2019-03-14

Lamprologini...part 1


An account of Lamprologini, including Neolamprologus sp. “Mwila”, a new shell-dwelling species from Mwila Island in the Kipili Archipelago, the facts about N. sp. “Eseki”, the whereabouts of Kinyamkolo, and much more (part 1 of 3)

Part 1 – Part 2Part 3

Partitioned into 32 sections, this account of Lamprologini includes facts about Neolamprologus sp. “Eseki” (§ 4–7), an introduction to N. sp. “Mwila” (§ 10–11), the rise of incipient species following divergent ecological adaptation, exemplified by the Telmatochromis temporalis species complex (§ 12) and Altolamprologus genus (§ 15), high and low lake stands in Lake Tanganyika (§ 16), the age of the species and variants in the Kipili Archipelago, including N. sp. “Mwila” and Lepidiolamprologus kamambae (§ 17–18), a species without a name (§ 21), the Lake Tanganyika expedition leader J. E. S. Moore (§ 22), the perception of N. modestus as a widespread species (§ 23), Lamprologus modestus nyassae (§ 25), an introduction to the N. modestus species complex (§ 30), and more. Regarding J. E. S. Moore and the type localities of N. modestus, the last section (§ 32), which is further partitioned into 27 subsections (§ 32.1–32.27), is an in-depth study of the historical locations Kinyamkolo and Mbity Rocks, including their whereabouts and linguistic meaning.


Fig. 1. Mwila Island was our team’s first dive site in the Kipili Archipelago in 1991, searching for Julidochromis marksmithi (Lundblad & Karlsson, 1993a; 1993b). At this small island, we found a new lamprologin species at a depth of 15 metres in April 2008, referred to as Neolamprologus sp. “Mwila”. See more in § 10.
Fig. 2. Neolamprologus sp. “Mwila” is distinguished from N. sp. “Eseki” by, among other things, the larger eye, absence of a yellow eye ring, its smaller size (5–7 vs. 10–11), and possibly a less indented caudal fin. Furthermore, N. sp. “Mwila” exhibits a different breeding behaviour and prefers another biotope (shell bed vs. sand with scattered rocks and stones). The picture was captured in 2008 at the eastern side of Mwila Island at a depth of 15 metres and shows a fully grown male. For more differentiating features, see § 11.

Contents
32. Where are Moore’s Kinyamkolo and Mbity Rocks?

Part 3
Appendix
Appendix 1 (fish species names derived from J. E. S. Moore)
Appendix 2 (the grammar behind the species epithets ‘moorii’ and ‘moorei’ as in Tropheus moorii and Nudospongilla moorei)
Appendix 3 (additional species associated with J. E. S. Moore)
Appendix 4 (collection locations, fish species, and number of specimens collected by J. E. S. Moore in Lake Tanganyika during his expeditions in 1895–1896 and 1899–1900)
Appendix 5 (Moore’s collection comprising 52 fishes from Lake Malawi obtained between 11 July and 15 August 1899 during his second expedition, including the misidentification of species)
Appendix 6 (a directory of geographical locations mentioned in the study with links to Google Maps)

Citation

References


1. Introduction to Lamprologini
Being one of the most diverse cichlid tribes regarding morphology, ecology, and behaviour, the Lamprologini is mainly found in Lake Tanganyika, Malagarasi River, and the Congo River drainage. In Lake Tanganyika, lamprologin species have colonised many of the lacustrine habitats, but are primarily found in the littoral zone where most of the fish species are lamprologins. Generally, this zone may be subdivided into three ecological categories with lamprologins living in sandy, muddy, and rocky habitats. Some species are large roaming predators, some are medium-size invertebrate pickers, while others are small enough to dwell inside empty gastropod shells. Furthermore, there is variation in nesting sites involving sandy pits and rocky caves, interspecific variation in clutch size [up to 3000 eggs per spawn in Neolamprologus cunningtoni and only about 12 in N. savoryi (Sturmbauer et al., 1994: 692)], complex patterns of parental care, various strategies in mating including serial and harem polygamy, alternative male reproductive phenotypes, and sneaker fertilisation (Nagoshi, 1983). In addition, the tribe Lamprologini comprises some of the very largest and smallest cichlid species in Lake Tanganyika: Lepidiolamprologus profundicola (>300 mm) and N. multifasciatus (30–40 mm).

Fig. 3. Lepidiolamprologus profundicola at Segunga. This species is the largest member of the tribe Lamprologini. Apart from its large size, it is easily recognised by its proportionally small eye. Its cautious behaviour makes it challenging to photograph in the lake.
Fig. 4. A sampled specimen of Lepidiolamprologus profundicola from Kansombo, total length 33 cm, NRM 61570. This species may reach a total length of about 35 cm.
Fig. 5. Neolamprologus multifasciatus at Mpando Point. Along with the similar looking N. similis, this species is the smallest cichlid in Lake Tanganyika (as well as the world’s smallest). In Tanzania, N. multifasciatus is found from Kala Bay south to Kalambo River, on the border between Tanzania and Zambia, whereas N. similis is found from the Malagarasi River south to Kekese, just north of Ikola. The geographical distribution of these species was established by the authors in 2008 (Karlsson & Karlsson, 2015b). In Tanzania, N. multifasciatus is commonly found among pebbles and rubble, while N. similis is mostly found in biotopes of empty shells. N. multifasciatus is also found in Zambia and N. similis in DR Congo.
Fig. 6. Neolamprologus savoryi at Mlowa Point, mainland Kipili. This species has the lowest brood size of the cichlids in the Lamprologini tribe. Often, only a few single fry are produced by a pair.

2. Current and future taxonomical status of the lamprologin species of Lake Tanganyika
In total, 112 Lamprologus-like species and subspecies have been formally introduced, with the taxonomically oldest, Lamprologus congoensis (Schilthuis, 1891: 85, Plate 6, Fig. 1), from Malepo Poll (previously Stanley Pool) in the Congo River, and the youngest, Julidochromis marksmithi (Burgess, 2014: 40), from Kerenge Island in Lake Tanganyika. Of these 112, eight are endemic to the Congo River drainage (Schelly & Stiassny, 2004: 37), and one, N. devosi (Schelly et al., 2003: 2), is known only from Malagarasi River in Tanzania. Two species were incorrectly introduced, L. ocellatus (Poll, 1952a: 15) and L. savoryi elongatus (Trewavas & Poll, 1952: 5); their names were preoccupied by L. ocellatus (Steindachner, 1909a: 402; Boulenger, 1915: 462) and L. elongatus (Boulenger, 1898a: 494; 1898b: 9), and formally replaced with L. kungweensis (Poll, 1956: 570) and N. brichardi (Poll, 1974: 109; 1986: 64). Furthermore, L. steindachneri (Boulenger, 1915: 461, 471) is an invalid replacement name and junior synonym for J. elongatus (Steindachner, 1909a: 403), see more below (§ 25). Of the remaining 100 lamprologin species and subspecies of Lake Tanganyika, 15 are frequently considered invalid due to synonymy. However, pending thorough re-examinations of voucher and live wild specimens, clearer diagnoses may be devised rendering some of these to be released from synonymy, including L. pleurostigma (Boulenger, 1914: 443) synonymised with Lepidiolamprologus attenuatus (Steindachner, 1909b: 445) by Poll (1956: 540; 1986: 53), Lamprologus lestradei (Poll, 1943: 317) synonymised with L. ocellatus (Steindachner, 1909a: 402) by Poll (1956: 537), and Telmatochromis burgeoni (Poll, 1942: 357) synonymised with T. temporalis (Boulenger, 1898a: 495; 1898b: 11) by Konings (1998a: 98), and Hanssens and Snoeks (2001: 650; 2003: 594).
Recently, some junior synonyms were suggested to be reinstated, such as J. macrolepis (Borodin, 1931: 51) synonymised with T. dhonti (Boulenger, 1919: 19) by Hanssens and Snoeks (2001: 652), but revived as T. macrolepis by Konings (2015a: 218), N. brichardi (Poll, 1974: 109) synonymised with N. pulcher (Trewavas & Poll, 1952: 6) by a series of authors, including Konings (2015a: 119), but argued to be a distinct species by Karlsson and Karlsson (2017a; 2017b: 66), and L. brevianalis (Boulenger, 1906: 555) synonymised with L. tetracanthus (Boulenger, 1899c: 118) by Poll (1946: 339), but treated as a valid species, referred to as N. brevianalis, by Karlsson and Karlsson (2018a: 106), who suggested some additional colour-related diagnostics to the species.
In addition to these 85–100 lacustrine and nine riverine valid lamprologin species, many more may be introduced pending further field observations and taxonomical research. Lake Tanganyika and the Congo River, as well as their larger tributaries, presumably harbour many yet unknown lamprologin species not easily discovered due to their cryptic morphology and behaviour. Recently, a very cryptic Julidochromis species, referred to as J. sp. “Transcriptus Tanzania”, was introduced (Karlsson & Karlsson, 2018a; 2018b; 2018c). In a riverine Lamprologus review, Schelly and Stiassny (2004: 3) “found that over 50% of Congo River Lamprologus material housed in museum collections was misidentified, including some of the vouchers representing specimens used in various molecular analyses”. Furthermore, peculiar distribution data suggest “that perhaps much of the Congo River lamprologine diversity remains to be discovered” (Shelly & Stiassny, 2004: 37). Based on similar problems in systematic research regarding the cichlids of Lakes Tanganyika, Malawi, and Victoria, “about 1000 species or more are still awaiting scientific description” (Snoeks, 2000: 17). The systematic complexity of Lake Tanganyika cichlids appears to be grossly underestimated (Snoeks, 2000: 25). Although it was reported when only 70 formal Lake Tanganyika cichlids were known, 21 of which were lamprologins, the following statement appears just as relevant today: “No doubt many more species remain to be discovered” (Boulenger, 1906: 539, 541–542).

Fig. 7. Neolamprologus cunningtoni is the member of Lamprologini with the largest number of fry produced in one spawn (often more than 3000). The photo was taken at Kungwe Point (about 13 km northwest of Halembe) at a depth of 15 metres. Repeated low water stands in Lake Tanganyika have resulted in geographical isolation and allopatric speciation, with many cichlids in the lake having at least a northern and southern representative, which have been given the status of distinct species, e.g. Cyphotilapia frontosa and C. gibberosa. The northern representative of N. cunningtoni may represent a separate species.
Fig. 8. Neolamprologus cunningtoni is the second largest member of Lamprologini, reaching a total length of nearly 30 cm. The photo shows a sampled specimen from Udachi with a total length of 27 cm, NRM 59523. Obviously, N. cunningtoni shares several features with Boulengerochromis microlepis, including colouration, general body shape, and a large number of eggs per spawn. Furthermore, both species are piscivores and breed in sandy habitats supplemented with rocks. Based on its overall appearance, N. cunningtoni may be regarded as a small version of B. microlepis.
Fig. 9. Boulengerochromis microlepis at a depth of 15 metres at Katondo Point, Cape Mpimbwe. Initially, B. microlepis was generically assigned to Tilapia (Boulenger, 1899a: 94), but soon thereafter moved to Paratilapia (Boulenger 1915: 370). Throughout much of the 1900s, Boulengerochromis was regarded as a close relative of Tilapia and included in Poll’s definition (1986: 30) of the tribe Tilapiini, until Takahashi (2003: 377) erected a new and monotypic tribe for it, referred to as Boulengerochromini, a revision later confirmed by DNA data (Meyer et al., 2015: 57). Besides molecular characters, Boulengerochromini is distinguished by several morphological features, including the infraorbital bones, oral teeth, caudal fin shape, and lateral scales (see details in Takahashi, 2003). Interestingly, based on the presence of gill-raker denticulations (with fine teeth), Takahashi (2003: 372) found similarities with Boulengerochromini and Lamprologini. Nuclear DNA data suggest that B. microlepis is part of the lacustrine assemblage and represents some of the oldest cichlid lineages in Lake Tanganyika (Koblmüller et al., 2014), seemingly with a similar age as Lamprologini. However, Boulengerochromini is currently positioned in a phylogenetic tree outside of a clade comprising Lamprologini and others (Meyer et al., 2015: 69).
Fig. 10. In the 1950s, a new species found in Kungwe Bay was incorrectly introduced as Lamprologus ocellatus (Poll, 1952a: 15). The name was preoccupied by L. ocellatus (Steindachner, 1909a: 402; Boulenger, 1915: 462) and the new species formally renamed L. kungweensis (Poll, 1956: 570). The picture shows L. ocellatus at Cape Kabogo, at a depth of 15 metres (exported by us in the 1990s as “Yellow – Cape Kabogo”).
Fig. 11. A newly discovered variant of Lamprologus ocellatus near its buried shell at a depth of 20 metres, found during our expedition in 2007/2008. With an intense copper coloured dorsal fin and bright yellow anal fin, the vernacular name “Red Gold” fits its characteristics perfectly well. Currently, L. ocellatus comprises variants found at Kigoma (its type locality) and throughout much of the lake, but future studies on L. ocellatus may remove the central and southern variants from the species.

3. Possible novel lamprologin genera
Lake Tanganyika not only has cryptic species that remain to be discovered, but also cryptic genera. The latest Lake Tanganyika genus that was erected is the haplochromin Interochromis (Yamaoka et al., 1998: 385), while that of Lamprologini is Altolamprologus (Poll, 1986: 66). The additional lamprologin genera are Neolamprologus, Paleolamprologus, and Variabilichromis (Colombé & Allgayer, 1985), Chalinochromis (Poll, 1974), Lepidiolamprologus (Pellegrin, 1904), Julidochromis and Telmatochromis (Boulenger, 1898a; 1898b), and Lamprologus (Schilthuis, 1891). Currently, there is “relatively high species richness of Neolamprologus”, which is thought to be “largely a consequence of taxonomists using this genus as a default repository for new species” (Day et al., 2007: 639). However, the introduction of about 20 new Neolamprologus species since the erection of the genus all perfectly conform to Poll’s (1986) definition of the genus (see more in § 20). The default repository status of Neolamprologus appears to relate mainly to tentatively recognised new species and temporarily reassignments of old lamprologin species, which have not been thoroughly investigated.
In contrast to the proven monophyletic nature of Lamprologini, it is widely recognised that some lamprologin genera, including Lamprologus and Neolamprologus, are compounds of superficially similar species, rendering them in need of taxonomical review (e.g. Stiassny, 1997; Day et al., 2007). Apparently, current generic classification poorly represents current research in evolutionary history. For example, the groups of riverine and lacustrine Lamprologus species are not resolved as their closest relatives (Sturmbauer et al., 2010). Nonetheless, the lamprologin-rich ichthyofauna of Lake Tanganyika includes some seemingly very uniform groups of species, each of which may share a recent common ancestor. Possibly, they constitute complete and natural groups of closely related species, which would become perfect genera. Based on molecular traits and brood-care helping behaviour, Sturmbauer et al. (2010) suggested the group of sand-dwelling and grossly N. modestus-like species to be a future lamprologin separate genus, which would include N. modestus (Boulenger, 1898a: 494; 1898b: 8), N. tetracanthus (Boulenger, 1899c: 118), N. cunningtoni (Boulenger, 1906: 557), N. mondabu (Boulenger, 1906: 557), N. petricola (Poll, 1949: 37), and N. christyi (Trewavas & Poll, 1952: 10). As N. tetracanthus is the type species for Neolamprologus, this future genus would not need a new name, only a narrower diagnosis which excludes non desirable species, i.e., those which are not N. modestus-like. Such a future genus of sand-dwelling species with mostly greyish colouration may include additional species, as well as species complexes which are currently known only by provisional names and diagnoses, including N. sp. “Eseki”, N. sp. “Mwila”, and the N. modestus species complex, see below (§ 4, 10, 30).

Fig. 12. Neolamprologus christyi at Chiloelo Point, just north of Kalepa. In Tanzania, this species is found at the southern islands in the Kipili Archipelago south to Kalambo River, on the border with Zambia.

4. The discovery of Neolamprologus sp. “Eseki” in 1990
In March 1990, at the southern rocky shore of Udachi, about two kilometres (km) south of Kabwe Village, Tanzania, we found an N. mondabu-like species in a shallow habitat comprising a mixture of rocks and sand. The collected specimens had an indented (concave) caudal fin with slightly pointed upper and lower corners, and a mostly greyish brown body. Initially, the species was referred to as N. mondabu (Boulenger, 1906: 557, Pl. 36, Fig. 3; 1915: 470), and some of the first exported specimens were shipped under this name to Åleds Akvarium in Sweden (22 May 1990) and Neil Hardy, Aquatica Ltd in Carshalton, London (5 July 1990). A few months later, we altered the name to N. sp. “Eseki”, a decision which was based on the fact that we found a second N. mondabu-like species living in the same habitat, a species we mistakenly thought was the true N. mondabu. Individuals of this second species mainly had a pearly greyish body and fins, and a truncate caudal fin with rounded corners. By this time, it was widely thought that N. mondabu usually has rounded caudal-fin corners, a perception derived from Poll’s (1956: 484) erroneous synonymisation of N. mondabu with N. modestus, which was indeed corrected later (Poll, 1978: 744), therefore, we applied the name N. mondabu to the pearly greyish species with rounded caudal-fin corners, and the newly invented name, N. sp. “Eseki”, to that with pointed corners. Similar misidentifications are illustrated in, for example, Konings and Dieckhoff (1992: 35, Fig. bottom); see more about misidentifications in § 23, 24, 26, 27, 28, 29. Subsequently, both species were found north and south of Udachi, at Kansombo and Cape Mpimbwe, respectively. Today, the species from Cape Mpimbwe and the surrounding area having rounded caudal-fin corners is probably best viewed as distinct from N. mondabu, and perhaps indicated as N. modestus, as suggested by Konings (2015a: 237), or, alternatively, N. sp. “Modestus Mpimbwe”, as in this article, see more below (§ 29). However, N. sp. “Eseki” is still referred to by the original name, for example, see Sturmbauer et al. (2010) and Breman et al. (2016). The name N. sp. “Eseki” was officially introduced with a consignment of 11 specimens of the species to Åleds Akvarium, Sweden (9 November 1990). One of us (Magnus) wrote in the accompanying shipping documents that “N. sp. ‘Eseki’ is similar to N. mondabu [N. sp. ‘Modestus Mpimbwe’], but adult fish are brown, eyes blue-green, fins yellowish, and caudal fin slightly emarginate. The size of the fish is only 7 cm”. This is the first unofficial provisional diagnosis of N. sp. “Eseki”, see an updated list of characteristics below (§ 7). Additional similar species included in that particular shipment were ‘N. christyi from Kasanga and N. modestus from northern Kasanga Bay, Kalambo, and Kala, the former being a solid black variant and the latter with conspicuous yellow margins in the dorsal and caudal fins’ (paraphrase of the said shipping documents). For the record, specimens of the Kalambo variant had bright yellow pectoral fins, perhaps the most typical feature of the true N. modestus.

Fig. 13. Udachi Point is a large rocky area, mostly submerged, stretching over 1.7 km, from Kabwe Bay to African Diving Fish Station. Here, Neolamprologus sp. “Eseki” was discovered in 1990. Another well-known fish discovered the same year in this rocky area is Tropheus ‘Kushangaza’, a mutant coloured form of T. sp. “Mpimbwe” (African Diving Ltd, 1990). Since 1989, when we first settled in the area, Udachi Point has been a stronghold for crocodiles and several accidents have occurred over the years, both to us and to nearby villagers, a dozen of which were fatal.
Fig. 14. Neolamprologus sp. “Eseki” was discovered by our team at the village Udachi in the beginning of 1990 (§ 4). The name, Eseki, is made up to sound interesting and be easy to remember, and has no linguistic meaning (§ 7). The first exports of this species were to Sweden and the United Kingdom in that same year (22 May and 5 July, respectively). In the picture is an individual at Udachi at a depth of four metres.
Fig. 15. Kampemba Point is located at the southern side of Korongwe Bay (Cape Mpimbwe at the northern, 4.5 km across the bay). Here, we first noticed the bright orange juveniles of Tropheus sp. “Kipili” in 1989. The Korongwe Bay area is a hybrid zone between the two adjacent Tropheus species; T. sp. “Mpimbwe” and T. sp. “Kipili”. In the beginning of the 1990s, we used to come here to collect the goby cichlid, Eretmodus cyanostictus, as Kampemba Point is the northernmost location where this goby cichlid is found in southern Tanzania. At the other side of the bay, only E. marksmithi is found. Deep inside the bay, at Kafwila, the two species hybridise.
Fig. 16. Neolamprologus sp. “Modestus Mpimbwe” at Kampemba, the southernmost population. This species is characterised by its azure blue fins and was first exported by us in November 1990. It lives sympatrically with N. sp. “Eseki”.
Fig. 17. Video clip of Neolamprologus sp. “Eseki”. This species is found from Isonga south to the northern islands of the Kipili Archipelago, a distribution range established by the authors in 2008 (§ 5). It was first discovered at the rocky shore of Udachi, about two kilometres south of Kabwe early in 1990 (§ 4). N. sp. “Eseki” feeds of shrimps, copepods, ostracods, snails, mussels, and/or insects (§ 6). Its closest relatives appear to be N. mondabu and N. christyi (§ 31).
Fig. 18 (left). N. sp. “Modestus Mpimbwe”, Udachi. Fig. 19 (right). N. sp. “Eseki”, Udachi.
Fig. 20 (left). N. sp. “Modestus Mpimbwe”, Katondo Point, Cape Mpimbwe. Fig. 21 (right). N. sp. “Eseki”, Katondo Point, Cape Mpimbwe.

5. Neolamprologus sp. “Eseki”: geographical distribution
Following our initial discovery at Udachi in 1990, N. sp. “Eseki” was found at additional locations. The Udachi variant was described as brown with yellowish fins and an intermediate of a ‘princess’ and N. modestus (Zadenius, 1991a: 38). The first exported N. sp. “Eseki” to the USA was reported as “newly discovered from southern Tanzania” (DeMason, 1992: 27). Subsequently, N. sp. “Eseki” became known in the aquarium hobby (e.g. Johansson, 1994a: 48; Smith, 1998: 12; 2007: 48; Konings, 2005: 116–117). At Nkondwe Island (Fig. 28) in the Kipili Archipelago, we found a slightly different colour variant, which was recorded by the importers as to have more yellowish eyes and fins (Zadenius, 1991b: 44). In an inventory diving log, Lundblad and Karlsson (1993b: 314) described the natural colouration and behaviour of the species at Kerenge Island: “Neolamprologus sp. ‘Eseki’ is found at the sand. Shaped like N. mondabu [N. sp. “Modestus Mpimbwe”], but slenderer. The body is grey-brown with yellow eyes. The fish species reaches a size of 7 cm”. This is the first official diagnosis of N. sp. “Eseki”, short and incomplete with a size measurement that needs to be adjusted (see characteristics below, § 7), yet good enough for the purpose of applying a provisional name. N. sp. “Eseki” was featured in yet another diving log, its territory at Katondo, Cape Mpimbwe, described as comprising small sandy pits adjacent to rocks and cliffs (Karlsson, 1998: 33; 2002; 2004).
Identifying the geographical distribution ranges of cichlid species and variants in Lake Tanganyika is no easy task. It requires time, usually weeks or months, money for renting or purchasing boats, engines, fuel, food, paying staff and transport, handling local bureaucracy, etc., and not least, mental and physical efforts. Referred to as “the kingdom of the Cichlidae” (Boulenger, 1901c: 142) and “the most mysterious lake in the world” (Moore, 1901b: 76–79), Lake Tanganyika is a vast inland freshwater sea. Its enormous size is hard to imagine when leafing through a book or magazine. Indeed, in good weather, it takes about 45 minutes for a regular motorised ship just to travel between the inside of a bay (Udachi, Utinta Bay) and its outermost point (Mpimbwe Point, Cape Mpimbwe). During most of the year, the lake usually changes abruptly in the afternoon becoming very windy and wavy. In bad weather, waves can reach five metres, with hundreds of local ships and small boats sunk annually due to sudden and unforeseen storms and rains. Crossing the lake is especially dangerous; in May 2008, a former member of our team crossed from Kabwe, Tanzania, to Moba, DR Congo and in open water, the ship sunk in a storm, drowning all 55 passengers except for three men, one of whom was the captain of the ship, a Moba resident. Our associate was not one of the other two survivors.
Along the coasts of Tanzania and DR Congo, only a few natural harbours exist, while man-made harbours are more or less completely missing. Travelling safely requires thorough planning, usually only mornings and forenoons are safe. Some accounts of the hardships of travelling on the lake are found in Moore (1901b: 94–97, 135), Büscher (1983: 161; 1987; 1998c), Karlsson and Lundblad (1991; 1992; 1994), Konings (1998b: 15), as well as Karlsson and Karlsson (2013: 7–8). “It must be nearly impossible for anyone who has not visited the African lakes to realize their huge size and oceanic character” (Moore, 1898a: 26). “[I]f there is anything I detest, both in principle and in practice, it is crossing a wide African water in dug-outs. [...] I don’t know how Fergusson felt, but I went across that six or seven miles of open water in terror of my life” (Moore, 1901b: 166).
Almost 20 years after the discovery of N. sp. “Eseki”, we conducted a hard-target search and identified the geographical distribution of the species to encompass the area between Isonga and the northern Kipili Archipelago, the latter which includes the islands of Kamamba, Kasisi, Kerenge, Mwila, and Nkondwe. These islands, the five northernmost of the archipelago, partly harbour a unique set of cichlid species and variants distinct from that of the three southernmost islands, i.e., Lupita, Mvuna, and Ulwile. For example, at the five northern islands, Ophthalmotilapia boops has a neon blue stripe across its body (vs. an all blackish body at the three southern islands), O. sp. “Whitecap” is all blackish (vs. bluish orange), O. nasuta has mainly a dark brown or blackish trunk (vs. yellow), Tropheus sp. “Kipili” has yellow (vs. grey) pectoral fins, Petrochromis famula has dark reddish brown (vs. blue) fins, a Lepidiolamprologus elongatus-like species is black and yellowish (vs. black and white), etc. Furthermore, Chalinochromis cyanophleps and N. furcifer exist only at the three southern islands, while Lepidiolamprologus kamambae has currently only been found at two of the northern islands (§ 18). Regarding a Cyprichromis coloratus-like species, commonly referred to as C. sp. “Jumbo Leptosoma” (cf. Takahashi, 2016), we have observed four distinct male colour morphs at the three southern islands, but only three such morphs at the five northern islands (cf. Karlsson & Karlsson, 2017k). This phenomenon of distribution has been described in more detail elsewhere (Karlsson & Karlsson, 2013; 2014a; 2015a; 2017a; 2017b; 2017c; 2017d; 2017e; 2017f). On a side note regarding the O. boops with a neon blue stripe across its body (the so-called ‘Neon Stripe’), this variant is more or less extinct, or at least critically endangered, at both Mwila (Fig. 1) and Nkondwe Islands, but still thriving at Kamamba, Kasisi, and Kerenge Islands.
In 2008, during our Kigoma-Kalambo expeditions, we visited the Kipili area and made several investigative scuba dives around the islands to clarify whether N. sp. “Eseki” and the similar N. christyi are sympatric (coexist in a certain place) or allopatric (geographically separated) in that area. At the five northernmost islands, we found only N. sp. “Eseki”, while N. christyi was found only at the three southernmost islands, in other words, there was no sympatry between the two species (e.g. Karlsson & Karlsson, 2017a; 2017b; 2017c). This was unsurprising as many other species and variants show the same pattern of distribution (see the nine references above). However, our observations contradicted those of Konings (2015a: 242), who reported seeing both species at Mvuna Island and Nkondwe Island. Possibly, Konings did not actually see individuals of both species at both islands but has rather based his report on two accompanying under-water photos taken by J. M. Artigas, captioned “Neolamprologus mondabu [N. sp. ‘Eseki’] at Mvuna Island” and “Neolamprologus christyi at Nkondwe Island” (Konings, 2015a: 239, Fig. bottom; 242, Fig. top). Presumably, the photographer has mistakenly switched the locality labels, ‘Mvuna’ should be ‘Nkondwe’, and vice versa. Regarding Nkondwe Island and mislabelling, Konings (2015b) claimed that the report of Kullander et al. (2014) concerning the geographical distribution of N. furcifer is incorrect, that, contrary to the report, Nkondwe Island harbours a population of the said species. This appears to be based on yet another labelling mistake by Konings and the photographer (J. M. Artigas). While N. furcifer does not exist at Nkondwe Island, it does exist at Mvuna Island (e.g. Karlsson & Karlsson, 2017a; 2017b; 2017c), so it appears that the locality label has been switched again; ‘Nkondwe’ should be ‘Mvuna’. Indeed, N. furcifer only exists at the three southernmost islands in the Kipili Archipelago, i.e., Lupita, Mvuna, and Ulwile, not at any of the five northernmost, i.e., Kamamba, Kasisi, Kerenge, Mwila, or Nkondwe. Furthermore, Konings (2015b) also stated, as a further contradiction to Kullander et al. (2014), that N. furcifer probably exists in the area north of Kipili and farther north. Even this statement is incorrect, not to mention remarkable, as it gives priority to speculations before empirical facts. While N. furcifer exists at Kampemba Point (Fig. 15; south of Cape Mpimbwe) and the three southern islands of Kipili, it does not exist anywhere between these localities, only N. timidus does. For the record, the similar N. timidus (Kullander et al., 2014) exists at all eight islands of Kipili as well as along the adjacent mainland, from Cape Mpimbwe southward via the Kipili area to at least Kisi Island, near Ninde. However, N. timidus and N. furcifer (Ulwile) in the Kipili area are easy to distinguish due to the rounded caudal-fin lobes of the latter.

Fig. 22. A juvenile, size 5 cm, of Neolamprologus sp. “Eseki” at the village Udachi. Inspired by the beautiful colours of the juvenile N. sp. “Cygnus” (another newly discovered cichlid in 1990), we bred N. sp. “Eseki” in our fish house in Dar es Salaam immediately after its discovery to determine the colouration of the juveniles. The result revealed no surprises; the fry looked similar to their parents.
Fig. 23. Neolamprologus sp. “Eseki” at Udachi, depth five metres. The body colouration of this species is beige to brown. The first unofficial diagnosis of it was included in the shipping documents to Sweden on 9 November 1990: “N. sp. ‘Eseki’ is similar to N. mondabu [N. sp. ‘Modestus Mpimbwe’, § 29], but adult fish are brown, eyes blue-green, fins yellowish, and caudal fin slightly emarginate. The size of the fish is only 7 cm”. For an updated list of characteristics, see § 7.
Fig. 24. The Kipili Archipelago constitutes two groups of islands, each harbouring a partly unique set of species and variants. The northern islands (yellow) are Kamamba, Kasisi, Kerenge, Mwila, and Nkondwe, and the southern islands (orange) are Lupita, Mvuna, and Ulwile. Map reference: Government of Tanzania, official maps, and African Diving, lake survey 2007/2008.
Fig. 25. A female of Neolamprologus christyi over a plant-bed of Najas marina armata at Mvuna Island, depth 4 metres. In the Kipili Archipelago, N. christyi is found only at the southern islands, i.e., Lupita, Mvuna, and Ulwile Islands, where Mvuna Island constitutes the most northern distribution of the species. The geographical distribution of N. christyi follows the same pattern as several other species and variants in the Kipili Archipelago, i.e., there is a barrier between the northern and southern islands, each harbouring a partly unique fish fauna with regards to species and colour variants (§ 5). For example, within the archipelago, Chalinochromis cyanophleps, N. furcifer, and N. christyi are found only at the southern islands, while N. sp. “Eseki” is found only at the northern islands, i.e., Kamamba, Kasisi, Kerenge, Mwila and Nkondwe Islands. Furthermore, N. sp. “Mwila” has only been observed at Mwila Island (§ 10, 11) and Lepidiolamprologus kamambae only at Kamamba and Kerenge Island (§ 5, 18). The same phenomenon is observed in the geographical variants of Ophthalmotilapia boops, O. sp. “Whitecap”, Tropheus sp. “Kipili”, Petrochromis famula, among others, where each group of islands harbour different colour variants.
Fig. 26. Neolamprologus timidus (Kullander et al., 2014a) from Ulwile Island, Kipili Archipelago. This species is distinguished from the similar N. furcifer by several morphological features, including cheek squamation (scaled cheek vs. naked), length of pelvic-fin soft ray (first soft ray longer than second vs. second soft ray longer than first), longer head and pectoral fin, as well as melanin pattern (absence of a dark spot at the caudal-fin base in adults and occasionally thin distinct horizontal stripes on the trunk vs. thick and vague). A molecular phylogenetic analysis (both nuclear and mtDNA) places N. timidus and N. furcifer in different clades, each containing different species of Lamprologini (Kullander et al., 2014a).
Fig. 27. Neolamprologus furcifer (Ulwile form) from Ulwile Island, Kipili Archipelago. This form has short rounded lobes in the caudal fin vs. elongated pointed tips in forms from other localities and was discovered in 2008 along with N. timidus (Kullander et al. 2014a: 313, 315). N. furcifer is not found at the northern islands in the Kipili Archipelago, contradictory to Konings (2015b), see more in § 5.

6. Neolamprologus sp. “Eseki”: natural diet
The feeding habits of N. sp. “Eseki” have not been fully elucidated, but it may feed on similar items as morphologically and ecologically similar species, primarily N. mondabu and N. christyi. For example, the 85-mm-long digestive tract of a 120-mm-long specimen of N. christyi from Mtosi in Tanzania contained various debris of crustaceans, molluscs, and worms (Poll, 1956: 586). Regarding the diet of another similar species, Yuma et al. (1998: 373) investigated the stomachs of 14 specimens of N. mondabu from Luhanga and Pemba (also referred to as Bemba/Mbemba), DR Congo and found shrimps (Limnocaridina latipes), Diptera larvae, and Ephemeroptera nymphs. Also, many investigated specimens of N. mondabu from DR Congo had small accumulations of snails in the gastrointestinal tract, which had been eaten unbroken (Büscher, 2002a: 11). The radiographic illustration in Büscher (2002b: 36, Fig. left, centre) shows an N. mondabu with numerous small snails and ostracods in the gastrointestinals. Regarding the ostracods (also known as seed shrimps), there are 80 described Lake Tanganyika species assigned to 25 genera, of which almost all are endemic to the lake (Park & Downing, 2000: 305).
At Luhanga, Pemba, and Uvira, DR Congo, adult individuals of N. mondabu were observed to feed on benthic gastropods, ostracods, shrimps, and insects (possibly mosquito larvae), and the presence of sand and detritus in their “stomach combined with observations of feeding behaviour indicates that this fish is a sand sifter” (Gashagaza & Nagoshi, 1986: 37–38, 40). Also at Pemba, individuals of N. mondabu were observed digging in the sand and feeding on chironomid larvae and other benthic invertebrates (Yuma, 1993: 217; 1994: 179). The 75-mm-long intestinal of a 100-mm-long specimen of either N. mondabu or N. modestus (Poll erroneously treated them as conspecific) was filled with fragments of small shells of gastropods (Poll, 1956: 487). Regarding the natural diet of N. modestus from Zambia, the gastric tract of investigated specimens comprised mosquito larvae, slug parts, filamentous algae, and 5–10% inorganic matter, such as sand (Büscher, 1988: 36–37). Although N. modestus has been observed hunting for juveniles of N. sexfasciatus (Brichard, 1982), its diet mainly comprises snails and mussels (Staeck, 2014: 135). Furthermore, while N. mondabu is known to turn over stones to find food (Hori, 1983), N. modestus rather vigorously wriggles its tail in the sand, occasionally standing vertically, to stir up edibles, both for themselves and their offspring (Ota & Kohda, 2014). Also, both species feed on planktonic copepods, especially during the early juvenile stage (Nagoshi, 1983; Gashagaza & Nagoshi, 1986; Ota & Kohda, 2014). Regarding the similar N. petricola from DR Congo, the ‘dweller among rocks’, small snails have frequently been found in the gastrointestinal tract of several specimens (Büscher, 2002a: 11), and the 70-mm-long digestive tract of a 105-mm-long specimen, which was caught with rod and line with worm as bait, contained many aquatic insect larvae (Poll, 1956: 557). In conclusion, the natural diet of N. sp. “Eseki” is likely to be shrimps, copepods, ostracods, snails, mussels, and/or insects.

Fig. 28. Nkondwe Island is the northernmost island in the Kipili Archipelago. The island harbours a similar set of species and geographical variants as the other islands within the northern group (Kamamba, Kasisi, Kerenge, and Mwila Islands; map, Fig. 24). For more details regarding the two groups of islands in the Kipili Archipelago, see § 5. DR Congo is visible in the background.
Fig. 29. Neolamprologus sp. “Eseki” at Nkondwe Island, depth five metres. The species is not sympatric with N. christyi in this locality (§ 5), contradictory to statements in Konings (2015a). The latter species, N. christyi, is only found at the southern group of islands (Lupita, Mvuna, and Ulwile Islands) and southwards (§ 5, 31; map, Fig. 24).

7. Characterisation and differentiation of Neolamprologus sp. “Eseki”
Based on a diagnostic species concept (e.g. Karlsson & Karlsson, 2018a: 35–57), N. christyi, N. mondabu, and N. sp. “Eseki” are clearly three distinct species. Konings (2015a: 239, 242) regards N. sp. “Eseki” to be a synonym of N. mondabu and appears to ponder on the possibility that also N. christyi may be a synonym of N. mondabu, but dismisses the latter mainly due to the distributional ranges of the two species, which he erroneously believes to overlap in the Kipili Archipelago, see above (§ 5). However, regardless of whether populations exist in the same habitat or are geographically isolated, ultimately their species status is determined by their characteristics.
N. sp. “Eseki” is a greyish to dark brown species that dwells in a shallow habitat comprising a mixture of sand and rocks, including gravel and stones. Individuals appear rather aggressive against conspecifics and both sexes of adult individuals presumably defend territories against same-sex rivals and food competitors. Possibly, like the similar N. mondabu, N. sp. “Eseki” is monogamous, or males are polygynous with their territories comprising several smaller female territories (cf. Gashagaza, 1991; Yuma, 1994; Takemon & Nakanishi, 1998).
Superficially, N. sp. “Eseki” is similar to N. mondabu, but appears to be shorter (10–11 vs. 11–12 cm), more deep-bodied, has a shorter, less pointed head, more indented (concave) caudal fin with more extended or pointed corners, dark brown (vs. light grey) adult body, greenish blue (vs. reddish blue) iris, and the ventral and dorsal margin of the orbit (eye) is frequently (vs. never) yellow. Furthermore, in N. sp. “Eseki”, the dorsal, caudal, and anal fins are usually dark yellowish brown (vs. reddish bluish), more precisely, in sub-adults the margin of the dorsal and upper part of caudal fin is yellowish or yellowish orange (vs. reddish orange) and in adults blackish yellow (vs. orange) with a greyish-bluish submargin (vs. lilac or dark blue). Regarding the dissimilarity with N. christyi, N. sp. “Eseki” is shorter (10–11 vs. 12–13 cm), more deep-bodied, has a less indented caudal fin (vs. deeply concave with elongated marginal rays), lacks a full yellow circle around the eye (prominent in adult N. christyi), lacks a juvenile bright blue fin colouration (Fig. 72), and becomes dark brown (vs. blackish). Although the dissimilarities between the three species are subtle, generally, the distinction between N. mondabu and N. sp. “Eseki” appears to be as clear as the distinction between N. sp. “Eseki” and N. christyi, especially when the caudal-fin shape and body length are concerned. Obviously, a future investigation of the taxonomical characters of N. sp. “Eseki” may reveal further dissimilarities with both N. mondabu and N. christyi. The three species never coexist and may constitute a group of closely related species, potentially referred to as a species complex, see more below (§ 30, 31).
In a mitochondrial DNA (mtDNA) study by Sturmbauer et al. (2010), N. sp. “Eseki” from an unknown sampling location was grouped in a subclade also containing, among others, N. petricola “Kapembwa, Zambia” [=N. cf. petricola “Zambia”] and N. mondabu “Mpimbwe, Tanzania” [=N. sp. “Modestus Mpimbwe”], but not, however, the similar-looking N. christyi. The study did not include the real N. mondabu (from the northern part of the lake), and N. sp. “Eseki” was included only in a phylogeny based on mtDNA, not on nuclear DNA. Furthermore, in a DNA barcoding study by Breman et al. (2016), two scientifically undescribed lamprologin species were identified as distinct, N. sp. “Eseki” and Chalinochromis sp. “Bifrenatus”.
In 1990, after the collection of the first specimens of N. sp. “Eseki”, we bred the species in our fish house in Dar es Salaam, Tanzania, to determine the colouration of the juveniles. By that time, N. sp. “Cygnus” having bright yellow juveniles had successfully been introduced to the aquarium hobby (Konings, 1991; Karlsson & Karlsson, 2017g; 2017h; 2017i), and we wanted to see if the juveniles of N. sp. “Eseki” were similarly colourful. They looked very similar to their parents, i.e., greyish brown and each spawn resulted in more than 100 fry. In the wild, such as at Katondo, Cape Mpimbwe, a pair of N. sp. “Eseki” produced a spawn of an estimated 100 fry (Fig. 17 [video]). In comparison, the clutch size of N. mondabu and N. christyi contains between 150 and 250 eggs (Konings, 2015a: 242–243). Also, an N. modestus-like species at Miyako Point, Tanzania, produced 180 eggs in a spawn (Kuwamura, 1986: 135).
The epithet Eseki is a fictional name constructed to sound intriguing and African. It is a mixture of letters without any link to an existing name or meaning. Initially, we did not recognise N. sp. “Eseki” as a new species, but referred to it as N. mondabu, see above (§ 4). Similarly, Konings (1998a: 161, Fig. bottom, left) first treated it as conspecific with N. sp. “Modestus Mpimbwe”, referred to as N. mondabu, later as a distinct species, N. sp. “Eseki” (Konings, 2005: 117), and more recently as distinct from N. sp. “Modestus Mpimbwe”, but conspecific with N. mondabu (Konings, 2015a: 239). Due to the uncertainty and confusion of the species in the past, the French interpretation of the name Eseki, meaning ‘who is this’, seems rather appropriate [E (Est=is) se (ce=this) ki (qui=who)].

Fig. 30. Neolamprologus sp. “Eseki” at Kekese.
Fig. 31. Neolamprologus sp. “Eseki” at Mlowa Point, Kipili Mainland.
Fig. 32. Neolamprologus sp. “Eseki” at Katondo Point, Cape Mpimbwe.
Fig. 33. Neolamprologus sp. “Eseki” at Kansombo.
Fig. 34. Neolamprologus sp. “Eseki” at Kabwe.
Fig. 35. Normally conditioned by a genetic variation, lamprologin cichlids occasionally possess irregular colouration marked by the absence or abnormal development of black pigments. An individual of Neolamprologus sp. “Eseki” exhibiting a pigmentary anomaly characterised by a pearly white head, photographed within a month of capture at our fish facilities in Dar es Salaam in the late 1990s.

8. The unusual caudal-fin shape: indented, emarginate, forked, crescentic, or lunate
In contrast to Poll’s argument (1956: 484) in the synonymisation of N. mondabu (indented or emarginate caudal fin) with N. modestus (rounded or subtruncate caudal fin), the caudal-fin shape is normally invariable in cichlid species and a reliable species character. Of the lamprologins, N. mondabu appears to have the least extreme indented caudal fin, while the most extreme is seen in N. furcifer (Kullander et al., 2014a: 325). Species with unusually indented caudal fins are sometimes named and mainly diagnosed, both formally and informally, based upon this character, including Tropheus sp. “Crescentic”, distinguished primarily from the similar T. brichardi by a “higher, longer, and crescentic tail” (Brichard, 1989: 166; Karlsson & Karlsson, 2015c); N. falcicula [=small sickle (a hand tool with a semi-circular blade)], “exaggerated crescentic caudal fin” (Brichard, 1989: 536); N. furcifer [=fork], “caudal deeply notched, crescentic” (Boulenger, 1898b: 9); Cyathopharynx furcifer, “caudal deeply emarginate, crescentic, the rays at the angles produced” (Boulenger, 1898b: 14); Lethrinops furcifer (of Lake Malawi), “caudal forked” (Trewavas, 1931: 149); L. lunaris [=of the moon], “caudal deeply emarginate” (Trewavas, 1931: 148); Otopharynx selenurus [=moon tail], “caudal scaly, deeply emarginate” (Regan: 1922: 679); Taeniolethrinops furcicauda [=fork tail], “caudal deeply emarginate” (Trewavas, 1931: 150). For etymological references, see Scharpf and Lazara (2018).

9. Mwila Island in the Kipili Archipelago
Mwila Island (Figs. 1 and 36) is located approximately 500 metres east of Kerenge Island. It is a small circular island with a circumference of about 850 metres, with trees and bushes as well as beaches of mostly different sized rocks. The distribution of the rocks extends into the water, where at depths between 10 and 15 metres, the habitat becomes more sandy and sediment-rich. The water is periodically cloudy due to the murky and turbid water of Mkamba River, which pours into Mbwilo Bay (Fig. 73) and spreads sediment to the nearby Mwila Island and Kerenge Island.

Fig. 36. South-eastern side of Mwila Island with Kasola Hill in the background, 10 km away (between Kalanswi Bay and Chongo Bay).

10. The discovery of Neolamprologus sp. “Mwila” in 2008
In 1991, our team first visited Mwila Island in the Kipili Archipelago (see an account of the cichlids in the Kipili area in Lundblad & Karlsson, 1993a; 1993b). At that time, two species were of particular interest, N. sp. “Cygnus” and J. marksmithi (see the Mwila Island variant in Karlsson & Karlsson, 2018a: 81, Fig. 206). Many more interesting cichlids inhabit the island, such as Tropheus sp. “Kipili” with orange juveniles, N. sexfasciatus, a bright yellow variant, N. brichardi, a variant with yellowish fins (Karlsson & Karlsson, 2017a; 2017b: 88, Fig. 51), etc. However, as Mwila Island is only one of eight islands in the archipelago, the exploration of the island in the 1990s was brief and limited to the rocky habitat.

“We shall not cease from exploration
And the end of all our exploring
Will be to arrive where we started
And know the place for the first time.
Through the unknown, remembered gate
When the last of earth left to discover
Is that which was the beginning”
(T. S. Eliot, 1943)

In 2008, we returned to Mwila Island for further underwater observations and investigations. In the afternoon of 30 April, at the sandy bottom at a depth of about 15 metres, we found an unknown N. mondabu-like species, subsequently referred to as N. sp. “Mwila”. Observed individuals were either patrolling over territories encompassing clusters of empty Neothauma tanganyicense shells, many of them in bad shape, partly covered by algae, and half buried in the sand, or guarding their offspring peeking out from some of the shells. No obvious excavation of sand was observed, nor did we see any adult individuals of N. sp. “Mwila” taking refuge inside the shells. When feeling threatened by our presence, they darted away in the opposite direction, temporarily leaving shells and fry. Seemingly, the empty snail shells were only nurseries and shelter for the offspring, not actual housing for the adults. Besides sand and empty shells of various gastropods and bivalves, the observed surrounding included gravel, layers of sediment, and many yellow or green Nudospongilla moorei-like sponges (Figs. 38–40) of a maximum size of about 10 cm. Additional fish species were observed in this habitat, including Lamprologus callipterus, T. dhonti (Figs. 52–53), an N. brevis-like species with an indented caudal fin (see more in the next blog post), and a Lepidiolamprologus meeli-like species, possibly L. sp. “Meeli-Boulengeri” (Figs. 50–51). On another side note, in the 1990s, we collected two different but very similar Lepidiolamprologus meeli-like species, one at Cape Tembwe, DR Congo, referred to as L. meeli, and the other along the opposite coast at Cape Mpimbwe, Tanzania, referred to as L. hecqui. When they were observed in aquaria, small differences were readily noticeable, the specimens from Cape Tembwe were shorter, had larger eyes, thicker black margins and stripes on their fins, and more blackish overall. While the Cape Tembwe variant is probably the true L. meeli (Poll, 1948: 29; Walschaerts, 1987: 32; cf. Herrmann, 1987: 77, Fig. bottom; Brichard, 1989: 353, Fig. top; 1989: 361, Fig. bottom, right; Büscher, 1998b: 58, Fig. 16), the Cape Mpimbwe variant, as well as that at Mwila Island, appears to be another species, presumably conspecific to L. sp. “Meeli-Boulengeri”, a species first identified by Van Wijngaarden (1995, cited in Konings, 2015a: 350) in reference to the southern or southeastern populations (cf. Schelly et al., 2006; Sturmbauer et al., 2010; cf. Karlsson & Karlsson, 2013: 16–17). Furthermore, L. hecqui, which hails from the Mtoa area, opposite to Kavala Islands, DR Congo, is likely an additional distinct species. Due to its blackish pelvic fins, it may be closely related to L. boulengeri (cf. Konings, 2015a: 350), or even conspecific to L. boulengeri. Indeed, Boulenger (1899c: 115) and Steindachner (1909a: 400) did not have access to living colours and Steindachner was not aware of Boulenger’s description of L. hecqui, at least he did not take this into account when describing L. boulengeri. In addition, the specimens collected by William Alfred Cunnington in 1904–1905 in the Mpulungu area, Zambia, which were grey above and lighter below, and identified as L. hecqui by Boulenger (1906: 559), accords better with L. sp. “Meeli-Boulengeri”, or perhaps it is yet another species (cf. DeMason, 2017: Fig. top, right). As for a short and provisional key, L. hecqui has blackish fins, including the pelvic fins, L. boulengeri has blackish pelvic fins, but yellow dorsal, caudal, and anal fins, both L. meeli and L. sp. “Meeli-Boulengeri” have lighter pelvic fins, the former possibly has a shorter body, larger eyes, thicker black margins, and a darker colouration.

Fig. 37. On the southern and eastern side of Mwila Island, at a depth of 15 metres, a vast open underwater landscape of sponges and battered Neothauma shells spreads out as far as the eye can see.
Fig. 38. Nudospongilla moorei-like yellow and green sponges are common in the habitat at Mwila Island.
Figs. 39–40. Moore (1898c: 162) spoke about deep-water crabs, prawns, and sponges as if they were particularly deep living. However, if the biotope is right, Nudospongilla moorei-like sponges may be found in less than 15 metres at Mwila Island, Kipili Archipelago (photos).
Fig. 41. A male of Neolamprologus sp. “Mwila” over a bed of empty Neothauma tanganyicense shells at Mwila Island, depth 15 metres. Total length of the individual is about 5–6 cm. For comparison with a juvenile N. sp. “Eseki” at Udachi of about the same size, see Fig. 22.
Fig. 42. A female Neolamprologus sp. “Mwila” at a depth of 15 metres guarding its offspring. The photo was taken at the eastern side of Mwila Island. Maximum recorded brood size is 8 fry (this photo), but a brood size of about 4–6 fry appears to be the most common.

11. Characterisation and differentiation of Neolamprologus sp. “Mwila”
N. sp. “Mwila” appears to be extremely cryptic to N. sp. “Eseki”, which has also been observed at Mwila Island, but in a shallower (5–10 metres) and different habitat comprising a mixture of sand and rocks of different sizes (30–300 cm). Furthermore, N. sp. “Eseki” does not appear to be as common at Mwila Island as it is at any of the other northern islands. Possibly, there is no established population, only occasional individuals arriving from neighbouring habitats. Nonetheless, at this island, both N. sp. “Mwila” and N. sp. “Eseki” have a beige or brown-beige body, mainly yellowish dorsal fin with a yellow margin and blue submargin, a yellowish indented caudal fin with a blue upper submargin, and white pelvic fins. Differential characteristics in favour of N. sp. “Mwila” seemingly include a shorter body (5–7 vs. 10–11 cm), shorter and wider snout, larger eye, the latter which is mainly bluish and lacks the yellow margin of the orbit as well as the yellow patch on the upper part of the eye. Possibly, the caudal fin of N. sp. “Mwila” is less indented. While N. sp. “Mwila” has been observed at Mwila Island with 4–8 offspring, spawning pairs of N. sp. “Eseki”, both in the wild and aquaria produced an estimated 100 fry. Furthermore, females of N. sp. “Mwila” were observed at Mwila Island guarding their offspring hiding in empty gastropod shells, while males were seen patrolling nearby. Seemingly, males patrol the outer boundaries of the territories, or, alternatively, the territories of the males are relatively large, and each encompass several smaller territories of brooding females, hence, N. sp. “Mwila” may be monogamous, or practising serial or harem polygyny. Also, the observed juveniles of N. sp. “Mwila” appeared to be somewhat larger than the more numerous juveniles of N. sp. “Eseki”. We did not observe any ‘alternative reproductive tactics’, as described for T. vittatus, T. temporalis, and N. modestus, involving territorial, pirate, satellite, and/or sneaker males (Katoh et al., 2005; Ota & Kohda, 2006; Hellmann et al., 2015). Furthermore, N. sp. “Mwila” was not observed being hunted or attacked by predators, which could potentially constitute nocturnal species, such as clariid (e.g. Dinotopterus cunningtoni), claroteid (e.g. Chrysichthys sianenna), and malapterurid (e.g. Malapterurus tanganyikaensis) catfishes, or mastacembelid eels (e.g. Mastacembelus ellipsifer).
When dealing with similar fishes of different sizes potentially being distinct species, one needs to take allometry into account. Allometric growth refers to the growth pattern in which different parts of the body grow at different rates (Mayr, 2004: 219). “The eye-diameter (measured as a percentage of the head-length) is a notorious allometric proportion: small fishes have relatively large eyes. Correlated with the eye-diameter, a number of other proportional measurements change during growth: e.g. cheek-depth and snout-length” (Barel et al., 1977: 251–352).
During our expedition in 2008, we collected many specimens of N. sp. “Mwila”, some which were fixed and preserved in formalin and ethanol, and 30–35 were placed for a year in aquaria for observation. The fishes spawned in the aquaria, with each clutch containing less than ten eggs and fry. During that year, the adults did not grow much larger, but maintained a size of about 5–7 cm. Also, the proportions of the eye and length of snout did not change, hence, specimens of N. sp. “Mwila” with a length of about 5–7 cm and a relatively large eye and short snout presumably represent adult specimens, which are comparable with those of N. sp. “Eseki” having a smaller eye and being about twice as large. We therefore regard N. sp. “Mwila” as a distinct species separate from the very similar N. sp. “Eseki”.

Fig. 43. Specimens preserved in ethanol, male and female, of Neolamprologus sp. “Mwila” from Mwila Island, deposited at the Swedish Museum of Natural History in 2008, NRM 59636. Photo: Sven O. Kullander
Fig. 44. A male of Neolamprologus sp. “Eseki” at Kerenge Bay, Kerenge Island, adjacent to Mwila Island. Seemingly quite rare, N. sp. “Eseki” is also found at Mwila Island and always encountered in the shallow rocky habitat. The individual in the picture is 9–10 cm.
Fig. 45. A preserved specimen of Neolamprologus sp. “Eseki” from Cape Mpimbwe, deposited at the Swedish Museum of Natural History in 2008, NRM 51532.
Figs. 46–47. A sampled specimen of Neolamprologus sp. “Eseki” from Kamamba Island, before and after preservation, deposited at the Swedish Museum of Natural History in 2008, NRM 60267.
Fig. 48. A male of Neolamprologus sp. “Mwila” sampled at Mwila Island in 2008. The size of the individual is about 6 cm, the average size of adult males, NRM 59636.
Fig. 49. A male (top) and female of Neolamprologus sp. “Mwila”. The number of sampled specimens was about 60, of which 25 were fixed in formalin and deposited at the Swedish Museum of Natural History in 2008, NRM 59636, with the remainder sent to Sweden for studies in aquarium.

12. Ecological forms and semi-distinct incipient species: the case of Telmatochromis temporalis
Micro allopatric speciation, the divergent natural selection of broken up stenotopic and philopatric populations following the fragmentation of their required habitats (Coulter, 1994), and sympatric speciation, partly the divergent adaptation to ecological niches, are hypothesised to be important processes in the evolution of Lake Tanganyika cichlids, including the Lamprologini. Ecological adaptation may lead to the modification of trophic morphology and behaviour, eventually resulting in new distinct species.
Ecological forms and semi-distinct incipient species have been studied regarding several lamprologin species. For example, the surroundings of Wonzye Point and Mbita Island, southern Zambia, and Chibwensolo, Ndole Bay, western Zambia, each contain a population of a T. temporalis-like species with two geographically interconnected ecological forms, referred to as rock and shell (or shell bed) dwelling individuals, or normal and dwarf morph (Takahashi, 2004; Takahashi et al., 2009). Shell beds are large areas of substrate completely covered in empty shells, normally of the gastropod snail Neothauma tanganyicense. While the area of Mbita Island (also known as Nkumbula/Kumbula Island, § 32.9) and Wonzye Point is about 80 km away from Chibwensolo, between which no shell beds or shell-dwelling populations are found, the distance between Mbita Island and Wonzye Point is 4.7 km (Takahashi et al., 2009: 3111, 3118). Specimens of both morphs were collected at all three locations and analysed by means of mtDNA and microsatellite (nuclear DNA segment) data as well as geometric morphometrics. Analyses of molecular variance were performed to assess the genetic differentiation within and between morphs and populations (Takahashi et al., 2009: 3112), supporting different origins of the shell-dwelling morphs from western and southern Zambia, i.e., they evolved independently, presumably from the rock-dwelling morph, which was regarded as more ancestral due to its more diverse mtDNA, closer resemblance to the similar T. brachygnathus (but see below, § 13), and a greater lake-wide distribution.
Regarding the southern populations, while the genetic differentiation between morphs (i.e., the rock-dwelling and shell-dwelling morphs irrespective of location, Mbita Island or Wonzye Point) was significant, that between populations within morphs (e.g., the shell dwellers of Mbita Island and Wonzye Point) was much smaller (Takahashi et al., 2009: 3114). In other words, the gene flow between the southern populations of the same morph was greater than the gene flow between the southern morphs from the same location. This prompted Takahashi (2004: 430) to suggest that the rock-dwelling and shell-dwelling individuals constitute two separable populations (cf. ‘sympatric populations’, § 15, Jorde et al., 2018). In each of the southern localities (Mbita Island and Wonzye Point), the geographical distribution of rock-dwelling and shell-dwelling morphs overlap, so gene flow between morphs at the same location is more likely constrained by ecological factors than geographical factors, which may be interpreted as a niche adaptation and initial step of sympatric speciation (cf. Takahashi et al., 2009: 3118).
While the rock-dwelling morph is referred to as T. temporalis (Takahashi, 2004; Takahashi et al., 2009), shell dwellers similar to those in the preceding studies, have been referred to as T. burgeoni (e.g. Poll, 1986; Herrmann, 1987; Konings, 1988; Tawil, 1988; 2010; Sato & Gashagaza, 1997; Büscher, 1998b: 53, ‘uncertain status’). Seemingly, T. burgeoni, which was described based on a single specimen from Nyanza-Lac, represents an artificial category of disconnected and distantly related shell-dwelling and T. temporalis-like populations dispersed throughout the lake. If T. burgeoni is a valid species, it would probably only comprise individuals and populations at Nyanza-Lac and the nearby surroundings. However, following the field observations by Konings (1998a: 98) and morphological comparative studies by Hanssens and Snoeks (2001; 2003), T. burgeoni is suggested to be a synonym of T. temporalis.
Takahashi (2004: 430–431) and Takahashi et al. (2009) identified the southern Zambian rock-dwelling morph as T. temporalis based on the teeth of the upper jaw, size of mouth and length of jaws, and other characters attributed to T. temporalis by Hanssens and Snoeks (2001; 2003). Recently, Konings (2015a: 356) suggested the studied shell dwellers of Takahashi (2004) to be T. dhonti. Indeed, Takahashi (2004: 430–431) counted 18–20 dorsal-fin spines in both the dwarf morph and T. dhonti, compared to 21 or more in T. temporalis. Also, in shell beds throughout the lake, miniature forms of T. dhonti or T. dhonti-like species appear to be common (Figs. 52–53). However, Takahashi (2004: 431) demonstrated that the southern shell dwellers are “different from T. dhonti, having significantly more dorsal and anal-fin spines and significantly fewer soft rays on these fins”. Furthermore, the southern shell dwellers had importantly fewer dorsal-fin spines, scales in a longitudinal row, gill rakers, vertebrae, shorter dorsal-fin base, anal-fin base, and caudal peduncle, than the rock dwellers (T. temporalis), but more pectoral-fin rays, premaxillary outer teeth, longer snout, eye, upper and lower jaws, and greater interorbital width. Subtly, the six anterior most premaxillary teeth of the shell dwellers (southern morph) were enlarged, clearly distinct from the small lateral teeth, whereas such a transition of tooth size was gradual and indistinct in the rock dwellers (southern morph) (Takahashi, 2004: 429). Also, the shell dwellers were distinguishable from T. brachygnathus (cf. § 13) in their morphology of jaws and teeth (Takahashi, 2004: 424, Table II, 428–429, 431). Morphometrically, regarding the Zambian morphs, the maximum size of rock-dwelling males is 75.7–88.1 mm, that of shell-dwelling males is 40.3–44.8 mm, while the minimum size of mature rock-dwelling males is 55–65 mm and that of mature shell-dwelling males is 25–35mm. The maximum size of rock-dwelling females is 53.1–62.0 mm and that of shell-dwelling females is 26.8–29.4 mm (Takahashi et al., 2009: 3114). The shell-dwelling morph has greater gonad size than the rock-dwelling morph, which suggests that differences in development of gonads have a genetic basis (Winkelmann et al., 2014: 2). Regarding the ovarian eggs of the females, these are small and more numerous in the rock-dwelling morph but large and fewer in the shell-dwelling morph. Apparently, the maximum body and maturity sizes of the shell-dwelling morph are much smaller than those of the rock-dwelling morph in both sexes. This shorter body accords with the fewer vertebrae, and the diversity of sizes closely matches the available shelter sizes in the respective habitats of the two morphs (Takahashi, 2004: 428, 430; Takahashi et al., 2009: 3114–3116). Also, the rock-dwelling morph showed a significant lunar synchronisation of ovary development, but the shell-dwelling morph did not (Takahashi, 2010).
Based on the above stated morphological and molecular results, Takahashi (2004: 432) concluded that the southern shell dwellers are “likely to be an undescribed species of Telmatochromis”.

Fig. 50. Lepidiolamprologus sp. “Meeli-Boulengeri” at Mwila Island. The name L. sp. “Meeli-Boulengeri” was first used in a master’s thesis by Van Wijngaarden (1995, cited in Konings, 2015a: 350) in reference to the southern form or group of small sand and shell-dwelling Lepidiolamprologus. It is distinguished from the closely related L. boulengeri by the lack of bright yellow fins and L. meeli by the lack of 2–4 lateral and horizontal rows of small whitish spots and a thick black margin on the dorsal, caudal, and anal fin (see more § 10).
Fig. 51. Fry of Lepidiolamprologus sp. “Meeli-Boulengeri” at Mwila Island.
Fig. 52. Telmatochromis dhonti dwells and breeds in empty Neothauma tanganyicense shells at Mwila Island. The picture shows a male of about 6–7 cm at a depth of 10 metres.
Fig. 53. A female Telmatochromis dhonti hiding under a single shell of a bivalve mollusc at Mwila Island. The female in the picture is about 5 cm, a common size for shell-dwelling individuals, but ultimately females may grow to about 8–9 cm.
Fig. 54. This individual of Telmatochromis cf. temporalis at Ngosa Point, just south of Kipili Archipelago, shows a camouflage colour pattern possibly caused by stress. Note the larger eye in this species compared to T. vittatus of the same locality (see next photo, Fig. 55).
Fig. 55. Telmatochromis vittatus at Ngosa Point, just south of Kipili Archipelago.
Fig. 56. Telmatochromis vittatus in Mtosi Bay, 14 km south of Ngosa Point. Adult males in full colouration often lose their stripes (compare with Fig. 55, see also Figs. 57–58).
Figs. 57–58. Telmatochromis bifrenatus at Luagala Point, the central part of the Mahale Mountains National Park. T. bifrenatus was described based on two specimens from Kigoma, both measuring only 4 cm in standard length. The specimen in the photos is about 10 cm (the photos show the same specimen).

13. Telmatochromis brachygnathus and the T. temporalis species complex
Konings (1998a; 2015), and Hanssens and Snoeks (2001; 2003) regard the miniature form that corresponds to the shell-bed dwellers and dwarf morph of Takahashi (2004) and Takahashi et al. (2009) as an ecological morph of T. temporalis, and therefore taxonomically synonymous with it. Konings (2015a: 158) also regards T. brachygnathus to be a morph and synonym of T. temporalis, claiming that certain populations consist of a mix of individuals which may be classified as either T. temporalis or T. brachygnathus, i.e., with short or long upper and lower jaws (the main distinguishing feature), and that these different individuals intermingle and form pairs. Despite the urge for only proposing synonyms with great caution (cf. N. petricola, § 26), we concur with Konings that T. brachygnathus (in the wide sense) is synonymous with T. temporalis (sensu Boulenger, 1898b), alternatively, a geographical replacement form with species status within the complex of T. temporalis-like species, hence, a synonymisation of T. brachygnathus (in the strict sense) with T. temporalis is perhaps not necessary. However, although a scientific species description is not always easy to comprehend, it “is the reader who is the important agent, the user of the information, not the paper, not the author” (Kullander, 2012: 3), therefore, one would expect a much clearer presentation of T. brachygnathus than what Hanssens and Snoeks (2003) delivered. Furthermore, some of our own observations, which partly conflict with those of Konings, may be worth pointing out. For example, throughout the Tanzanian coast, we have observed under water more than 100 populations of the T. temporalis species complex, many of which we have sampled for scientific studies, in addition to the many specimens we have collected from throughout the lake and exported to the aquarium hobby, but never have we found a population with a mix of T. temporalis-like and T. brachygnathus-like individuals, possibly with one exception, the population at Kasola Island in which one large individual was observed to have a protruding lower jaw while the others, mostly smaller individuals, had equal jaws. However, a protruding lower jaw is not explicitly stated to be a differentiating feature between these two species, but rather a set of both upper and lower protruding jaws (cf. Hanssens & Snoeks, 2003). Furthermore, the fact that a larger individual had a longer lower jaw than the rest of the observed individuals may be linked to allometry, which indeed appears to be the case (cf. Hanssens & Snoeks, 2003: 610). In addition, not even the rock-dwelling and shell-dwelling forms in southern Zambia appear to have different kinds of jaws, despite the divergent ecological adaptation (cf. Takahashi et al., 2009: 3111, Fig. 1a). Also, we do not support Konings’ claim of the existence at Nkondwe Island of a population of individuals with both long and short lower jaws combined with bodies of great and small height, as illustrated in his publication (Konings, 2015a: 158). Based on our experience, the lower photo on page 158 appears to illustrate an individual from Mvuna Island, not Nkondwe Island. However, we agree with Konings (2015a: 159) that T. temporalis, or the complex of T. temporalis-like species, “is morphologically very plastic”, regarding the diversity of populations.
Obviously, removing Konings’ main argument for the invalidation of T. brachygnathus, i.e., the occurrence of both short and long jaws within T. temporalis-like populations, renders the species still valid. Nonetheless, there are other details which may invalidate T. brachygnathus s.l., including the potential misidentification of T. temporalis by Hanssens and Snoeks (2003). In part, their T. temporalis appears to be distinct from Boulenger’s T. temporalis (1898b), presumably representing an undescribed species. While they seemingly have identified two similar but distinct and occasionally sympatric T. temporalis-like species, Hanssens and Snoeks (2003) possibly chose the wrong species to represent T. temporalis and applied the new name T. brachygnathus to the real T. temporalis, hence the synonymisation. The full story, including new Telmatochromis species will be presented in another article.
Geographically and morphologically distinct groups of T. temporalis-like populations were found throughout much of the lake by Hanssens and Snoeks (2003: 606, 611–612), suggesting the potential for more species to be identified. Eventually, all these geographical replacement forms may receive scientific names, being included in what Hanssens and Snoeks (2003) refer to as the T. temporalis complex, for example, T. sp. “Temporalis Elongated Yellow Top” from southern Tanzania (Fig. 65).

Fig. 59. A relatively elongated variant of Telmatochromis cf. temporalis found at Namansi. This and several other variants were sampled by the authors in 2008. The picture was captured at a depth of six metres at Namansi Point. Some of the Telmatochromis species, including T. temporalis, T. dhonti, and similar species, have always been among the hardest group of lamprologin species to properly identify, which is partly because several basically unknown species still await characterisation (§ 13).
Fig. 60. Telmatochromis cf. temporalis at Mwaka, Korongwe Bay. Variants of the T. temporalis complex are allopatrically distributed throughout the lake, many of which presumably constitute distinct species. The introduction of T. brachygnathus by Hanssens and Snoeks (2003) [differentiated from T. temporalis by a smaller mouth (shorter upper and lower jaws)] is confusing as there are seemingly no T. temporalis-like species with a shorter pair of jaws than T. temporalis. Possibly, the name T. brachygnathus should be restricted to the T. temporalis-like populations at Cape Chaitika, the type locality for T. brachygnathus. See more in § 13.
Fig. 61. A yellowish-white Telmatochromis vittatus (left) and an orange-beige T. cf. temporalis (right) at Kisi Island. At some locations, these species are quite similar as both vary geographically in body colouration and length. Following more research on the genus, within the taxa, T. vittatus and T. temporalis, additional species will very likely be recognised. An in-depth analysis of Telmatochromis will be presented in a future article (§ 13).
Fig. 62. Telmatochromis cf. temporalis at Kitawe at a depth of five metres. The species and variants of the T. temporalis species complex differ primarily in the length and height of the body. Commonly, there is also variation in the body and fin colouration, occasionally in the length of the lower jaw.
Fig. 63 (left). Telmatochromis cf. temporalis, Lamvya Bay, has a short stocky body. Fig. 64 (right). T. cf. temporalis, Mtosi Bay, has a long slender body. Both specimens (Figs. 63 and 64) were found in a typical rocky habitat.
Fig. 65. Telmatochromis sp. “Temporalis Elongated Yellow Top”. Discovered in southern Tanzania in 2008, this variant is extremely elongated and regarded as a distinct species.
Fig. 66 (left). Telmatochromis cf. temporalis, Lamvya Bay, NRM 51555. Fig. 67 (right). T. cf. temporalis, Popo Point, NRM 60280. There is a large morphological difference between these two variants, with the Popo Point variant almost twice as long as the Lamvya Bay variant. Preserved specimens are from the collection of the Swedish Museum of Natural History.

14. Colouration as a cause of divergent selection in Lamprologini
Dichromatism comprising light beige and dark greyish or blackish individuals has been observed regarding the T. temporalis species complex (Hanssens & Snoeks, 2003: 610; pers. obs.). This dichromatism is not geographically clustered, but occurs within populations, appearing to be a step in divergent ecological niche adaptation. While blackish individuals have a camouflage better adapted to shadowy spaces, such as under rocks and inside caves, light beige individuals better blend in on top of rocks in well-illuminated areas (cf. Mboko & Kohda, 1995; Maan & Sefc, 2013). For a video sequence of a pair of a light beige T. temporalis-like species confronting a blackish pair at Kitawe, south of Samazi, see Karlsson and Karlsson (2017j: 1 min 46 sec).
Partly due to the limitation of obvious sexual and individual variation in colouration, speciation by sexual selection on polymorphic colouration as opposed to micro allopatric and sympatric speciation (see above, § 12) is generally thought to be rare in Lamprologini. Relatively few geographical colour variants exist, but the lamprologins may have evolved more variation in colouration had it not been for the piscivores, such as lates, spiny eels, and catfishes, who more easily detect and catch conspicuously coloured prey. Indeed, brightly coloured lamprologin individuals are occasionally observed, including a bright yellow specimen of a T. temporalis-like species at Muloba Bay, Mahale Mountains (Fig. 69). The enhanced colouration of some lamprologins, such as Julidochromis marksmithi, or the peculiar colouration of some other congeneric species (cf. Karlsson & Karlsson, 2018a), may be a result of sexual selection as opposed to natural selection and ecological adaptation. Furthermore, in Lufubu River, where the predation is presumably different from that in Lake Tanganyika, a T. dhonti-like species with a partly bright red dorsal fin was recently discovered (Indermaur, 2014; Büscher, 2016; 2017). The fact that both male and female are the same colour does not affect the above reasoning. In Lake Tanganyika, many haplochromins, which generally are a symbol of sexually distinct colouration, have a unisexual colouration, including some of the Tropheus and Petrochromis species, yet, they have evolved spectacular colouration. In conclusion, processes such as sexual selection, natural selection, and ecological niche adaptation may all contribute to the evolution of the Lamprologini.

Fig. 68. Two differently coloured Telmatochromis cf. temporalis at Nkonkonti Bay. Dichromatism in the T. temporalis species complex is not geographically clustered but occurs within populations and appears to be a step in divergent ecological niche adaptation.
Fig. 69. A colourful bright yellow-orange individual of Telmatochromis cf. temporalis at Muloba Bay, Mahale Mountains National Park.

15. Ecological forms and semi-distinct incipient species: the cases of Neolamprologus mondabu and species of Altolamprologus
Regarding N. sp. “Mwila”, had it not been for the larger adult eye and maximum size of 5–7 cm, observed and collected specimens of N. sp. “Mwila” could have been sub-adults of N. sp. “Eseki”. Although field observations and visible morphological characteristics suggest it to be a distinct species, pending a proper taxonomic investigation, N. sp. “Mwila” may alternatively be regarded as an ecological miniature form of N. sp. “Eseki”. Similarly, at Rumonge, Burundi, Gashagaza et al. (1995) discovered an N. mondabu-like population of small individuals dwelling inside or near empty shells, which appeared specifically distinct. The maximum size (standard length) of investigated specimens was 48.2 mm for males and 41.2 mm for females (vs. 84.1 and 68.9 in N. mondabu), with 4–9 lower lateral line scales (vs. 9–14), and a shorter dorsal-fin base. However, based on overall appearance, the Rumonge specimens appeared very similar to N. mondabu (cf. Gashagaza et al., 1995: 299, Fig. 6). Awaiting a taxonomic study, the population was temporarily concluded to represent an ecological morph of N. mondabu, or simply mature youngsters of that species.
Likewise, Büscher (1998b: 56) reported the existence of a Lamprologus speciosus population at Kasenga, DR Congo, comprising unusually small individuals, which did not live in empty shells of Neothauma, as the species usually does, but in shells of the smaller Lavigeria. When these individuals were placed in aquaria and had access to larger shells, they grew larger, eventually attaining a similar size as regular L. speciosus. We have made similar observations of species such as Lepidiolamprologus attenuatus and T. vittatus. Typically, for many lamprologins, they reach sexual maturity well before reaching the maximum body size, but this may be different for certain lamprologin species, including those of Altolamprologus. In the mid-1980s at Sumbu in Cameron Bay, Horst Walter Dieckhoff discovered a small A. compressiceps-like shell-dwelling form (Herrmann, 1987: 70), which was first exported in 1986 by Adrian Carr (Eysel, 1987: 630), referred to as A. sp. “Sumbu” (Konings, 1988: 214). Alternative names for this potentially distinct species include A. sp. “Shell”, A. sp. “Sumbu Shell”, A. sp. “Sumbu Dwarf”, A. sp. “Compressiceps Shell”, A. sp. “Compressiceps Sumbu”, A. sp. “Compressiceps Sumbu Shell”, A. compressiceps “Sumbu”, A. compressiceps “Sumbu Shell”, among others. For an online photo, see, for example, DeMason (2017: Fig. second from top, left). The maximum observed size was 5.5 cm for males and 4 cm for females, with a more purplish body and more yellowish fins than the regular A. compressiceps coexisting in the same area, i.e., western Cameron Bay. Apparently, A. sp. “Sumbu” is smaller, has different colouration, and has adapted to a different habitat type than A. compressiceps. Possibly, there may be additional differences, such as a shorter and shallower dorsal fin with fewer hard and soft rays. Yet, the two species/forms are commonly regarded as conspecifics (e.g. Konings, 2015a: 141), an opinion which is largely based on mtDNA studies, such as that of Koblmüller et al. (2016), where A. sp. “Sumbu” appears to be closely related to a rock-dwelling Altolamprologus species from the same geographical area, presumably A. compressiceps. However, there is also very low degree of genetic differentiation between A. compressiceps and A. calvus, suggesting the latter to have evolved from the former about 67–142 thousand years ago (Spreitzer et al., 2012: 143, 144; Koblmüller et al., 2016). Due to the DNA similarity, genes are regarded as being exchanged between the two species at low frequency (Spreitzer et al., 2012: 144; cf. Koblmüller et al., 2016). Presumably, A. calvus split off from A. compressiceps due to ecological adaptation and restricted gene flow; rocky habitat borders and/or ecological niche partitioning were likely involved in the process. As for A. sp. “Sumbu”, the evolutionary lineage represented by this species is seemingly on a similar evolutionary path as that of A. calvus. However, while the journey of the latter appears to have come to an end, or a temporary pause, that of the former has just begun. Distinct trophic adaptation and the strong bond to a specific type of habitat may eventually lead to speciation (Sturmbauer, 1998: 25). If the shell-dwelling A. sp. “Sumbu” is not yet a distinct species, over time it probably will be. Recently, Jorde et al. (2018) asked if science is “underestimating the occurrence of sympatric populations”, a question which appears to relate to A. sp. “Sumbu” among others.
Besides A. sp. “Sumbu”, there are additional shell-dwelling Altolamprologus along the shores of the lake, which are presumably not closely related to A. sp. “Sumbu”, but rather to the larger rock-dwelling form of A. compressiceps found in the same area. In most cases, these shell-dwellers are not as much ecological forms, let alone incipient semi-distinct species, as they are simply sexually mature young A. compressiceps, or A. compressiceps-like, that cannot yet compete in the rocky habitat, and as such, they are better suited to the sandy habitat with empty shells.
However, in 1990 at Cape Mpimbwe, at a depth of 25–30 metres, we found a seemingly partly isolated sand/shell dwelling population of Altolamprologus, initially exported as ‘Altolamprologus calvus spec.’ (Zadenius, 1991a: 37), subsequently referred to as A. sp. “Micro” or A. sp. “Calvus Micro” (Fig. 71). Specimens were reported to have spawned in aquaria at a size of 25–40 mm, which appeared to be their maximum size (Bengtsson, 1993). The jaws of the specimens pointed straight forward, similar to A. calvus, with a mostly whitish body with brown or greyish lateral bars, the dorsal fin had small white distinct dots, while there was a bluish field on the anal fin. Presumably, these individuals were young but sexually matured A. compressiceps, but they may have belonged to an ecological miniature form and sympatric population of A. compressiceps, representing the dawn of a new species.
Recapitulating the general message, a cichlid lineage may sympatrically split by means of ecological niche separation, one sublineage adapting to a rocky habitat, the other to a sandy habitat. As soon as divergent ecological adaptation begins, divergent reproduction will follow, initially leading to sympatric conspecific populations, eventually to distinct species.

Fig. 70. A defensive male of Neolamprologus sp. “Mwila” at Mwila Island. In comparison with N. sp. “Eseki”, the caudal fin is less indented (concave), almost subtruncate (straight-cut) with slightly pointed corners, the upper one being longer than the lower.
Fig. 71. A miniature Altolamprologus at Katondo Point, Cape Mpimbwe, depth 30 metres. It was exported by us (African Diving) as A. sp. “Micro”. From the first and second export to Sweden in 1991, seven individuals were obtained by Ingemar Bengtsson, who, six months later, successfully bred the fish in aquarium using empty Neothauma shells (Bengtsson, 1993). The size of the male and female was 40 and 25 mm, respectively. Bengtsson (1993) reported that the spawning took place in a manner like other shell-dwelling cichlids. This photo was previously published in the Italian aquarist magazine Hydra (Karlsson, 2004).

16. Historical high and low lake stands in Lake Tanganyika
The age of Lake Tanganyika is thought to be 9–12 million years (Cohen et al., 1993), or, alternatively, 20 million years (Tiercelin & Mondeguer, 1991). Due to climate change, including extended periods of extreme dry and cold climate, the lake level has repeatedly fluctuated, occasionally dropping several hundred metres (Gasse et al., 1989). For example, seismic data revealed that for about 25 thousand years ago the water level of Lake Tanganyika was more than 600 metres below current level (Scholz & Rosendahl, 1988: 1645). Due to the basement topography, which includes long and narrow ridges running across the lake basin, Lake Tanganyika has during the most extreme water level low stands been split into three or four smaller sublakes (Lezzar et al., 1996; Cohen et al., 1997a: 116, Fig. 2L). In periods of extreme low stand, species and lineages of fish, including cichlids, were able to spread across the lake along these ridges. The most recent low stand prevailed between 18 and 14 thousand years ago when the lake level appears to have been 160–180 metres below current level (Lezzar et al., 1996: 1; Cohen et al., 1997a: 119; 1997b; Delvaux & Williamson, 2008: 586), or, alternatively, as suggested by molecular data on Tropheus, about 500 metres below the current level (Sturmbauer et al., 2005: 347). The subsequent warmer climate, beginning about 12 thousand years ago, resulted in a rapid rise of the lake level by more than one hundred metres in less than three thousand years (Scholz et al., 2003: 148; Delvaux & Williamson, 2008: 586–587).
For the record, a high-stand condition predominated in Lake Tanganyika approximately between 170 and 40 thousand years ago, when the water level may have been about 55 metres higher than that of today (cf. Cohen et al., 1997a: 107, 123 and Fig. 5).
During the past two hundred years, there have been some rather moderate fluctuations. For example, due to a natural debris blockage of Lake Tanganyika’s only outlet – the Lukuga River – the lake level rose by 8 metres between 1846 and 1876, which significantly altered the geography of the lake (Camus, 1965: 1244; cf. Carson, 1892; Sieger, 1893: 579; Cunnington, 1920: 515, Worthington & Ricardo, 1937: 1065).
In 1857–1859, during the expedition of Richard F. Burton and John H. Speke, Ubwari Peninsula was referred to as Ubwari Island, while a small village in its northernmost part was referred to as Mzimu (Burton, 1860: map facing page 384). David Livingstone referred to the island as Mozima (Livingstone, 1874: map). On the map in ‘Across Africa’ by Cameron (1877: foldout map after page 389), Ubwari Peninsula is indicated as an island. During the lake surveys by Livingstone and Henry M. Stanley, between November 1871 and March 1872, Ubwari was referred to as Muzimu Island (Stanley, 1902: 482, 493, 510, etc., and map). Stanley (1878a: 21; 1899a: 17) acknowledged that Burton and Speke referred to the island as Ubwari, while Livingstone and himself referred to it as Mozima/Muzimu. Nevertheless, ten years later in 1882, the island had become a peninsula, referred to as Ubwari or Ubwari Peninsula (Hore, 1882; map; Chavanne, 1885), the same name which was also applied by Belgian and German colonials in 1892 (Janssens & Cateaux, 1908: 356, 358; Luscombe, 2018a: map), Hore (1892: 168, map), and Cunnington (1906a: 130) during the ‘Third Tanganyika Expedition’, 1904–1905. On the map of Hore (1882: map), there is a small island off the north eastern coast of Ubwari Peninsula, referred to as Mzimu Island. Today, a remnant of the name Muzimu exists, as one of the northernmost capes of Ubwari Peninsula is referred to as Cape Muzimu.
Following the collapse of the Lukuga blockage in 1877–1878, the lake level rapidly decreased by nine metres in six years and then again by two metres from 1886 to 1894, after which it reached its current level, i.e., a surface elevation of about 773 metres above sea level (Camus, 1965: 1244). However, minor lake level fluctuations still prevail in Lake Tanganyika. For example, between November 1997 and April 1998, during the so-called El Niño rains, the lake level rose by at least two metres (pers. obs.; cf. Büscher, 1998a: 789; cf. Sturmbauer et al., 2003: 52). This information was also posted on our website in 1998 as a reference to diminishing rocky habitats, exemplified by the natural habitat of Tropheus “Red Namansi” in Mtosi Bay, which, prior to the El Niño rains, had shrunk by more than 50% compared to the situation in 1991 when that Tropheus variant was discovered (see more in Karlsson & Karlsson, 2016a; 2016b). Furthermore, currently (July 2018), the lake level appears to be exceptionally high, about one metre higher than that of April 1998, at about the same level as that of 1989 (pers. obs.). Apparently, in the past 30 years, Lake Tanganyika has fluctuated by about three metres.

Fig. 72. Telmatochromis vittatus (left) and a sub-adult Neolamprologus christyi (right) at Kasola Island. Young individuals of N. christyi have beautiful blue fins, one of several features distinguishing the species from the similar N. sp. “Eseki” (§ 7).

17. The age of Neolamprologus sp. “Mwila”
N. sp. “Mwila” was only observed at Mwila Island but is potentially also found at the eastern shore of the neighbouring Kerenge Island and along the mainland opposite to the islands, referred to as Mbwilo Bay (Fig. 73), where the habitat possibly includes empty Neothauma shells. The depth around the Kipili islands is normally 50–100 metres, except between the two groups of northern and southern islands, where it is probably a bit deeper. Furthermore, at the western edge of the archipelago, it quickly gets deeper than 100 metres [the maximum depth of the lake, 1470 metres, is located near the Congolese coast opposite to the Kipili Archipelago (Tiercelin & Mondeguer, 1991: 10, Fig. 2)]. If N. sp. “Mwila” only exists in the Mwila Island area, it is supposedly a young species not older than the latest lake fluctuation that reduced the lake level by about 100 metres or more. If the lake level was 160–180 metres below the current level about 14 thousand years ago, beginning to rapidly rise two thousand years later, Lake Tanganyika may have reached its current level approximately nine thousand years ago, see references above (§ 16). By 14 thousand years ago, the Kipili Archipelago appears to have been dry land, but about five thousand years later, its current geography rapidly took form. As it appears, all endemic species and variants in the Kipili area may be no more than about nine thousand years, which includes N. sp. “Mwila”. This short duration of only nine thousand years is probably not enough for any major trophic, morphological or behavioural adaptation to have evolved. Due to this limited age, N. sp. “Mwila” is expected to resemble and behave like N. sp. “Eseki”, from which it presumably derives. Also, the genetic divergence between these two species is expected to be limited. Low levels of mtDNA sequence divergence have recently been found for the Alcolapia species flock from Lakes Natron and Magadi, Tanzania and Kenya. These so-called Soda tilapias of the highly alkaline and hot small lakes are estimated to be no more than nine thousand years (Seegers et al., 1999, cited in Takahashi, 2004: 427). A similarly small genetic divergence was detected between the rock-dwelling and shell-dwelling morphs of T. temporalis in southern Zambia (Takahashi, 2004: 427), see above (§ 12). Furthermore, it is suggested that Lake Victoria and most of its drainage more or less dried out about 18 to 14 thousand years ago and that one or a few closely related founding haplochromin cichlid species thereafter evolved into the lake’s current species flock of at least 500 distinct species (Meyer et al., 1990; Nagl et al., 2000; Verheyen et al., 2003; 2004; Elmer et al., 2009; Pennisi, 2018; cf. Fryer, 2001).
In conclusion, it appears that N. sp. “Mwila”, the rock-dwelling and shell-dwelling T. temporalis-like species in southern Zambia, the species of the Alcolapia flock, and the cichlids of the entire species flock of Lake Victoria are all young divergent distinct forms of similar age, some with species status, others referred to as morphs.

Fig. 73. Dusk approaches over the shallows in Mbwilo Bay near Kirando, with Mswa Hill visible in the distance. This area is to a large extent unexplored. About 9000 years ago, the Kipili Archipelago was mostly dry land, so the endemic species to this area, e.g., Lepidiolamprologus kamambae, Neolamprologus sp. “Mwila”, and N. sp. “Cygnus Mwila” are therefore presumably young species. For a characterisation of N. sp. “Cygnus Mwila”, see an upcoming article.

18. Endemic species of the Kipili area, including Lepidiolamprologus kamambae
Apparently, the modern Kipili Archipelago is geologically and hydrologically rather young, about nine thousand years. The area constitutes a barrier of distribution for many species and variants, see references above (§ 5). Furthermore, it is home to some locally endemic variants, such as Ophthalmotilapia boops “Neon Stripe”, and species, such as N. sp. “Mwila”, Cyprichromis sp. “Kipili Zebra” (Karlsson & Karlsson, 2017k), and Lepidiolamprologus kamambae (Kullander et al., 2012; Karlsson & Karlsson, 2013). Confirmed observations of L. kamambae have been made at Kamamba Island and the adjacent Kerenge Island, both islands which are part of the northernmost Kipili islands (§ 5). However, during our expeditions in 2008, we briefly noticed and managed to take a snapshot of a black and white torpedo and L. kamambae-like fish in the shallow rocky habitat at Mswa, just north of the Kipili area. Contradictory to our observations, a photo in Konings (2015a: 153, Fig. bottom) is described as to illustrate L. kamambae at Lupita Island, which is part of the southernmost Kipili islands. Due to its relatively large eye, the specimen in the photo appears to be a juvenile or sub-adult. However, we strongly oppose the specific identification, the photo does not illustrate L. kamambae, but a very cryptic L. elongatus-like species found between the southern Kipili islands and Mtosi Bay, possibly farther south. This cryptic species appears to be a distinct species and a member of what may be referred to as the L. elongatus species complex. For example, the black blotches on the trunk have more of a vertical than horizontal orientation, and it lacks the lateral rows of pearly dots possessed by L. elongatus. Furthermore, due to the blackish blotches on the silvery white body, it is very cryptic to L. kamambae, but adults have a deeper body and yellowish submarginal stripe on the dorsal and anal fins (compare illustrations in Karlsson & Karlsson, 2014b; Mierzeńska, 2016: 25, Figs. top and centre). This L. elongatus-like species, which here is tentatively referred to as L. cf. elongatus “Lupita–Mtosi”, is not sympatric anywhere with L. kamambae.
In the northern part of the Kipili Archipelago, including Kamamba Island and Kerenge Island, where L. kamambae naturally occurs, there is another L. elongatus-like species. The latter has a yellowish colouration, therefore is easily distinguishable from the sympatric L. kamambae (compare illustrations in Karlsson & Karlsson, 2013: 2–3, Figs. 1 and 2).
Besides the incorrect identification of the L. elongatus-like species at Lupita Island, Konings (2015a) has chosen to rename the type locality of L. kamambae (Kamamba Island). This is a very peculiar proceeding as it implies that the authors of the taxon, L. kamambae (Kullander et al., 2012), did not seriously investigate the correct and proper name for the type locality prior to the publication. For the record, the name Kamamba is the official name of the island featured in the official maps of Tanzania (Government of Tanzania, 1984). In addition, most local Tanzanian citizens that we, the authors of L. kamambae, have spoken with during our stay in the area (from 1989 onwards), refer to the island as Kamamba. Furthermore, prior to our introduction of L. kamambae and Kamamba Island, no one in the cichlid community, neither the scientific nor the hobby, had ever reported anything about L. kamambae or Kamamba Island, at least not in any well-known and easily accessible publication (obscure communities in social media do not count), hence, there was no history of multiple and conflicting names like there is with many other Lake Tanganyika locality names (§ 32.1, 32.8, 32.9). Then why introduce an alternative name/spelling? Why not just accept this seemingly sensible, perfectly appropriate, and officially valid name? It is just bizarre. For additional maps indicating Kamamba Island, see, for example, Janssens et al. (2005a: map), Basel Mission Archives (2018a: map), Luscombe (2018d: map), UT (2018: map).
Furthermore, for a correct identification of a juvenile (3 cm) L. kamambae, see Karlsson and Karlsson (2013: 7, Fig. bottom). The preceding publication features about 40 illustrations of mostly different torpedo-like species and variants, including L. kamambae, L. kendalli, and many L. elongatus-like species.
In addition, due to its confined geographical distribution, L. kamambae appears to be, like N. sp. “Mwila”, a young species, with an age of about nine thousand years or less.

19. Species
Many of the disagreements about the validity of certain species taxa appear to derive from the confusion of what such species really are. A species taxon is nothing more than its recognisable characters, whether these are morphological, physiological, molecular, behavioural, etc. Two or more isolated populations or other aggregates of organisms are never naturally one, they are only one on grounds that make sense to an observer, therefore subjectively synthetic. Giving names to organismal objects is the ultimate foundation in taxonomy. While populations are the true evolutionary entities, the species is a tool to sort them into genealogically related taxonomical classes, also known as taxa, which importantly contribute to the understanding of the organismal world (see more in Karlsson & Karlsson, 2018a: 35–57).

Fig. 74. What constitutes a species is addressed in detail in ‘Tanganika Magazyn’, no. 22 (Karlsson & Karlsson, 2018a: 35–57). Apparently somewhat ahead of his time, Darwin (1859) regarded species as mere labels applied to a divergence process. Unfortunately, the advent of the so-called ‘Modern Synthesis’ turned Darwin’s advance in the opposite direction, attributing a superior position to the species rank and introduced the species as the unit of evolution (Mishler, 2010). More information about ‘Tanganika Magazyn’, no. 22, here, here, and here (video).

20. Generic assignment of Neolamprologus sp. “Eseki” and N. sp. “Mwila”
Neolamprologus has three different definitions, one which is widely adhered to (Poll, 1986), another which apparently conforms to a more natural classification (Stiassny, 1997), and a third, rather modest definition, which is the original (Colombé & Allgayer, 1985). The generic assignment of N. sp. “Eseki” and N. sp. “Mwila” to Neolamprologus follows all three definitions. (1) Both species have pelvic fins in which the first ray is the longest, hence following Poll (1986). (2) N. sp. “Eseki” and N. sp. “Mwila” closely resemble N. mondabu, N. christyi, N. modestus, and N. petricola. These four species all lack a fully ossified labial cartilage (a discrete bone within the dentary), which supposedly also apply to N. sp. “Eseki” and N. sp. “Mwila”, thereby conforming to the definition of Stiassny (1997). (3) Furthermore, regarding the same four species, the infraorbital section of the cephalic lateral line has a reduced series comprising a single element, the lachrymal, and never more than 60 lateral scales in a longitudinal series, which is how the genus originally was defined by Colombé and Allgayer (1985). While Poll (1986: 55) initially noted the polyphyly and mixture of various lamprologin species in his definition of Neolamprologus, the classification of Colombé and Allgayer (1985) was a rather modest contribution, gathering almost all the former lacustrine Lamprologus species in a new single genus, Neolamprologus. Nonetheless, restricting the genus Lamprologus to the Congo River drainage is undoubtedly a step in the right systematic direction (Colombé & Allgayer, 1985). The monophyly of Lamprologini is firmly established (Sturmbauer et al., 1994; Kocher et al., 1995; Stiassny, 1997; Takahashi et al., 1998; Salzburger et al., 2002a; 2002b; Clabaut et al., 2005; Day et al., 2007; Koblmüller et al., 2008; Muschick et al., 2012; Meyer et al., 2015; Irisarri et al., 2018), but the intra-tribal relationships remain problematic and require revision (Schelly et al., 2006; Sturmbauer et al., 2010; Weiss et al., 2015).

Fig. 75. Neolamprologus christyi at Molwe, depth 5 metres.

21. A species without a name
Over the years, we have discovered new species in the Kipili area, including Lepidiolamprologus kamambae (Kullander et al., 2012; Karlsson & Karlsson, 2013), N. timidus (Kullander et al., 2014a), Chalinochromis cyanophleps (Karlsson & Karlsson, 2012b; Kullander et al., 2014b), Cyprichromis sp. “Kipili Zebra” (Karlsson & Karlsson, 2017k), and N. sp. “Mwila”, as well as variants including the more yellowish N. sp. “Eseki” (Zadenius, 1991b). In the early 1990s, a deep-living new lamprologin species was found, which has the intermediate appearance of N. ventralis and N. caudopunctatus, including a silvery deep body with bluish lateral streaks. It was found together with Cyphotilapia gibberosa along the southern islands. The species was collected only once, a male and a female, but no photos of them exist. Currently, the species has no name and the exact location is forgotten.

22. J. E. S. Moore and the Tanganyika expeditions
The occurrence of marine-like animals in Lake Tanganyika was more or less the sole reason for the realisation of the two East African expeditions in 1895–1896 and 1899–1900. Referred to as the ‘First and Second Tanganyika Expedition’, the first was financed by the Royal Society and the British Association for the Advancement of Science, the second by the Royal Geographical Society (only £600) and private donors (at least £4,400) (Moore, 1898c; 1903: 3, 7; Cunnington, 1920; Troyer, 1991: 43). Both expeditions were led by J. E. S. Moore, a cancer researcher and zoologist of the Royal College of Science (now Imperial College London) (Troyer, 1991: 41), whose rather famous initials stand for John Edmund Sharrock (Fig. 139), see more below (§ 32.24). The expeditions were organised by Professor Edwin Ray Lankester (Moore, 1903; 3, 8), a friend of Moore and supporter of his theory regarding a marine origin of the fauna of Lake Tanganyika, also a future director of the Natural History branch of the British Museum. According to Troyer (1991: 38), nearly 200 new species were added to the known fauna of the region.
At Kinyamkolo (Mpulungu, Zambia, § 32) in 1895–1896, Moore collected fish, frogs, sponges, snails, crabs, among others (§ 32.19) and his collections of fish, mostly cichlids and catfish, were eventually deposited at the museum and to the zoologist George Albert Boulenger (British Museum, Natural History, 1906: 544). “Mr. Moore’s first collection, made in 1895–96, contained about 90 fishes, referred to 33 species, 25 of which were described as new” (Boulenger 1906: 539), of which, 20 were cichlids and six were lamprologins. Of these six, one was given the name Lamprologus modestus, now referred to as Neolamprologus modestus (Boulenger, 1898a: 494; 1898b: 8, Plate 1, Fig. 5; 1901a: 400).

23. Neolamprologus modestus and the perception of a lake-wide distribution
Characteristically, N. modestus in a strict sense is a black, dark brown or beige species with yellow pectoral fins and a rounded or subtruncate caudal fin with rounded corners. The natural distribution lies in the southern part of the lake, between western Zambia (Cape Chipimbi) and southern Tanzania with gaps between Cameron Bay and Mbete Bay, and Kilewani and Kala Bay, see more below (§ 27, 30). Besides the type localities of Kinyamkolo and Mbity Rocks (§ 32.22, 32.25, 32.27), where Moore collected one specimen each, additional localities where N. modestus allegedly has been found were subsequently reported, hence, the perception was that N. modestus is a lake-wide species. However, while some collections have indeed been N. modestus or N. modestus-like, others have been N. mondabu, a distinct species described by Boulenger (1906: 557) eight years later, but repeatedly confused and occasionally synonymised with N. modestus.
Célestin Hecq, a captain in the ‘Force Publique’ of the Congo Free State, collected a single specimen at Albertville (Mtoa, DR Congo, see, for example, Poll, 1946: 351; Coosemans, 1948: 461; Konings, 2013a: 24–25; 2013d), which was assigned to N. modestus by Boulenger (1901a: 401). Not necessarily N. modestus, but most likely at least N. modestus-like, this presumably badly preserved specimen of 84 mm was deposited in 1899 in the collection at the ‘Museé du Congo’, the precursor of ‘Musée royal de l’Afrique centrale’, Tervuren, Belgium (Poll, 1946: 333). Eventually, the specimen was indicated as ‘skeleton’ in the ‘Catalogue of the fresh-water fishes of Africa’ (Boulenger, 1915: 470).
Moore’s “series of sketches executed from fresh specimens” enabled Boulenger (1898b: 2) “to represent some of the new species in their natural colours”, such as Cyathopharynx furcifer (see more below, § 32.20). However, N. modestus was apparently not one of these species, because its conspicuous yellow pectoral fins were not mentioned in the description (cf. Boulenger, 1898a: 494; 1898b: 8, Plate 1, Fig. 5; 1901a: 400). Likewise, a young specimen of N. modestus from Komba Bay (Nkamba Bay, cf. Boulenger, 1906: 538, map) collected by Cunnington in 1904–1905 was not reported as to have yellow pectoral fins (cf. Boulenger, 1906: 558), which individuals from this locality indeed have, at least adults (cf. Staeck, 2014: 135, Fig. bottom).
Specimens collected at Uvira, DR Congo, by Guy Babault were assigned to N. modestus by Pellegrin (1927: 500). Presumably, they were N. mondabu, whose type locality is Kaboge, about 45 km farther south. Similarly, a small collection of fishes was obtained by Harry C. Raven during the Universal Films Co. expedition to East Africa in 1920. Caught at Kigoma by means of a small seine in shallow water, one specimen of 32 mm in standard length (SL) was assigned to N. modestus and four (60–80 mm, SL) to N. mondabu (Myers, 1936: 14). The single specimen identified as N. modestus may have been a badly preserved N. mondabu. Collected by Arthur Lestrade at Rumonge, Burundi, two specimens of a maximum length of 53 mm were identified as N. modestus by David and Poll (1937: 273), but should likely have been identified as N. mondabu. Partly debatable, Poll (1952b: 120–121) reported the sighting of N. petricola at Mtoto, N. modestus at Mtoto and Kabimba, and N. mondabu at Mtoto, Kabimba, Kalemie, and Kigoma.

Fig. 76. Serenity prevails in the early morning at Moyobozi (30 December 2007), a lakeshore fishing village located in the Uvinza District (Kigoma Region), just north of the Malagarasi River estuary. The short rocky coastline just south of the village is the endpoint for the distribution southwards for several species, including Tropheus brichardi. For a photo and video sequence of T. brichardi south of Moyobozi, see Karlsson and Karlsson (2015c: Fig. 7; 2017j: 6 min 11 sec).
Fig. 77. Neolamprologus mondabu at “Moyobozi”, depth three metres.
Fig. 78. Neolamprologus sp. “Modestus Mpimbwe” at Mwaka, Korongwe Bay.
Fig. 79 (left). N. mondabu at Cape Kabogo. Fig. 80 (right). N. mondabu at Segunga.

24. Neolamprologus modestus and further confusion involving N. mondabu and N. petricola
According to Boulenger (1906: 558), N. mondabu is distinguishable from N. modestus “by the slightly emarginate caudal fin [vs. subtruncate with rounded corners] and the shorter mouth”. Furthermore, Herrmann (1987: 87) stated that N. mondabu has a shallower body than N. modestus. In addition, Herrmann (1987: 87) considered classifying the variant with the yellow pectoral fins as another species, distinct from N. modestus. However, the one with the yellow pectorals is the original N. modestus, as illustrated by a photo of the species at or near its type locality, Mbity Rocks (Mbete/Kasakalawe) in Konings (2015a: 237, Fig. centre).
Between 1956 and 1978, when Poll and his followers regarded N. mondabu as synonymous with N. modestus, obviously most illustrations and descriptions of N. mondabu in books and articles are referred to as N. modestus. For example, Berglund’s field observation (1976) of “N. modestus” as one of the most common fish species in the shallow, sandy, and rocky biotope of Kigoma likely relates to N. mondabu, though his statement is accompanied with a photo taken in an aquarium of an N. modestus-like species. Indeed, N. mondabu is common among rocks (cf. Jonas, 2015: 7). This synonymisation, as well as common misidentifications, continued long after the reinstatement of N. mondabu in 1978, which is evident in a long series of books, articles, and scientific papers (see below, this §). The photos in Axelrod (1973: 42, Fig. bottom; 1975: 274, 275), partly reproduced in Axelrod and Burgess (1986: 316, Fig. bottom), illustrate specimens with a somewhat rounded caudal fin, accordingly entitled N. modestus. However, due to the obviously damaged caudal fins, which supposedly will be slightly emarginate upon recovery, the dark brown and drab body, yellowish dorsal margin, and only moderately pearly scales, they should have been entitled N. mondabu. Likewise, Kahl (1971: 17, Fig. top) illustrated N. mondabu or an N. mondabu-like species, recorded as N. modestus, while Brichard (1978a: 208, Fig. top) stated that “the southern race of Lamprologus modestus has yellow pectoral fins”, illustrating a specimen of an N. mondabu-like species. Regarding another specimen, obviously of N. mondabu or an N. mondabu-like species, Brichard (1978a: 229, Fig. top) referred to it as an unidentified species. In an inventory study of the fish species at Cape Chipimbi, Brichard (1978b) stated that N. modestus is found throughout the lake, including at Cape Chipimbi, where it was frequently observed. Furthermore, at that same location, Brichard (1978b: 204) also reported the presence of a species referred to as L. staecki, a pseudo-scientific non-available species name, synonymous with the original N. modestus with yellow pectoral fins (Poll, 1978: 745), see more below (§ 28). Hence, it appears from Brichard’s report (1978b) that both N. mondabu or an N. mondabu-like species (referred to as N. modestus) and N. modestus (L. staecki) occur at Cape Chipimbi. However, at Moliro, just about 10 km farther north, there is an all-black N. modestus-like species, lacking the yellow pectoral fins, which we have exported and referred to in the 1990s as N. petricola, and at Livua, 13 km north of Moliro, the Belgian Hydrobiological Mission caught two specimens in March 1947, which were later designated as paratypes in the description of N. petricola (Poll, 1949: 37, 40); see more about N. petricola in § 26. The yellow pectoral fins of N. modestus are also found in the nearby populations of Ndole Bay (Maßmann, 1986), Nkamba Bay (Staeck & Linke, 1994: 105–106), and Chisanse (Konings, 2005: 123, Fig. top), western Zambia.
Further reports regarding the misidentification of N. mondabu (referred to as L. modestus or N. modestus) include Kawabata and Mihigo (1982: 139, Table 1), who listed “N. modestus” as occurring at Uvira, DR Congo. Furthermore, Hori (1983: 129, cited in Lowe-McConnell, 1987: 95), Hori et al. (1983: 30, Table 4), and Nagoshi (1983: 39; 1985; 1987) reported the sighting of “N. modestus” at Luhanga, about 15 km south of Uvira. Yamaoka (1985: 254) and Gashagaza et al. (1986: 37–38) made observations of “N. modestus” at Uvira, Luhanga, and Pemba (26.5 km south of Uvira). Nakanishi and Nagoshi (1987) reported the mating system of “N. modestus”, Takemon (1989) described the feeding, territorial, and reproductive behaviour of “N. modestus” at Pemba, while Gashagaza (1991) reported the reproductive behaviour of “N. modestus” at Pemba. Eventually, Nakai (1993) acknowledged that “N. modestus” in most previous reports regarding the northern part of the lake have been misidentified and were in fact N. mondabu, reporting the spatial distribution of spawning sites of N. mondabu at Pemba (referred to as Mbemba). Yuma (1993: 217) reported that N. mondabu at Pemba digs in the sand and feeds on chironomid larvae. Likewise, Burns (1993: 243), in a study on ecological communities, described how both N. mondabu (the predator) and chironomid larvae (the prey) are more numerous when they coexist.
Conversely, Ota and Kohda (2014) described the food habits of N. modestus at Mbita Island, Zambia (referred to as Nkumbula Island), but referred to the species as N. mondabu. Similarly, in the same locality, Ota et al. (2014) made an observation of the reproductive behaviour of N. modestus, erroneously referred to as N. mondabu.
In a molecular study on Lamprologini, Sturmbauer et al. (2010) referred to the N. modestus-like populations west of Mbete Bay as N. petricola, but to those east of the bay as N. modestus. In more detail, the populations at Kapembwa and south of Katoto were designated as N. petricola, while whose at Kasakalawe (map, Fig. 138) and ‘Kalambo Lodge’ (south of Kalambo River) were designated as N. modestus. Morphologically, the populations are largely identical, but those referred to as N. petricola have black pectoral fins, while those referred to as N. modestus have yellow pectoral fins (Sturmbauer et al., 2010). However, Kohda et al. (1996: 241, Fig. 2) observed the same Zambian populations and referred to all of them, regardless of pectoral-fin colouration, as N. petricola. As a means to separate them in communication, the N. modestus-like Zambian populations with black pectorals may be referred to as N. cf. petricola “Zambia”.
Furthermore, importers of Lake Tanganyika cichlids occasionally referred to N. modestus-like species as N. mondabu (e.g. Reiber, 1995: 5), while the ornamental fish collectors at Kigoma referred to N. mondabu or an N. mondabu-like species as N. sp. aff. modestus (Seehausen, 1991: 13). In pure aquaristic literature, many misidentifications occur, such as the illustration of N. cunningtoni in Mayland (1978: 239), referred to as N. modestus.

Fig. 81 (left). Neolamprologus sp. “Eseki”, Cape Mpimbwe, NRM 57982. Fig. 82 (right). N. sp. “Modestus Mpimbwe”, Cape Mpimbwe”, NRM 59564.
Fig. 83 (left). Neolamprologus sp. “Eseki”, Kansombo, NRM 60257. Fig. 84 (right). N. sp. “Modestus Mpimbwe”, Kansombo”, NRM 51515.
Fig. 85 (left). Neolamprologus petricola, male, Tembwe, NRM 70258. Fig. 86 (right). Neolamprologus petricola, female, Tembwe, NRM 70258.
Fig. 87 (left). Neolamprologus cf. modestus, Kitawe, NRM 51588. Fig. 88 (right). N. christyi, Kitawe, NRM 60279.
Fig. 89. Its specific name meaning ‘dweller among rocks’, Neolamprologus petricola was described by Poll (1949) partly based on specimens he collected at Mtoto, DR Congo in December 1946. Despite a rather unclear diagnosis, the species potentially being synonymous with N. modestus, Poll (1978: 755) listed N. petricola as valid pending further systematic work, pronouncing synonyms involving cryptic species must only be proposed with great caution (Poll, 1978: 727). Distinguishing features of N. petricola include a steeply sloping head profile, obviously displayed by the young individual in the photo, which was collected by our team at Mtoto early in 2000, photographed in aquarium within a month of capture at our facilities in Dar es Salaam.

25. Lamprologus modestus nyassae (Borodin, 1936: 23)
Irresponsibly regarded as a subspecies of N. modestus, Borodin (1936: 23) introduced Lamprologus modestus nyassae based on specimens collected by Arthur Loveridge. The type locality was recorded as Mwaya, Lake Malawi, where Loveridge collected frogs and snakes in March 1930. The location of Mwaya is frequently verified or indicated to be in Tanzania (e.g. Loveridge, 1960: 280; Eschmeyer et al., 2018), but according Loveridge (1933a: 15; cf. 1933b: 394), “Mwaya is just north of Karonga, Nyasaland [Malawi] on the northwest shore of Lake Nyasa, actually the village is separated from the lake by about a mile of swamps”. Based on this statement, one may get the impression that Mwaya is located in Malawi, but Mwaya is indeed located in present-day Tanzania, as evident in the following three maps: Loveridge (1933a: Plate 1), REAAA (1926), and Eccles and Trewavas (1989: 15, 17).
In 1936, Trewavas (1946: 243; Poll, 1946: 350) examined the types described by Borodin as L. modestus nyassae and found them to be small specimens of L. reticulatus (Boulenger, 1906: 560, Pl. 35, Fig. 4), the latter name which is now widely regarded as a synonym of L. callipterus (Boulenger, 1906: 559, Pl. 36, Fig. 4). As Lamprologus is not found in Lake Malawi, the attribution of Loveridge’s specimens to that lake is obviously an error of labelling. The scientific expedition of Loveridge (§ 32.7) made several stops along the shores of Lake Tanganyika, including Mpulungu, Kasanga, Kipili, Sumbwa (near Ikola), and Ujiji. As most or all of Loveridge’s Tanganyika fish described by Borodin were indicated as being collected at Ujiji, also L. modestus nyassae likely hails from Ujiji, which is in the northern half of the lake. However, all types of L. callipterus and L. reticulatus were collected in the southern half of the lake, i.e., Mpala and Niamkolo (Mpulungu, § 32), and Livua, respectively (Boulenger, 1906: 559–560).
The repeated low lake stands, which resulted in the temporary formation of at least three smaller sublakes, one each in the northern, central, and southern part of the lake basin (§ 16), likely had a major impact on the evolution of the fish fauna of Lake Tanganyika (Poll, 1952c; Poll & Matthes, 1962). Due to geographical isolation and allopatric speciation, many cichlids, including the lamprologins, have at least a northern and southern representative of a particular lineage or species. Some of these representatives have the status or potential status of distinct species, such as Cyphotilapia frontosa (in the north) and C. gibberosa (in the south), Telmatochromis bifrenatus and T. vittatus, N. tetracanthus and N. brevianalis, and Petrochromis orthognathus and P. sp. “Orthognathus Ikola”. Possibly, also L. callipterus has a northern and southern representative.
In a molecular study to describe the evolutionary history of certain species, including L. callipterus (Nevado et al., 2009), mtDNA data revealed that L. callipterus includes two divergent and geographically disjunct mtDNA lineages, one in the north and another in the south. These two mtDNA lineages diverged roughly about one million years ago, coinciding with a major water level low stand in Lake Tanganyika, which divided the lake into at least two isolated sublakes. It was concluded that the two mtDNA lineages originated as the result of ancient hybridisation, as well as the separation and isolation of L. callipterus populations in the two sublakes (Koblmüller et al., 2007; Nevado et al., 2009).
If these northern and southern representatives of L. callipterus prove to be distinct in more standardised taxonomical characters as well, they may be regarded as two distinct species. Konings (1988: 202) observed a distinction between them, with those in the south having a distinct black margin in the caudal fin, while those in the north lack this feature, hence, it appears that the northern and southern groups of populations of L. callipterus can be separated by both mtDNA and melanin characteristics, and potentially regarded as two distinct species. If so, then L. callipterus represents the southern populations and L. modestus nyassae the northern. However, although L. modestus nyassae is not synonymous with L. callipterus and L. reticulatus, it may still be a synonym, because in 1909, Steindachner (1909a: 403) described Julidochromis elongatus, an L. callipterus-like species (Konings, 1988: 202) from Ujiji. Indeed, Boulenger (1915: 460, 471) reassigned J. elongatus to Lamprologus, which resulted in the new name combination, L. elongatus (Steindachner, 1909a: 403) being a secondary homonym, preoccupied by L. elongatus (Boulenger, 1898a: 494; 1898b: 9), therefore, the replacement name L. steindachneri was suggested (Boulenger, 1915: 471). However, following the introduction and establishment of Lepidiolamprologus (Pellegrin, 1904: 163, 295; Colombé & Allgayer, 1985), to which Lamprologus elongatus (Boulenger, 1898a: 494; 1898b: 9) was designated type species, L. elongatus (Steindachner, 1909a: 403) is no longer a secondary homonym, therefore available to be applied to the northern populations of “L. callipterus” should these one day be more widely regarded as a species distinct from L. callipterus (see more below*). In conclusion, L. modestus nyassae (Borodin, 1936: 23) is a junior synonym of L. elongatus (Steindachner, 1909a: 403), the northern representative of “L. callipterus”.

* According to the nomenclatural code of ICZN (1999: Art. 59.3) regarding “Secondary homonyms replaced before 1961 but no longer considered congeneric”, it is stated that “[A] junior secondary homonym [Lamprologus elongatus (Steindachner, 1909a: 403)] replaced before 1961 is permanently invalid unless the substitute name [Lamprologus steindachneri (Boulenger, 1915: 471)] is not in use [it is not in use] and the relevant taxa are no longer considered congeneric [while Lamprologus elongatus (Boulenger, 1898a: 494) is assigned to Lepidiolamprologus, Julidochromis elongatus (Steindachner, 1909: 403) will be assigned to Lamprologus, should the species be regarded as distinct], in which case the junior homonym [Lamprologus elongatus (Steindachner, 1909a: 403)] is not to be rejected on grounds of that replacement”.

Fig. 90 (left). Neolamprologus sp. “Eseki”, Kansombo, male. Fig. 91 (right). N. sp. “Eseki”, Kansombo, female.
Fig. 92 (left). Neolamprologus sp. “Modestus Mpimbwe”, Kansombo, male. Fig. 93 (right). N. sp. “Modestus Mpimbwe”, Kansombo, female.
Fig. 94 (left). Neolamprologus christyi, Kitawe, male. Fig. 95 (right). N. christyi, Kitawe, female.
Fig. 96 (left). Neolamprologus cf. modestus “Lwasase Point – Kilewani”, Kitawe, male. Fig. 97 (right). N. cf. modestus “Lwasase Point – Kilewani”, Kitawe, female.

26. Neolamprologus petricola and confusion with N. modestus and N. mustax
If N. modestus is restricted to the phenotype found at the type localities (Kinyamkolo and Mbity Rocks, § 32.27), consisting of a black, dark brown, or beige body with yellow pectoral fins, additional species may be recognised. Obviously, they will all be N. modestus-like, presumably closely related, and potentially constitute a species complex (see below, § 30). One such species is N. petricola, a stenotopic species found in a rocky habitat, which grows to about 12 cm and feeds mostly on insect larvae (Hori, 1991). The type series of N. petricola was collected at Mtoto, Albertville (Kalemie), and Livua (Poll, 1949: 40; Walschaerts, 1987: 33). In the 1990s and 2000s, we (African Diving) collected and exported several N. petricola variants from DR Congo, labelled according to their collection location, including N. petricola “Cape Tembwe”, “Kiku”, “Moliro”, “Mtoto”, “Tembwe”, and “Zongwe”. We did not observe any conspicuous distinction between them; they were nearly identical with a blackish body and fins. Characteristically, these seven populations all have the same general body depth and sharply descending head profile (pers. obs.) like the holotype from Mtoto (cf. Poll, 1949: 38, Fig. 19), the latter feature which may be the most typical for N. petricola (Fig. 89). In addition, adults frequently have a frontal hump. Seemingly, N. modestus and very similar species/variants either lack an obvious frontal hump or have only an indication of one (§ 27). They also appear to lack the steep head profile normally possessed by N. petricola. Our collected specimen of N. petricola from Tembwe (Figs. 85–86) is more or less identical to that of Poll’s (1956: Pl. 9, Fig. 4) from Mtoto (station 108). Also, the Tembwe population was identified as N. petricola by Büscher (1998a), but the illustrated specimen from Moliro in Konings (2015a: 240, Fig. 4) appears to have a rather elongated head. Nonetheless, pending further research, the geographical distribution of N. petricola, which was restricted to the south western coast of DR Congo by Maréchal and Poll (1991: 288), presumably lies between Kalemie and Moliro. While the populations north of Kalemie, of which there are insufficient data, may temporarily be referred to as N. cf. petricola “Congo North”, those south of Moliro (at Cape Chipimbi) and farther south and southeast are referred to as N. modestus or N. cf. petricola “Zambia” (§ 24).
According to Poll (1956: 557), N. petricola resembles N. modestus, but is easily distinguishable from it due to the occasionally indented caudal fin and more elongated body of N. modestus. However, these characteristics obviously concern N. mondabu, which Poll (1956: 484) erroneously regarded as synonymous with N. modestus. In fact, N. modestus has an indistinguishable caudal fin and, contrary to Poll’s statement, a seemingly less elongated body than N. petricola (cf. Figs. 85 and 87). Interestingly, Poll chose the shape of the caudal fin as a distinguishing character between N. modestus and N. petricola, where the shape of the caudal fin is very similar or possibly identical, but not between N. modestus and N. mondabu, where it is conspicuously different.
After Poll (1978: 727) had realised his mistake regarding the synonymisation, he expressed that some species had not been diagnosed with sufficient clarity, including N. petricola. Nonetheless, based on subtle characters, N. petricola, as well as other poorly diagnosed species, was indicated as valid and included in the synopsis of lamprologin species. N. petricola, which obviously is very similar to N. modestus, may be such a cryptic species that it is better considered valid pending further systematic work. Indeed, due to the presence of cryptic species, synonyms must only be proposed with great caution (Poll, 1978: 727). Of course, synonymisation should not be based on feeling but on thorough examination of voucher specimens and field observations, ideally “[c]riticism of particular papers must be restricted to scientific publications, and should be backed up with facts” (Kullander, 2012: 2). For a detailed differential diagnosis concerning N. modestus and N. petricola, see the synopsis in Poll (1978: 755).
Brichard (1978: 282–283) described N. petricola as reaching a maximum size of 130 mm, possessing molar-shaped pharyngeal teeth, comprising a yellow body, and probably being restricted to the southern part of the lake. While the description of the length and teeth refers to N. petricola, that of the body colour and geographical distribution refers to N. mustax (Poll, 1978), a species which was repeatedly confused with N. petricola prior to its description. Presumably, when Brichard collected the type series for N. mustax in 1976–1977, he must have thought they were specimens of N. petricola, although a photo of a collected specimen was described as an unidentified species (Brichard, 1978: 209, Fig. top). The molar-shaped pharyngeal teeth of N. petricola possibly indicate a diet composed of molluscs, but examination of the stomach contents revealed the presence of insect larvae (Brichard, 1978: 283).
Additional misidentifications include Neergaard (1976: 95, Fig. top; 1977: 92, Fig. bottom; 1982: 92, Fig. top), Staeck (1977: 155, Fig. 99, 100), Mayland (1978: 241, Fig. top), and Stawikowski (1979), who illustrated specimens of N. mustax, but referred to them as N. petricola. Also, as a more current error (as of 1 January 2019), the species profile of N. petricola by FishBase (Froese & Pauly, 2019) is accompanied with a photo of N. mustax.
Prior to its description, most N. mustax specimens in the trade were referred to as N. petricola (Forsberg, 1982: 16). Based on colours and body shape, some authors (e.g. Thurén, 1982: 14) thought that they looked rather similar. Partly due to this general confusion, Poll (1978: 730) compared the proportions of N. mustax with N. petricola, and to some extent with N. modestus. Herrmann (1987: 88) pointed out the previous confusion regarding N. petricola and N. mustax, the great similarity between N. petricola and N. modestus, and that N. petricola, as of 1987, never had been collected and exported. The first collection of N. petricola by our team (African Diving), including the variant of Mtoto, was made in late 1997 and exported the following year. In addition, at a few occasions in the early 1990s, we exported the variant from Kansombo, north of Kabwe, Tanzania, as N. petricola “Kansombo”. The populations at Kansombo and that at the opposite coast of the lake, at Mtoto, DR Congo, both have a black body and frontal hump, and appear rather similar. However, today we refer to the Kansombo variant as N. sp. “Modestus Mpimbwe”, see below (§ 29). Regarding some N. mustax-like variants, around 2000 we found and exported several interesting variants from southern DR Congo, including one with a blue chin, N. sp. “Leleupi Blue Chin” (Fig. 101), and another with a solid bright orange body, the latter which hails from Lunangwa and appears to conform to a species currently referred to as N. sp. “Leleupi Kapampa” (Fig. 99) by Konings (2015a: 132).

Fig. 98. In the 1970s, Neolamprologus mustax was frequently confused with N. petricola (§ 26). The picture shows a Zambian specimen of N. mustax from the collections of the Swedish Museum of Natural History, NRM 61025.
Fig. 99 (left). Neolamprologus sp. “Leleupi Kapampa” from Lunangwa. Fig. 100 (right). N. sp. “Leleupi Kapampa” from Kiku. Both individuals (Figs. 99 and 100) photographed at our fish facilities in Dar es Salaam in 2001, prior to export. The species name derives from Konings (2015a: 132).
Fig. 101. Neolamprologus sp. “Leleupi Blue Chin” collected by our team near Lupota, DR Congo in the early 2000s.

27. Neolamprologus modestus-like species in southern Tanzania
Besides the gap between Cameron Bay and Mbete Bay, where N. cf. petricola “Zambia” is found, the geographical distribution for N. modestus includes the area between Cape Chipimbi, Zambia, and Kilewani, Tanzania, plus the area between Lusekese, Kala North and Mikongolo, Kala South, Tanzania (§ 30). North of Kilewani and Kasanga Bay at Muzi, the N. modestus-like population lacks the yellow pectoral fins, these fins, as well as the dorsal, caudal, anal, and pelvic fins, appear solid black and non-transparent. In the early 1990s, this variant was exported by our team as N. modestus “Black – Kasanga” (Johansson, 1994a; 1994c). Today, we tentatively refer to this variant and potential species as N. cf. modestus “Lwasase Point – Kilewani”, alternatively as N. sp. “Modestus Lwasase Point – Kilewani”, which appears distinguishable based on its solid black body and fins. As the name implies, its geographical distribution lies between Lwasase Point and Kilewani.
Between Mbofula Points (Fig. 103) and Kalambo River in Tanzania, a straight line of about 135 km, many groups of N. modestus-like populations occur, which mainly differ with respect to the colouration of the dorsal, caudal, anal, and pectoral fins. Owing to these differences, they are referred to as N. cf. modestus, the abbreviation ‘cf.’ (Latin ‘conferre’, English ‘compare’) which here implies a provisional classification pending further taxonomic research and signifies a somewhat divergent appearance compared to that of the typical N. modestus. In addition to a description of their colours, they are referred to by names conforming to their geographical distribution in Tanzania, listed in an order based on their occurrence from north to south:
1. N. cf. modestus “Mbofula Points – Mtosi Bay” (Figs. 102, 104–105); brown or beige body and fins, the upper half of the caudal fin usually has a thick greyish-blue margin and a thin bright blue submargin, occasionally the margin and submargin of beige individuals particularly are subtly yellowish.
2. N. cf. modestus “Kalandasi Point – Wampembe” (Fig. 106); brown or beige body and fins, the upper half of the caudal fin has a wide bright yellow margin and the rear and soft parts of the dorsal and anal fins have bright yellow dots, sometimes forming large blotches. In the 1990s, this variant was exported by our team (African Diving) as N. modestus “Fire Tip – Wampembe” (Johansson, 1994b).
3. N. modestus (Fig. 107); geographical distribution between Lusekese, Kala North, and Mikongolo Island, Kala South, and between Kilewani and Kalambo River; brown or beige body and fins, pectoral fins are bright yellow and the upper half of the caudal fin has occasionally one or two vague bluish submargins. In the 1990s and 2000s, we exported and referred to this variant, which is the original variant or phenotype of the species, as N. modestus “Yellow Fin – Kantalamba”. In addition, throughout the western and eastern Zambian coasts, populations of this species appear to have yellow pectorals, which are considered as a hallmark for N. modestus, at least in the strict sense.
4. N. cf. modestus “Nausingili Island + Singa Island” (Fig. 108); brown or beige body and fins, pectoral fins have a small, sometimes faint yellow dot, the upper half of the caudal fin has a thick bright yellow margin and thin submargin between which there may be a thin bluish stripe, and the dorsal fin has a thin bright yellow margin.
5. N. cf. modestus “Maleza Island” (Fig. 109); brown or beige body and fins, the upper half of the caudal fin has a thick bright yellow margin and thin submargin.
6. N. cf. modestus “Kasola Island” (Fig. 110); brown or beige body and fins, the pectoral fins occasionally have a small bright yellow dot.
7. N. cf. modestus “Lwasase Point – Kilewani” (Fig. 111); dark brown to beige body, usually with dark brown or black fins.
No detailed data are currently available regarding the N. modestus-like species in the areas between Mtosi Bay and Kalandasi Point, as well as between Wampembe (Fig. 141) and Lusekese, Kala North.
From the above list of N. modestus-like populations and colour characteristics of the different variants or potential species, it seems that certain colour patterns reappear geographically. For example, the original blackish, brown or beige N. modestus with yellow pectorals is found south of the bay at Kilewani, but also in the Kala Bay area, while those populations between Kilewani and Kala Bay exhibit other colour patterns and are referred to as separate variants and potentially distinct species. Obviously, a species may comprise several geographically isolated populations, see for example, Chalinochromis brichardi, which occurs in the north and south, but not in the central parts of the lake, where it is replaced by distinct congeners. Furthermore, the list above does not necessarily comprise seven distinct species, but rather seven distinct colour patterns, which, however, are undoubtedly the most common hallmarks for distinct species.

Fig. 102. Neolamprologus cf. modestus “Mbofula Points – Mtosi Bay” at Mbofula Points, 10 km south of Kipili. This is the most northern locality where we found this variant, or any variant of N. cf. modestus. Along a large portion of the coastline, between Kampemba and Mbofula Points, including the Kipili Archipelago, there are no N. modestus-like species.
Fig. 103. About 10 km south of Kipili, three strips of land project into the lake, forming the rocky habitat of Mbofula Points, an area where the northernmost representative of Neolamprologus cf. modestus is found (Fig. 102).
Fig. 104. Neolamprologus cf. modestus “Mbofula Points – Mtosi Bay” at Mtosi Point.
Fig. 105. Neolamprologus cf. modestus “Mbofula Points – Mtosi Bay” in Mtosi Bay. Some individuals may attain a brown colouration, while most individuals observed are light beige (see previous photo, Fig. 104). The dark individuals of this variant are distinguished from a similar variant at Muzi (N. cf. modestus “Lwasase Point – Kilewani”) by the frequent thick greyish-blue margin and thin bright blue submargin of the upper half of the caudal fin. Occasionally, the margin and submargin of beige individuals of the “Mbofula Points – Mtosi Bay” variant are subtly yellowish, while individuals of the “Lwasase Point – Kilewani” variant usually have thick black, non-transparent pectoral fins, especially the variant at Muzi.
Fig. 106. Neolamprologus cf. modestus “Kalandasi Point – Wampembe” at Mongwe, near Msamba. This may be the most colourful variant of the N. modestus-like species; parts of the pectoral, dorsal, and caudal fins are thick yellow. In the 1990s, we exported this colour variant from Wampembe with the epithet “Fire Tip”.
Fig. 107. In Tanzania, the true Neolamprologus modestus is found around Kala Bay and in the extreme south from Kilewani to Kalambo River. The photo was taken at Kambwimba, 7.5 km north of Kalambo River.
Fig. 108. The variant of Neolamprologus cf. modestus found at Nausingili and Singa Island is similar to the Kasola Island variant, but has a bright yellow trim in the dorsal and caudal fin compared to the usual blue in the latter. The picture shows a male with fry at Singa Island at a depth of five metres.
Fig. 109. Neolamprologus cf. modestus at Maleza Island lacks the yellow marks in the pectoral fin, observed in the colour variant at the adjacent islands, Nausingili and Singa.
Fig. 110. At Kasola Island, the Neolamprologus modestus-like species has a few yellow marks at the central part of the pectoral fin, similar to the variants found at Nausingili Island and Singa Island, but lacks the bright yellow trim in the dorsal and caudal fin found in the latter two.
Fig. 111. Neolamprologus cf. modestus “Lwasase Point – Kilewani” at Kitawe. The variant from Muzi, 8 km southeast of Kitawe, was exported by us in the 1990s as N. modestus “Black”. The thick black and non-transparent pectoral fins are typical for this variant.

28. Lamprologus staecki (Brichard, 1978a: 224)
Brichard (1978a: 224) published an illustration of N. modestus, or at least an N. modestus-like species, referring to it as ‘Lamprologus staecki’ with the full name in italics. Obviously, the name appears scientific and one may think that his intention was to scientifically introduce it. However, his extremely short description of “slate gray body and yellow pectorals are typical” (Brichard, 1978a: 224) clearly confuses it with N. modestus, on which he commented: “the southern race of Lamprologus modestus has yellow pectoral fins” (Brichard, 1978a: 208). Presumably, ‘L. staecki’ was not seriously intended as a scientific name, the name was simply carelessly italicised, hence, it should not be regarded as a ‘nomen nudum’ [a formally published name that fails to conform to the nomenclatural code (ICZN: 1999)], but rather a pseudo-scientific trade name. Subsequently, Brichard (1989: 352) referred to the cichlid as “staecki” (with quotation marks and non-italicised, but without a generic name), a less ambitious and more subdued reference to the fish. However, he disagreed with Poll (1978: 745) that it was synonymous with N. modestus, but without giving any further details. At the end of the 1970s, ‘L. staecki’ was occasionally imported to Sweden (e.g. Menander, 1979).

29. Neolamprologus modestus-like species in central Tanzania
Two more N. modestus-like variants or species occur in central Tanzania. The southernmost of these, N. sp. “Modestus Mpimbwe”, which occurs between Karema and Kampemba Point, has erroneously been referred to as N. mondabu (Lundblad & Karlsson, 1992: 271; 1995: 46; Karlsson, 1998: 40; 2002; 2004; Konings, 1998a: 162; Sturmbauer et al., 2010), see more above (§ 4). A provisional description of N. sp. “Modestus Mpimbwe” includes a brown, beige, or slightly bluish beige body, yellowish brown or yellowish beige dorsal and caudal fin, both fins with a very thin yellow margin and blue submargin, the latter which is medium thick on the dorsal fin but usually very thick and indistinct on the caudal fin. The pelvic fins and anal fin are brown or beige and besides the pectoral fins, which are hyaline greyish, all fins are pearly in appearance. Adult specimens have a frontal hump, indented lower head profile, moderately elongated snout, and a rather truncate or straight-cut caudal fin with slightly rounded corners (vs. normally very rounded caudal fin in N. modestus). In contrast to N. modestus and all its variants (§ 27), N. sp. “Modestus Mpimbwe” has not been observed beating its caudal fin on the sandy or sediment-rich substrate to expose prey. This behaviour combined with the bluish fins, truncate caudal fin, and adult frontal hump distinguishes it from N. modestus and partly also from the most similar and geographically adjacent congeners. Differences within N. sp. “Modestus Mpimbwe” include the submarginal blue caudal-fin stripe in the variants of Kabwe, Udachi, and between Kansombo and Karema, which is occasionally concentrated to the upper half of the fin. Furthermore, the variants along the shores of Cape Mpimbwe and northern Korongwe Bay are never or only rarely dark brown, but normally slightly bluish beige. Currently, no N. modestus-like species have been observed between Kampemba Point (Fig. 15) and Mbofula Points (Fig. 103), Tanzania, a distance of 40 km. N. sp. “Modestus Mpimbwe” was included in the diving log by Lundblad and Karlsson (1992: 271; 1995: 46), referred to as N. mondabu and described as large and blue with a frontal hump, and Karlsson (1998: 40; 2002; 2004) observed specimens to be searching for eggs from spawning Lamprichthys tanganicanus. Throughout its geographical distribution, N. sp. “Modestus Mpimbwe” is sympatric with N. sp. “Eseki”.
North of Karema, between Ikola and Katumba Point, a differently coloured group of populations is found. Referred to as N. sp. “Modestus Mahale”, these populations represent a potential species which primarily differ from N. sp. “Modestus Mpimbwe” by the colouration of the dorsal and caudal fin. In N. sp. “Modestus Mahale”, the dorsal fin has an orange margin, lilac submargin, and an orange horizontal stripe just below, while the caudal fin has a somewhat thick orange margin, lilac submargin, and an orange vertical stripe or indistinct field posterior to that submargin. Furthermore, the caudal fin of N. sp. “Modestus Mahale” is not divided into upper and lower parts, like most of the N. modestus variants in the south (§ 27), but the vertical caudal-fin margin stretches across the entire fin. Interestingly, a similar colouration on the fins may be observed in N. mondabu, at least regarding the populations between Moyobozi (Fig. 76) and Segunga. Within N. sp. “Modestus Mahale”, there is no distinguished colour variation, but the caudal fins of the populations at Sibwesa and northwards are slightly indented as opposed to truncate or straight-cut. In particular, the variant represented by the populations at Sibwesa and Lyamembe is very different from the real N. modestus from the south. At Halembe, north of the Mahale Mountains, an N. modestus-like population has been observed (§ 31) (Fighiera et al., 2011: 22; Jonas, 2018), which represents a variant whose colour characteristics conform relatively well to N. sp. “Modestus Mahale”.
In two inventory studies at Miyako in the Mahale Mountains National Park, Tanzania, Kuwamura (1986; 1987) reported the sighting of N. sp. “Modestus Mahale”, referred to as N. modestus. The species was recorded to be a polygynous harem breeder (Kuwamura, 1986: 138), and found to be very abundant in both the rocky and sandy habitat, in depths down to 35 metres (Kuwamura, 1987: 4). At a nearby location, referred to as Mahali, Rossiter (1995: 198) found N. modestus (N. sp. “Modestus Mahale”) to be common among boulders and sand. Like N. sp. “Modestus Mpimbwe”, N. sp. “Modestus Mahale” has not been observed uncovering prey with its caudal fin. Throughout the southern quarter of its geographical distribution, N. sp. “Modestus Mahale” is sympatric with N. sp. “Eseki”.
The characteristics of the illustrated fish in Konings (1988: 52, Fig. bottom), captioned as N. mondabu, conforms very well to the description of N. sp. “Modestus Mahale”. Presumably, the photo, which does not have a locality label, was taken somewhere between Ikola and Katumba Point.

Fig. 112. Neolamprologus sp. “Modestus Mahale” at Kekese, a small inlet between rocks about 20 km north of Ikola.
Fig. 113. Bulu Point (left) and Karilani Island are popular collection locations for ornamental fish collectors. Partly due to substandard collection equipment, one may frequently come across displaced species and variants in the waters of these locations, fish which have been collected elsewhere in the lake but escaped from cages or deliberately released here. For example, commonly found are Tropheus “Red Rainbow”, T. “Double Blot” (Kirschfleck), and T. “Kaiser” (Ikola), all originating far south of these locations. The growing problem with displaced cichlids is of increasing concern among environmentalists. A few times a year, the lake witnesses the capsizing of ornamental fish collectors’ boats carrying thousands of cichlids of different, remotely collected, locally endemic species, with the result of minor ecological disasters. Admixture of alien and indigenous fish species should not be taken lightly as invasive species may lead to the extinction of species and loss of biodiversity (Harrison & Stiassny, 1999; 2004; cf. Karlsson & Karlsson, 2014c).
Fig. 114. Neolamprologus sp. “Modestus Mahale” at Karilani Island.

30. Introduction to the Neolamprologus modestus species complex
While Maréchal and Poll (1991: 284) concluded that the geographical distribution of N. modestus is confined to the “south end of Lake Tanganyika” (§ 27), we regard the rest of the N. modestus-like species found throughout much of the lake as partly distinct species, all of which are closely related, allopatrically distributed, and members of what may be referred to as the N. modestus species complex. For more details of what a species complex is, see Karlsson and Karlsson (2018a: 26–27, 90–100, caption Fig. 306, and § 55–57, 60, 61); see also Ribbink et al. (1983: 157) and Spreinat (1995: 208). Alternatively, this species complex may be referred to as a species, and all its members regarded as geographical colour and shape variants as opposed to distinct species. Consequently, N. petricola, N. sp. “Modestus Mpimbwe”, N. sp. “Modestus Mahale”, etc., would be geographical variants and synonyms of N. modestus. However, the trend in systematics appears to be splitting as opposed to lumping. Sturmbauer et al. (2010) suggested N. tetracanthus, N. cunningtoni, N. christyi, N. petricola, N. modestus, and N. mondabu to be a good candidate group for a future separate genus (§ 3), within which, the group of N. modestus-like species may be a good candidate for a species complex.
Currently, the N. modestus species complex features the following taxa: N. modestus, N. cf. modestus “Mbofula Points – Mtosi Bay”, N. cf. modestus “Kalandasi Point – Wampembe”, N. cf. modestus “Nausingili Island + Singa Island”, N. cf. modestus “Maleza Island”, N. cf. modestus “Kasola Island”, N. cf. modestus “Lwasase Point – Kilewani”, N. sp. “Modestus Mpimbwe”, N. sp. “Modestus Mahale”, N. petricola, N. cf. petricola “Zambia”, and N. cf. petricola “Congo North”.
One should keep in mind that similar groups of populations need not be sympatric to be identified as representatives for separate species. Most botanical and zoological species, including cichlid fish species, are allopatrically distributed and, increasingly acknowledged, the occurrence of cryptic species and speciation are highly underestimated. “Allopatric diversification is certainly a major pathway of speciation in many of Lake Tanganyika’s cichlid lineages” (Koblmüller et al., 2008: 11). “Since many species are extremely young, locally distributed and morphologically similar, inferring species status in cichlid fishes is often problematic” (Koblmüller et al., 2008: 10). Previously, the morphological “yardstick” method was suggested (Mayr, 1964: 167; 1992: 229; 1996: 274; Mayr & Ashlock, 1991: 100–105; Stauffer & Kellogg, 1998: 27), which implies giving species status to geographically separated populations if they are as morphologically distinct as sympatric species. Of course, colour characteristics may be included in such a differentiation method. However, sympatric species that are nearly identical, morphologically, behaviourally, and based on colouration, such as N. furcifer and N. timidus at Cape Mpimbwe (cf. Kullander et al., 2014a: 309, Fig. 4a, 4b) render the “yardstick” method useless (Karlsson & Karlsson, 2017i). Currently, “there’s also no agreed-upon yardstick for how much morphologic or genetic difference separates species” (Gibson, 2011: 394). Identifying species appears to be a rather subjective and pragmatic business, each observer, a hobbyist or professional taxonomist, applying a “personal yardstick” based on prior individual experiences.
In general, many new animal species are discovered and described “all the time”. For example, on average, ten new shark species are discovered every year (Bernvi, 2018). Some of these are represented by newly discovered populations, while others are morphologically distinct known populations that were previously lumped together under a single species name. Recently, it was revealed that what was thought to be a single widespread cobra species, the forest cobra, comprises five separate species, two of which were described as new, with the other three previously regarded as subspecies (Wüster et al., 2018). Hence, what was recently referred to as the forest cobra species is now the forest cobra species complex. Seemingly, the trend is that subspecies and variants become species, species become complexes, and complexes become genera.

Fig. 115. Lake Tanganyika map showing the geographical distribution in Tanzania of the individual species and variants of the Neolamprologus mondabu species group and N. modestus species complex based on field observations. In Tanzania, N. mondabu was observed from Segunga Bay northwards, and N. sp. “Modestus Mahale” (or another N. modestus-like species, § 29, 31) between Ikola and Magambo, possibly occurring north to Segunga Bay, a potential geographical boundary for the species. See enlargement of the Kipili area, Fig. 24.
Fig. 116. Neolamprologus christyi at Mtosi Bay, the type locality of the species.

31. The Neolamprologus mondabu species group
What may be referred to as the N. mondabu species group constitutes a group of rather speculatively related species, including N. mondabu, N. christyi, N. sp. “Eseki”, and N. sp. “Mwila”. The former three may even constitute additional species, being cryptic and allopatric, especially N. mondabu, which appears to be distributed throughout much of the northern half of the lake, but without being taxonomically and properly investigated. For the definition of a species group and species complex, see Karlsson and Karlsson (2018a: 90, § 55). The members of the N. mondabu species group share some obvious characteristics, including the moderately elongated dark brown or greyish body with an indented or lunate caudal fin. They usually reside in a mixed rocky and sandy habitat, except N. sp. “Mwila” which has not been found among rocks, but only sand and empty shells. However, Trewavas and Poll (1952: 12–13) did not observe any particular similarities between N. christyi and N. mondabu, but rather compared N. christyi with N. furcifer, N. petricola, N. savoryi, N. pulcher, and N. brichardi. Indeed. N. christyi is distinguished from N. mondabu by some rather clear meristic data, including the number of scales in a longitudinal series (50–60 vs. 34–37) (Poll, 1978: 753, 756). Furthermore, N. christyi is at least 4 cm larger than N. mondabu. Besides N. sp. “Mwila”, the other three species do not coexist anywhere in the lake, but are allopatrically distributed, a common feature for closely related species. The latter is due to the fact that allopatric speciation is thought to be more common that sympatric speciation.
While N. christyi is named after Cuthbert Christy, first collector of the species in 1926–1927, the specific epithet of N. mondabu is its local name. In Tanzania, N. mondabu occurs between the Burundi border and Segunga. From Segunga south to Halembe we have not observed N. mondabu, but an N. modestus-like species (§ 29) has been reported at Halembe (Fighiera et al., 2011: 22; Jonas, 2018). Seemingly, along the Tanzanian coast, N. mondabu does not coexist with any N. modestus-like species as it does along the opposite coast in DR Congo (cf. Konings, 2015a: 240). Since the 1970s, field observations in Lake Tanganyika have revealed that interspecific relations, including mutualism and commensalism, are important elements in the fish communities of the lake (Kawanabe et al., 1993; 1997; 1999). Researchers have shown that species with apparently similar ecological requirements are well segregated in resource utilisation, and the coexistence of closely related species composes a beneficial relationship that presumably promotes their existence within the community (Nakai et al., 1994). Speculatively, coexisting N. mondabu-like and N. modestus-like populations may represent different species than N. mondabu-like and N. modestus-like populations that do not coexist.
N. sp. “Eseki” occurs between Isonga and the five northern islands of Kipili (Kamamba, Kasisi, Kerenge, Mwila, and Nkondwe Islands) (§ 5). Along its southernmost distributional range, it does not coexist with any N. modestus-like species as it does farther north. N. sp. “Mwila” has so far only been found at Mwila Island (§ 10–11), where there are no N. modestus-like species, only occasional specimens of N. sp. “Eseki”. N. christyi occurs between the three southern islands of Kipili (Lupita, Mvuna, and Ulwile Islands) and Chituta Bay, Zambia (pers. obs.; Kohda et al., 1996: 239). Seehausen (1990: 9–10; 1991: 12–14) reported the sightings of N. christyi and N. cylindricus at Kigoma, but these were undoubtedly displaced specimens, escaped or released from cages by local ornamental fish collectors. In its southernmost range, N. christyi is sympatric with N. modestus, or N. modestus-like species, but geographically isolated from such species in its northernmost range, including the southern islands of the Kipili Archipelago.
Max Poll and his team of native fishermen, which were attached to the great Belgian Hydrobiological Mission of 1946–1947, obtained the holotype of N. christyi and 14 of its paratypes from the southern rocky coast of Mtosi Bay. Specimens were caught on 2 April 1947 by angling with worm at a depth of 2–4 metres (Trewavas & Poll, 1952: 12; Walschaerts, 1987: 32). For photos taken by Poll on the same date depicting the rocky environment of Mtosi Bay and the fishing gear the team used, see Leloup (1949: Pl. 13, Fig. 5) and Poll (1952a: Pl. 35, 36, 38). As for the locations of the N. christyi specimens collected by Christy, who, incidentally, was killed by a buffalo in DR Congo in 1932 (Nature, 1932; Royal African Society, 1932; Worthington, 1933: 285), his notebooks were mislaid, therefore the locations are unknown (Worthington & Ricardo, 1937: 1062).
Until the beginning of the 1980s, all notes and articles about N. christyi (e.g. Neergaard, 1976; 1977; 1982; Brichard, 1978; 1989; Mayland, 1978) were more or less transcriptions of the original descriptions by Trewavas and Poll (1952) and Poll (1956) partly because the species was not exported for the ornamental fish trade until around 1980. While one of the very first exported N. mondabu was illustrated in Kahl (1971: 17, Fig. top), perhaps the very first exported N. christyi was illustrated in Bialkowski (1981: 18). Another early photo of a live N. christyi from Zambia was featured in Zadenius (1983: 28, Fig. bottom), but due to the allegedly deviant lips of the specimen in that photo, which was considered not to completely conform to the original sketch by Trewavas and Poll (1952), the author was not convinced of its identity and speculated it as being a specimen of N. mondabu with elongated caudal-fin corners. Others, like Axelrod and Burgess (1986: 317), illustrated specimens of N. cunningtoni and referred to them as N. christyi.
Throughout the 1990s and beginning of the 2000s, we (African Diving) collected and exported N. christyi mainly from the southern Kipili islands, Mtosi Bay, Kala Bay, the Maleza Island area, and Kasanga area. There is hardly any noticeable colour variation between the many populations; they are all greyish or blackish with yellowish and bluish fins.
Funded by the ‘Tanganyika Exploration Committee’ and referred to as the ‘Third Tanganyika Expedition’, Cunnington conducted his expedition between March 1904 and June 1905 (Cunnington, 1906c; Boulenger, 1906). Besides investigating the flora and fauna of Lake Tanganyika, he also explored Lakes Malawi and Victoria. On Lake Tanganyika, for about 8 months, he sailed with a dhow from south to north and made collections of fish, snakes, crabs, prawns, parasitic crustaceans, sponges, plants, algae, and much more. For the exact route he sailed on Lake Tanganyika, see detailed map (Fig. 119). For illustrations of a characteristic Lake Tanganyika dhow of the time, see Brown (1893: 256) and Swann (1910a: 83, Fig. top). At Kaboge, about 43 km south of Uvira, DR Congo, on 1 March 1905, the expedition obtained specimens of new copepod and ostracod species, including Cyclops cunningtoni (Sars, 1909: 37, 52, 54–55; 1910: 734). At the same time, the two types of N. mondabu appeared to have been collected. Less than three weeks later, on 18 March 1905, the expedition left Lake Tanganyika and headed for Bukoba at Lake Victoria. After Cunnington’s initial catch, additional collections of N. mondabu followed. In 1920, at the expense of the Unites States National Museum, Washington, D.C., Raven collected four specimens of N. mondabu at Kigoma (§ 23) (Myers, 1936: 1, 14). Later, in the 1930s, the entomologist Louise Burgeon collected reptiles, amphibians, and fish in East Africa, including Lake Tanganyika. At Nyanza-Lac, Burundi, nine specimens of N. mondabu were obtained (David, 1936: 157; Poll, 1946: 341). At Rumonge in 1935 and Nyanza-Lac in 1937, the anthropologist Arthur Lestrade collected more than 70 N. mondabu specimens, and at Rumonge in 1939, another two specimens of the same species were collected by Maurice Van Hemelrijck, colonial Minister for the Belgian Congo in 1958–1959 (Poll, 1946: 341). Much later, observations of N. mondabu were made near Cape Tembwe (Konings, 1996: 77), at or near Tembwe (Büscher, 1998a: 790; 2002a: 8, Fig. top), and Lunangwa (Konings, 2015a: 238, 240). Whether these three latter sightings really relate to N. mondabu or cryptic similar species remains to be seen. Speculatively, they may relate to species that are phylogenetically closer to N. sp. “Eseki”.
The type locality of N. mondabu is Kaboge, DR Congo (Boulenger, 1915: 471), which is close to Mboko Island (cf. Konings, 2015a: 238), and should not be confused with Cape Kabogo, Tanzania. At Kaboge, at least two battles between Belgian colonial forces and slave traders have taken place, one in April 1898 and another on 11–12 October 1899, the latter led by Captain Hecq (Fig. 127) on the Belgian side (Janssens & Cateaux, 1908: 145, 153; 1911: 432–436). An excerpt from Moore’s narrative (1901b: 145), ‘To the Mountains of the Moon’, reads: “[M]uch to our surprise we also found Captain Haec [Hecq], the Belgian commandant at Usambura [Bujumbura]; he had encountered the rebel soldiers west of the Rusisi river, and had totally defeated them, killing several hundreds and entirely dispersing the rest. He had then pushed over the mountains to Usambura and joined Captain Bethe [the official head of the German Tanganyika district], where we found him almost dancing with delight”.

Fig. 117. Located at the northern side of Utinta Bay, African Diving Fish Station was established in 1989 following the discontinuation of a small ornamental fish operation with a temporary and portable set-up at Kigoma in 1987 and 1988. The station has been the starting point for several fish expeditions on Lake Tanganyika in recent years.

Part 1 – Part 2Part 3