Description of a rock-dwelling cichlid that re-invaded the sand substrate in Lake Malaŵi, Africa Author Konings, Adrianus F. Cichlid Press, PO Box 13608, El Paso, TX 79913. Author Miller, Torin A. 0000-0003-0840-5814 Penn State University, 432 Forest Resources Building, University Park, Pennsylvania 16802. tmiller9492@gmail.com Author Stauffer, Jay R. Penn State University, 432 Forest Resources Building, University Park, Pennsylvania 16802. & Honorary Research Associate, South African Institute for Aquatic Biodiversity, Makhanda, RSA. text Zootaxa 2024 2024-01-11 5399 2 181 189 http://dx.doi.org/10.11646/zootaxa.5399.2.7 journal article 285347 10.11646/zootaxa.5399.2.7 4fa82fe5-0e6a-47e3-9c95-8ec168e16a3b 1175-5326 10494391 8A52D565-700B-49BC-B4E6-BB234A7345D1 Pseudotropheus likomae , new species Figs. 2 & 3 Pseudotropheus livingstonii ‘likoma’ Ribbink et al . 1983 Pseudotropheus livingstonii Konings 1995 Holotype . PSU 13452 , adult male , 66.3 mm SL, Mbuzi Island , Likoma Island , Lake Malaŵi , Malaŵi , Africa ( S 12 04.677’ , E 34 44.883’ ), 2 July 1991 , J. R. Stauffer Jr. Paratypes . PSU 13367 , 17, 54.7–76.5 mm SL, data as for holotype . Diagnosis. The possession of a large number of small scales on the nape and chest region, an abrupt transition from large flank scales to small chest scales, and the possession of true ocelli aligns P. likomae with the other mbuna species in Lake Malawi . A basic melanin pattern consisting of 5–7 bars (5 or fewer beneath dorsal fin) and two horizontal stripes distinguish adult P. likomae from all other mbuna species since they either have no bars, have greater than five below the dorsal, or lack the horizontal stripes. Juvenile P. likomae , which exhibit the horizontal stripes only vaguely or not at all, are distinguished from Metriaclima ngarae Miller et al . , M. gallireyae Miller et al . , Pseudotropheus crabro Ribbink and Lewis , Chindongo demasoni (Konings) , and C. saulosi (Konings) , mbuna that have five or fewer bars belo w the dorsal fin, by a light tan to hyaline dorsal fin vs. a dorsal fin heavily pigmented with black. Juvenile P. likomae that lack the horizontal elements can be distinguished from Metriaclima lanisticola (Burgess) , another mbuna with five bars below the dorsal, by a shallower body depth (30.4–33.4 %SL vs. 37.0–41.6 %SL). Pseudotropheus likomae appears most closely related to P . livingstonii and like the latter has five or fewer distinct vertical bars below the dorsal fin, but P. livingstonii does not possess at any stage in life horizontal elements in its melanin pattern, while a mid-lateral and a dorso-lateral stripe distinguish mature P. likomae from P. livingstonii and other sand-dwelling mbuna. However, juvenile P. likomae can not reliably be distinguished from juvenile P. livingstonii . Description. Morphometric and meristic data in Table 1 . Medium-sized mbuna, ovoid body (mean BD 32.0% SL) with greatest depth between third to fifth dorsal spine. Dorsal body profile with gradual curve downward posteriorly, more pronounced towards posterior of dorsal fin and beginning of caudal peduncle; ventral body profile slightly convex to almost straight between pelvic fins and base of anal fin with upward curve to caudal peduncle mirroring dorsally. Dorsal head profile straight between snout tip and interorbital, rounded with smooth curve between interorbital and dorsal-fin origin; horizontal eye diameter (mean 34.6% HL) greater than preorbital depth (mean 18.3% HL); eye positioned in anterior half of head with posterior orbit margin on or near vertical median of head; short, straight snout with isognathous jaws; teeth in upper jaw in 4–5 rows and 4–5 rows in lower jaw; teeth in outer row bicuspid (some lateral teeth unicuspid), in inner rows tricuspid. FIGURE 2. Pseudotropheus likomae . Holotype , PSU 13452, adult male, 66.3 mm SL, Mbuzi Island, Likoma Island, Lake Malaŵi , Malaŵi , Africa ( S 12 04.677’ , E 34 44.883’ ), 2 July 1991 , J. R. Stauffer Jr. Dorsal fin with XVII or XVIII (mode XVIII) spines and 8–9 (mode 9) rays. Anal fin with III spines and 7–8 (mode 8) soft rays. First 6 or 7 dorsal-fin spines gradually increasing in length posteriorly with sixth spine about twice as long as first spine; last 12 increasing only slightly in length posteriorly with last spine longest; soft dorsal fin with subacuminate tip, third or fourth ray longest. Anal-fin spines longer posteriorly; fourth or fifth ray longest, length equal to or slightly longer than dorsal fin. Caudal fin subtruncate to emarginate. Pelvic fin reaching to second or third anal-fin spine. Pectoral fin rounded, paddle-shaped, short, to vertical through base of 10th or 11th dorsal-fin spine. Flank scales ctenoid with abrupt change to small scales on breast and belly; 31–33 lateral line scales, cheek with 4–5 rows of small scales. Dorsal fin and anal fin scaleless; tiny scales over proximal ¼ of caudal fin. The angle of the vomer with the parasphenoid of the holotype (PSU 13452) is 67° ( Fig. 3A ). Male coloration: head brown with light blue interorbital bar; cheek white and with blue outline; white opercle with green highlights and dark brown spot; throat white. Flank brown dorsally with six brown bars which fade out on the lower half of the flank; a mid-lateral stripe up to three scales wide where it intersects with a vertical bar to less than one scale wide in between bars; a dorso-lateral stripe between the first and fourth vertical bar with an even thickness of about one scale. Flank, lighter ventrally with blue highlights; breast brown and belly white. Dorsal fin light tan with yellow/orange lappets. Caudal fin membranes light blue. Proximal ¼ of anal fin white and distal ¾ black with white lappets; 1–2 orange ocelli. Pelvic fins with white leading edge; first spine white with black distal ¼. Pectoral fins clear. Female coloration: head brown with light-blue interorbital bar; opercle with green highlights and black spot; throat white. Flank dorsally same as males; very light tan ventrally; breast and belly white; no blue highlights. Juveniles with similar color as females but without horizontal stripes on flank. Distribution. Pseudotropheus likomae is found near Likoma Island and its surrounding cluster islands ( 12°04’40.6”S 34°44’53.0”E ), Malaŵi . The species has not been recorded from Likoma’s sister island, Chizumulu, which is about 10 km west. A transplanted population of P. likomae is present around Thumbi Island West ( 14°01’29.7”S 34°49’21.5”E ). FIGURE 3. A. CT scan of Pseudotropheus likomae , holotype, PSU 13452, adult male, 66.3 mm SL, Likoma Island, Lake Malaŵi, Malaŵi, Africa (S 12 04.677’, E 34 44.883’), 2 July 1991, J. R. Stauffer Jr.; angle of vomer is 67° with parasphenoid. B. A mouthbrooding female of P. likomae at Maingano Island (Likoma, Malawi). C. A male P. likomae at Ndomo Point (Likoma Island). Photographed by Larry Johnson on 11 August 2014. D. A male P. likomae at Thumbi West Island (Malawi). Photographed by Juan Miguel Artigas Azas on 4 November 2008. Etymology. The specific epithet likomae is derived from Likoma Island, the locality where the type specimens were collected. Remarks. The morphometric and meristic data of P. livingstonii from Cape Maclear, Malaŵi are given in Stauffer et al . (2016 : Table 1 , 172–173). There is no overlap of the minimum polygon clusters of the sheared second principal components (SPCA2) of the morphometric data plotted against the first principal components (PC1) of the meristic data for P. likomae and the population of P. livingstonii from Cape Maclear ( Fig. 4 ). The first principal component (size variable) of the morphometric data explained 90% of the observed variance and SPCA2 explained 2.9%. Variables that had the highest loadings on SPCA2 were snout length (-0.52), postorbital head length (0.45), and caudal peduncle length (-0.40). The first principal component of the meristic data explained 40.7% of the variance. Variables with the highest loadings on the first principal components of the meristic data were number of gillrakers on the outer ceratobrancial (0.36), pored scales posterior to the hypural plate (0.36), and number of lateral-line scales (0.35). The holotype of P. livingstonii fell between the polygons formed by P. likomae and P. livingstonii collected from Cape Maclear. We speculate that this intermediate position occurred because the specimen is a juvenile and the location of the type locality is given as Zambesi River (most likely Shire River, the outflow of Lake Malaŵi , which flows into the Zambesi). A similar situation was found when the principal components of this holotype were plotted against that of P. livingstonii from Cape Maclear and Pseudotropheus elegans Trewavas from Chitande ( Stauffer et al . 2016:175 ). Field observations. P. likomae is common in intermediate and sandy habitats and usually found in water less than 15 m deep. Juveniles have been seen inhabiting the shells of Lanistes nyassanus but individuals with the two horizontal stripes in addition to the vertical bars have not been seen associated with empty shells. The latter normally group together in small schools of up to 20 individuals when they feed from the sand or from rocks in their habitat.An individual picks at items on the sand or rock but open mouth combing through the algal matrix on a hard substrate commonly seen in the superficially similar Metriaclima lanisticola (Burgess) has not been observed. Ribbink et al . (1983) examined the stomach contents of 13 individuals caught in Khuyu Bay ( Likoma Island) which contained on average, 68% benthic invertebrates, 20% blue-green algae, 8% plankton, and 4% green algae. When plankton is plentiful P. likomae also feeds from the water column. FIGURE 4. Sheared second principal components (morphometric data) plotted against the first principal components (meristic data) of specimens of Pseudotropheus likomae from Likoma Island and surrounding cluster islands ( + red), specimens of P. livingstonii from Cape Maclear ( x green), and the holotype of P. livingstonii ( o blue). Similar to P. livingstonii , P. likomae appears to lack feeding territories and is usually found in groups of foraging individuals of various sizes. Spawning has not been observed but Ribbink et al . (1983) report the following observation: A large group of territorial fish was found in 4–12 m depth at Ponyemba ( 12°06’02.3”S 34°43’45.1”E ), where parallel ridges of rock run out towards Masimbwe Islet. The territorial fishes were about 2 m apart, were highly aggressive intraspecifically and their defended areas were usually at the base of the rocks and centered around sand-scrape bowers. Approximately 25% of the territorial fishes defended areas which were entirely on the rocky ridges. Associated with these males were numerous females and non-territorial males which remained 2–3 m above the males in the water column, apparently feeding on plankton. Mouthbrooding females ( Fig. 3B ) have occasionally been encountered and always were single, finding shelter between some rocks or resting on the sandy substrate. Fry-guarding females have not been seen and likely the fry are set free once they are mature and abandoned by the female.