 20 August 2006, 10.00 h, 19.8 ◦ C, 45.4 µS cm−1; 21 August, 10.18 h, 19.4 ◦ C, 44.2 µS cm−1; from 18 August 2006 to 23 August 2006.

ZSM 41761 (11), specimens R1 – R11, from the Cunene River mouth, 17 ◦ 15.606 ′ S, 11 ◦ 45.892 ′ E, altitude 2 m, 15 December 2009, coll. F.H. van der Bank; ZSM 41762 (2), specimens 49 and 49, 17 ◦ 16.325 ′ S, 11 ◦ 47.177 ′ E, 8 November 2010, coll. S. Voges; ZSM 41765, specimen C113, same place, 17 January 2011, coll. S. Voges; ZSM 41763, specimen Ü7, same place, 13 July 2011, coll. S. Voges; ZSM 41764 (9), specimens Ä110–Ä118, same place, 22 November 2011, coll. S. Voges. The specimens from the Cunene mouth were not studied for EOD.

Samples examined for genetics. DNA samples are stored at the Institute of Pharmacy and Molecular Biotechnology, Heidelberg University (IPMB). IPMB 57459–57469, Namibia: Cunene River Mouth, 17 ◦ 15.606’ S, 11 ◦ 45.892’ E, coll. F.H. van der Bank, 15 December 2009;

IPMB 43971–43974, Namibia: Cunene River: Epupa Falls, 17 ◦ 00’07” S, 13 ◦ 14’57” E, coll. E. Swartz and B. Kramer, 14 August 2006, 17 ◦ 00’07” S, E 013 ◦ 14’57” E; IPMB 43975–43978, as before, but 15 August 2006; IPMB 43993, as before, but 17 August 2006;

IPMB 43980, Namibia: Cunene River: Ruacana Falls, 17 ◦ 24’24” S, 014 ◦ 13’01” E, coll. E. Swartz and B. Kramer, 19 August 2006; IPMB 43986, 43988, as before, but 21 August 2006; IPMB 43990, as before, but 22 August 2006; GenBank accession numbers: (KC 202227 - KC202230; KC202238 - KC 202258).

Cunene River just above the Epupa Falls (Angolan/Namibian border, locality no. 9 on Figure 1.

Body moderately long, prominent mobile and forward-extending mental lobe on lower jaw, median fins set well back with dorsal fin shorter than and originating behind anal fin, depth of caudal peduncle 38% (34–43%) of its length, 24 (23–25) dorsal fin rays, 30 (28–31) anal fin rays, 59 (56–64) scales in lateral series, 13 (12–16) scales around caudal peduncle, HL (head length) 20% (19–21%) of SL, BD (body depth) 29% (27–32%) of SL, LD (dorsal fin length) 19.4% (18.1–22.3%) of SL, LSo (length of snout) 48% (45–50%) HL, LA (anal fin length) 24.2% (22.7–25.3%) of SL, CPL (length of caudal peduncle) 18.3% (16.3–19.9%) of SL. (See also Remarks.)

Head with terminal mouth well in front of eye, mental lobe on lower jaw protruding beyond upper jaw. Head and body dorsolaterally compressed. Dorsal fin situated about two-thirds of standard length from snout, obliquely oriented, anteriorly higher and posteriorly lower, distal margin sometimes only slightly crescent-shaped with anterior two or three rays longer than posterior rays, number of rays 23 (n = 5), 24 (n = 4), 25 (n = 6); anal fin opposite dorsal fin with distinctly more anterior origin, obliquely oriented, anteriorly lower and posteriorly higher, anterior rays longer than posterior ones, especially in males where they also appear stronger and often darkened, distal margin crescent-shaped (in males only posterior to rounded, elongated anterior part of fin), number of rays 28 (n = 1), 29 (n = 4), 30 (n = 7), 31 (n = 3). Scales cycloid with reticulate striae, scales extending anteriorly to operculum and pectoral fins (beyond pelvic fins). Scales on caudal peduncle circumference, 12 (n = 5), 13 (n = 4), 14 (n = 5), 16 (n = 1) Caudal peduncle relatively deep, subcylindrical entire length, usually 18.3% (16.3–19.9%) in SL (Table A1, in Appendix A). Electric organ discharge biphasic with weak pre-potential (Figure 5). Males approaching sexual maturity develop a kink in the base of the anal fin (e.g. Figure 4C) that is absent in juveniles and females where the anal fin base is straight. Colour in life: brownish grey with many distinct dark-brown blotches, except on head and belly, purple hue depending on the angle of light incidence, paired fins light and transparent.

Medium brown, with darker, irregular blotches.

The Cunene is a major, perennial and independent river that arises from the Angolan central highlands of Bié and flows southward towards the Namibian border, shortly before it turns west and breaches the coastal mountain ranges (Zebra and Baynes Mountains) to drain into the Atlantic. In the section between Ruacana Falls and Epupa Falls, water level was regulated by a hydroelectric company (NamPower) at Ruacana Falls. The Ruacana Falls were bare rock and completely dry, apparently because the water dammed above Ruacana Falls (Calueque Dam) was all fed into the hydroelectric power turbines. When the water level below the dam was kept high, fishing with gill nets and other methods generally proved unproductive. The Epupa Falls consist of a main fall with many lesser falls beside this over a wide front, and mormyrid EODs were demonstrated with an electro-acoustic, custom-built “fish detector” also below the Falls although the fish were not caught. Although we were warned of a high incidence of crocodiles we saw only a few, and no hippopotami. River borders were covered mainly by dense semi-aquatic shrubs at Epupa Falls, and dense reed beds also with shrubs at Ruacana Falls. Palm trees (Makalani palms) were common at Epupa, much less so at Ruacana where dicotyledonous trees dominated.

At present known only from the lower Cunene River, from just below Ruacana Falls to the Cunene mouth. This river section forms the Angolan/Namibian border.

Closest relationships are assumed with M. altisambesi to the east of M. multisquamatus sp. nov. on the basis of morphological similarity, EODs and genetics.

Marcusenius multisquamatus sp. nov. refers to the highest number of lateral line scales among the different forms of southern African bulldog fish (excluding the three Mossamedes / Cunene specimens (BMNH 1907.6.29.231–233) from any location on the Cunene up to 300 km north of the Angolan/Namibian border).

Compared with the M. angolensis holotype, M. multisquamatus sp. nov. specimens had lower counts in nD, no. of dorsal fin rays (maximum, 25 in the latter versus 26 in the former, that is, no overlap) and nA, number of anal fin rays (maximum, 31 versus 33), shorter LA, anal fin length (maximum, 0.253 versus 0.258 of SL) and PDL, predorsal length (maximum, 0.665 versus 0.674 of SL), smaller ratio HL/Na, head length/separation of nares (maximum, 15.52 versus 15.96), but a greater BD, body depth (minimum 0.271 versus 0.266 of SL).

When compared with the other Marcusenius species within the Okavango – Kwando– Zambezi System, M. multisquamatus sp. nov. is characterized by a specific morphology and EOD in multivariate analysis, specific bands in genomic ISSR fingerprinting, and as a monophyletic taxon in mitochondrial DNA (mtDNA) cytochrome b analysis. To identify a specimen in hand it is best to rely on several characters in combination to exclude mistakes due to outliers. The 90th percentile of the distribution of HL (measured as HL/SL, head length to standard length) of M. multisquamatus sp. nov. specimens, is shorter than the 10th percentile for M. macrolepidotus (together with M. angolensis, shortest HL of all). The BD/SL ratio (body depth to standard length) of M. multisquamatus sp. nov. overlaps with that of M. macrolepidotus by less than one quartile; the same holds true for the distributions of nA (no. of anal fin rays), nD (no. of dorsal fin rays), LD/SL (ratio of dorsal fin length to standard length) in which the means or medians are greater for M. multisquamatus sp. nov., and SPc, number of scales around caudal peduncle in which the median for M. multisquamatus sp. nov. is smaller. The EOD of M. multisquamatus sp. nov. has a leading head-negativity of miniature amplitude that is usually not present in the M. macrolepidotus EOD. Marcusenius multisquamatus sp. nov. and M. macrolepidotus are clearly differentiated in ISSR bands 2, 6 and 8 (Table 4).

There is less than a 10% overlap of distributions between the greater LSo/HL (length of snout to head length) of M. multisquamatus sp. nov. specimens compared with that of M. altisambesi from the Upper Zambezi, and less than 25% overlap for M. altisambesi from the Okavango. Also, there is less than one quartile overlap for LA/SL (anal fin length to SL), LSc/HL (length of snout to head length) and CL/HL (chin length to head length) of M. altisambesi from the Upper Zambezi compared with M. multisquamatus sp. nov. and less than one quartile overlap for lower SLS (no. of lateral line scales) of M. altisambesi from the Okavango compared with M. multisquamatus sp. nov. Okavango bulldog fish are distinguished by an SPc (no. of scales around caudal peduncle) of exclusively 12 (median 12, same median for Upper Zambezi bulldog fish) whereas SPc ranges from 12–16 in M. multisquamatus sp. nov. (median, 13). ISSR band 2 is specific for M. multisquamatus sp. nov. and band 6 for M. altisambesi.