The taxonomic diversity of the cichlid fish fauna of ancient Lake Tanganyika, East Africa

Abstract

Ancient Lake Tanganyika in East Africa houses the world's ecologically and morphologically most diverse assemblage of cichlid fishes, and the third most species-rich after lakes Malawi and Victoria. Despite long-lasting scientific interest in the cichlid species flocks of the East African Great Lakes, for example in the context of adaptive radiation and explosive diversification, their taxonomy and systematics are only partially explored; and many cichlid species still await their formal description. Here, we provide a current inventory of the cichlid fish fauna of Lake Tanganyika, providing a complete list of all valid 208 Tanganyikan cichlid species, and discuss the taxonomic status of more than 50 undescribed taxa on the basis of the available literature as well as our own observations and collections around the lake. This leads us to conclude that there are at least 241 cichlid species present in Lake Tanganyika, all but two are endemic to the basin. We finally summarize some of the major taxonomic challenges regarding Lake Tanganyika's cichlid fauna. The taxonomic inventory of the cichlid fauna of Lake Tanganyika presented here will facilitate future research on the taxonomy and systematics and the ecology and evolution of the species flock, as well as its conservation.

Introduction

Ancient lakes, defined here as lakes that have continuously existed for much of the Quaternary period or longer, are well known as biodiversity hot-spots. These long persisting freshwater bodies are typically very deep and rather isolated and usually house extremely species-rich biological communities featuring exceptional levels of endemism (Brooks, 1950; Martens, 1997). The extraordinary species richness of these lakes is often the product of intralacustrine adaptive radiations, in the course of which a common ancestor diversifies rapidly into new, phenotypically distinct, species that occupy the available ecological niche space (Schluter, 2000; Salzburger et al., 2014). As a matter of fact, some of the most impressive cases of adaptive radiations are known from ancient lakes, as exemplified by the species flocks of cichlid fishes in the East African Great Lakes (Fryer and Iles, 1972; Seehausen, 2015; Salzburger, 2018) or the amphipods in Lake Baikal (Macdonald et al., 2005). Besides being hot-spots of organismal diversity, ancient lakes may also serve as species reservoirs over time (Salzburger et al., 2002; Schelly and Stiassny, 2004; Wilson et al., 2004).

Scientific interest in ancient lakes and their faunas is manifold (e.g. Albrecht and Wilke, 2008; Larson and Schaetzl, 2001; Salzburger et al., 2014; Timoshkin et al., 2016; von Rintelen et al., 2014); yet, the different ancient lakes have received different levels of scientific attention. While Lake Baikal and the Laurentian Great Lakes are considered the best studied lakes in the world, the East African Great Lakes are under-studied in various aspects, for example with respect to their faunas and especially when it comes to taxa other than the cichlids (Salzburger et al., 2014). But even for the cichlid species flocks of the East African Great Lakes, which have been in the focus of taxonomic and speciation research for more than a century, the basic taxonomic structure is often poorly investigated. In Lake Malawi, for example, less than half of the estimated number of 800–1000 species are nominally described (Snoeks, 2000, Snoeks, 2004). Likewise, in Lake Victoria, only about 25% of the estimated amount of endemic species are described (Snoeks, 2000).

The situation is somewhat different for Lake Tanganyika, for which a much more comprehensive taxonomic record for cichlids is available (Snoeks et al., 1994). This is – at least to some extent – because the Tanganyikan cichlid species show greater differences to each other facilitating their classification (Snoeks, 2000), which can in turn be attributed to the relatively greater age of the lake's species flock compared to those of lakes Victoria (ca. 100–150 ka; Verheyen et al., 2003) and Malawi (ca. 700–800 ka; Malinsky et al., 2018; Meyer et al., 2017) and because of the polyphyletic nature of the Tanganyikan cichlid assemblage (Salzburger et al., 2002, Salzburger et al., 2005). Besides, there have been distinct periods of increased collection and classification activities with respect to the Tanganyikan cichlid fauna (see below).

Lake Tanganyika is the oldest (~9–12 Ma) of the East African Great Lakes and represents – by means of water volume – the largest body of freshwater in Africa (32,600 km² with a maximum depth of 1470 m) (Cohen et al., 1993; Salzburger et al., 2014). Lake Tanganyika's markedly diverse ichthyofauna is composed of 22 different fish families (Koblmüller et al., 2006; Fermon et al., 2017), including what is arguably the phenotypically most diverse cichlid assemblage in the world (Fryer and Iles, 1972; Salzburger et al., 2014). Apart from the cichlids, Lake Tanganyika is unique among the East African Great Lakes in having the by far highest proportion of endemic and morphologically diverse genera in groups of organisms other than cichlids (Salzburger et al., 2014). It is the Tanganyikan cichlids, however, that rank among the most noted model systems in evolutionary and speciation research (e.g. Irisarri et al., 2018; Muschick et al., 2012; Salzburger, 2018; Theis et al., 2017; Winkelmann et al., 2014), behavioural biology (e.g. Jungwirth et al., 2015; Theis et al., 2012; Young et al., 2019), and the study of the molecular mechanisms of trait evolution (e.g. Böhne et al., 2016; Santos et al., 2014). Despite the general interest in Tanganyikan cichlids, most previous studies have either focused on one particular species, on a sub-group of species (e.g. a genus or a tribe), or on a subset of taxa occurring in a particular area of the lake as a representative for the Tanganyikan cichlid radiation. As a consequence, some species and/or geographic regions are thoroughly investigated, whereas others remain understudied. Overall, the scientific literature is vague when it comes to the actual number of cichlid species found in Lake Tanganyika, and even more so for other African Great Lakes. Well established online databases – such as FishBase (Froese and Pauly, 2019) or the Catalog of Fishes (Fricke et al., 2019) – are of moderate help in this context as these are restricted to contain information about described species and their level of completeness depends on their curation, whereas undescribed species and varieties of existing species have mainly been discussed in extensive monographs (Konings, 2015) and/or hobbyists' journals.

Here we provide a concise overview of the currently described, valid cichlid species of Lake Tanganyika and list so far undescribed species as well as local varieties, taking into consideration the available literature including all first descriptions of cichlid species from the lake, as well as personal observations during many years of field collections (1980–2018) covering the majority of the shoreline of Lake Tanganyika (see Fig. 1). Note that we only considered species which we observed, and/or which have been reported to occur in the lake itself (i.e. lacustrine species), whereas purely riverine species are not discussed.

Fig. 1.

Map of Lake Tanganyika with indicated localities visited for collection and diving activities. Orange circles represent sites visited before 1998 by only one author (HHB), red circles indicate locations sampled between 2007 and 2018 by all authors. Darker areas in the lake illustrate the three sub-basins of Lake Tanganyika. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

We do not aim to challenge or revise the taxonomic status of any of the described cichlid species from Lake Tanganyika. Instead, we (i) provide an up-to-date inventory of all Lake Tanganyika cichlid species considered valid in the light of the International Code of Zoological Nomenclature; (ii) report candidate taxa for future descriptions as new species based on personal observations and opinions; and (iii) identify the major areas of taxonomic uncertainty with regard to the cichlid species flock of Lake Tanganyika. The species inventory of Lake Tanganyika cichlids, compiled to the best of our knowledge, will serve as valuable resource for the scientific community interested in the Tanganyikan cichlid fish fauna.

Cichlid taxonomy

Before reviewing the current taxonomy of the cichlid fauna of Lake Tanganyika, we briefly discuss some of the general problems associated with the delineation of species in cichlids, in which, for various reasons, the classic species concepts are often not effective (reviewed in Salzburger, 2018). A first challenge emerges from the species-richness of the cichlid species flocks themselves, as it is often difficult for taxonomists to keep track of this diversity and to identify unambiguous characters on the basis of which species can be distinguished (Kornfield and Smith, 2000; Snoeks, 2000). The various cichlid species in the East African Great Lakes are very closely related, as a consequence of their origin via rapid adaptive radiation, which adds another level of complexity to taxonomic work (see e.g. Shaffer and Thomson, 2007; Van Steenberge et al., 2018). Furthermore, since the cichlid adaptive radiations are still ongoing, the boundaries between populations of the same species and two distinct species are often transitional (see e.g. Theis et al., 2014; Pauquet et al., 2018), making it difficult to draw a line between the alternatives that two sister-taxa are two species, rather than just one. Species delineation is further complicated by past introgressive hybridization events and ongoing gene flow between species, which appear to be rather common in cichlids (Anseeuw et al., 2012; Gante et al., 2016; Meier et al., 2017; Meyer et al., 2017; Irisarri et al., 2018). Finally, different approaches towards cichlid classification have been adopted over time, among taxonomists, and also among the radiations. What is considered a species thus differs among the cichlid species flocks of lakes Victoria, Malawi and Tanganyika. For example, whereas in lakes Victoria and Malawi, a difference in male nuptial colouration can be the sole diagnostic character distinguishing two species, different ‘colour-morphs’ are typically combined into the same species in Lake Tanganyika. This situation might partially reflect differences in the contribution of underlying evolutionary processes among the cichlid adaptive radiations in the East African Great Lakes (Van Steenberge et al., 2018). However, also within Lake Tanganyika, different criteria have been used in the delineation of cichlid species, and some valid species are separated by minor differences only. For example, Neolamprologus longior (Staeck, 1980) differs from its congener N. leleupi (Poll, 1956) by slight differences in body proportions and colouration only. Note that N. longior, among many other species, has initially been described as a sub-species. However, Poll (1986) refuted this concept for Lake Tanganyika cichlids and elevated all previously existing sub-species to the species level.

Species delineation in general, and in cichlids in particular is not an easy task and should incorporate the available suite of methods in an integrative framework (see Van Steenberge et al., 2015, Van Steenberge et al., 2018). Clearly, a uniform treatment in species delineation would be desirable; even if, at the end, each case has to be studied thoroughly and assessed individually. A re-evaluation of the Tanganyikan cichlid species and/or the revision of the taxonomic status of certain species is beyond the scope of this work. Instead, we aim to provide an overview of the current taxonomic status of the cichlid fish fauna of Lake Tanganyika. In the following, we subdivide the taxonomic diversity of Tanganyikan cichlids into the two categories ‘described’ and ‘undescribed’ species, whereby the former category includes what we classify as ‘questionable species’ and ‘museum species’. This subdivision is to account for the situation that some Tanganyikan cichlids have been studied in much more detail than others, with many of them still awaiting formal description, while again others have not been observed since their first description.

Described Tanganyikan cichlid species

It took a bit more than 30 years after Richard F. Burton (1821–1890) and John H. Speke (1827–1864) – in search of the source of the Nile – discovered Lake Tanganyika in 1858 (Burton, 1860) until the first lacustrine cichlids of Lake Tanganyika were described (Günther, 1894). Among them was Astatotilapia burtoni (Günther, 1894), a haplochromine species inhabiting the vegetated littoral zone of the lake as well as adjacent rivers and swamps. This widespread species has become one of the best studied cichlids and a common model species for behavioural, developmental and molecular studies (e.g. Böhne et al., 2016; Santos et al., 2014; Theis et al., 2012; Weitekamp and Hofmann, 2017).

After the first species descriptions by Albert K. L. G. Günther (1830–1914) in 1894, the number of formally described species increased rapidly around 1900 due to the comprehensive taxonomic work by George A. Boulenger (1858–1937) based on collections from expeditions to Lake Tanganyika conducted between 1894 and 1905 (see Fig. 2 and Table 1). A second major increase in species descriptions occurred between the 1940s and the 1980s through the extensive work of Max Poll (1908–1991) on the collections of the Belgian expedition to the lake between 1946 and 1947 (see Fig. 2 and Table 1). It was also Poll (Poll, 1986) who grouped the – at the time – 173 described Tanganyikan cichlid species into 12 tribes based on meristic and anatomical characters (note that in taxonomy a tribe is the rank between the genus and the family level). Subsequent taxonomic and molecular phylogenetic work erected additional tribes for some genera, while merging other tribes (Takahashi, 2003; Takahashi and Koblmüller, 2011; Dunz and Schliewen, 2013). According to our accounts, 208 cichlid species belonging to 57 genera and 16 tribes are described from Lake Tanganyika to date (including valid, lacustrine species only), while new taxa are added nearly every year (see Fig. 2 and Table 1). Our assignment of species into tribes largely follows the molecular phylogenetics-based studies by Muschick et al. (2012) and Dunz and Schliewen (2013).

Fig. 2.

Cumulative number of described cichlid species over time. The number of described species increased over the years steadily with a major increment around 1900, by the extensive work by George Boulenger, followed by a second steep ascent between the 1940's and the 1980's due to Max Poll's contributions. Note that only currently valid species are included, all later synonymized species are not considered in this study (see Table 1 for references).

Table 1.

List of described, valid, lacustrine cichlid species of Lake Tanganyika. For each of the 208 species the tribe assignment, the initial name upon description, and the reported type locality are provided. If no holotype has been assigned, the sampling localities of the syntypes are listed. Note that for the two species L. kungweensis and N. brichardi the species name has changed and therefore both names and references are listed. Only native and still valid species have been considered. Superscript notation:¹ Species we consider as ‘museum species’; ² Species we consider as ‘questionable species’; ³ Species with affinity to rivers, occurring in the lake and in the Lake Tanganyika basin; ⁴ Species not endemic to the Lake Tanganyika basin. LT = Lake Tanganyika; BUR = Burundi; DRC = Democratic Republic of Congo; TAN = Tanzania; ZAM = Zambia.

We would like to note that this compilation only contains those species, which are still valid; whereas, species that were synonymized subsequent to their description are not considered. Furthermore, we only report native species. Therefore, we did not include the Nile Tilapia, Oreochromis niloticus (Linnaeus, 1758). This species was introduced on several occasions in and around Lake Tanganyika but failed to successfully colonize the lacustrine zone of the lake and is mainly found in adjacent rivers. On the other hand, the subspecies O. niloticus eduardianus (Boulenger, 1912) was included in our list (see Table 1), as this taxon is considered native in the northern part of Lake Tanganyika.

‘Museum species’

Most of the 208 described cichlid species of Lake Tanganyika can more or less readily be encountered while Scuba diving or snorkelling, or bought on local fish markets. For example, in the last five years alone, we were able to collect specimens of 182 out of the 208 described Tanganyikan cichlid species during fieldwork campaigns in Burundi, the Democratic Republic of Congo (DRC), Tanzania and Zambia, and a similar number of species was photographically documented by a single biologist during ca. 750 h of underwater observations (Konings, 2015). On the other hand, there are five cichlid species that, following their initial description, have never been reported again from the wild (to the best of our knowledge). Here, we refer to these species as ‘museum species’, since they are only known from the type material in museum collections (see Table 1).

Three of these species, L. stappersi Pellegrin, 1927(a), Neolamprologus hecqui (Boulenger, 1899a), and N. wauthioni (Poll, 1949) have been collected from the western shoreline of Lake Tanganyika and only very little is known about the species' ecology, behaviour or distribution. The assessment of L. stappersi and N. hecqui is further complicated by the fact that for these species only the holotypes exist in museum collections. This makes it difficult to compare them to other taxa, as no within-species variance can be determined. On top of this, the only available specimen of N. hecqui was collected from the mouth of a catfish (Poll, 1956) and is, hence, not in particularly good shape. Subsequent specimens collected as N. hecqui were all re-assigned as L. meeli and L. boulengeri, respectively (Van Wijngaarden, 1995; Konings, 2015). For N. wauthioni, a paratype series comprising 13 specimens collected between 1946 and 1947 has been deposited. Still, this species has never been collected again (except for some incorrectly identified specimens later assigned to L. ocellatus (Steindachner, 1909b) by Büscher (2007)). At this time and without new collections, it is difficult to judge whether these three species have unusually small distribution ranges restricted to under-explored sections of the shoreline or may, given their similarity to species described later, be senior synonyms of other taxa.

In contrast, there are two species supposedly occurring in well accessible areas of Lake Tanganyika, which have not been reported again after their descriptions and which we consequently list as additional ‘museum species’. Pseudosimochromis margaretae (Axelrod and Harrison, 1978) was described on the basis of four specimens collected at a depth of three to six meters in the bay off Kigoma, Tanzania. While members of this genus are generally fairly easy to observe while snorkelling, we failed to collect or observe this species, despite intensive sampling, diving, and snorkelling activities at the reported type locality or elsewhere. The other species is Lamprologus finalimus Nichols and LaMonte, 1931 for which only the holotype exists. Intensive collection and research activities at and around the type locality in more recent years (see e.g. Van Steenberge et al., 2011; Mushagalusa et al., 2014; Fermon et al., 2017) did not reveal any further specimen of this species. In both cases the type material indicates a clear distinction from their congeners. This suggests that P. margaretae and N. finalimus are either extremely rare, have a very cryptic life style, or might have become extinct.

Additionally, we would like to mention here Xenotilapia burtoni (Poll, 1951a), although, according to our definition, this species does not entirely qualify as ‘museum species’. A substantial type series for this species was collected between 1946 and 1947 in the Burton Bay, DRC. However, to our knowledge this species was only reported again once after its initial description (Fermon, 2007).

‘Questionable species’

Three out of the 208 formally described cichlid species of Lake Tanganyika are categorized as ‘questionable species’ here: Tropheus kasabae (Nelissen, 1977), T. polli Axelrod, 1977, and N. cancellatus Aibara et al., 2005. The former two species were previously suggested, based on literature but not on morphological measurements, to be junior synonyms of T. moori Boulenger, 1898 and T. annectens Boulenger, 1900, respectively (Konings and Dieckhoff, 1992; Konings, 2013). We here agree that their species status is questionable, as in both cases the newly described species was never directly compared to the type material of T. moori and T. annectens, respectively, for which additionally the certainty of their type localities is under debate (see Konings, 2013; Konings and Dieckhoff, 1992 for details). At this stage, these two species should be considered valid until a solid revision of the genus Tropheus is available, which is currently in preparation (Van Steenberge, personal communication). The third species we consider a ‘questionable species’, N. cancellatus, is reported from a single location in Zambia only. It has previously been suggested based on morphological grounds that this species might represent a hybrid between members of the genus Telmatochromis and Lamprologus (sensu lato) (Konings, 2015). Recent genetic data (Ronco et al., unpublished) lend support to this hypothesis, so that we consider N. cancellatus an occasional, natural hybrid and thus list it as ‘questionable species’, needing further investigation.

Undescribed Tanganyikan cichlid species

In addition to the 208 formally described cichlid species (including ‘museum species’ and ‘questionable species’), a substantial number of so far undescribed species have been identified, partly in the scientific literature, yet to a much larger extent in hobbyists' journals and in the ornamental fish trade (note that cichlids are very popular among aquarists). Lacking any proper scientific description, these putative species (or local varieties) are usually referred to under cheironyms, such as trade names or the names of their location of origin. Quite a number of these undescribed species have been incorporated in scientific studies so that data on their morphology, ecology and/or behaviour as well as on their phylogenetic position and/or population structure exist (see e.g. Koblmüller et al., 2004, Koblmüller et al., 2007; Egger et al., 2007; Meyer et al., 2015). However, their taxonomic status remains undefined. In Table 2, we list 55 undescribed cichlid species or local varieties reported from Lake Tanganyika in the scientific and/or popular literature, all of which we were able to observe and collect in the field and were subject to subsequent examinations. We have classified these taxa into the two categories ‘local variety’ or ‘potential new species’, based on personal observations and opinion (see Table 2). We do not claim here that this has any nomenclatural implications. Instead, our main intention is to emphasize the urgency of taxonomic revisions of many genera of Lake Tanganyika cichlids to clarify the status of the taxa mentioned in Table 2.

Table 2.

List of undescribed species and local varieties. The categorization is based on our personal opinions and observations from fieldwork and collection activities. The notation of the cheironyms follows the conventions explained in Snoeks (2000). LT = Lake Tanganyika; aff. = species affinis, suggesting that the taxon is similar, but distinct from the mentioned nominal species; cf. = conferre, suggesting the taxon to be comparable with the mentioned nominal species (Table 2).

Taxonomic challenges in Lake Tanganyika cichlids

The taxonomy of cichlid fishes in general, and that of the cichlid species flocks in the East Africa Great Lakes in particular, is highly challenging (Snoeks et al., 1994; Snoeks, 2000), which is partly due to the sheer number of species present and their close relatedness. In Lake Tanganyika this is further complicated by cases of convergent evolution within the radiation (Muschick et al., 2012), which might have contributed to several generic misplacements. Additionally, many Tanganyikan cichlid species show complex distribution patterns, presumably shaped by the patchy distribution of habitats along the lake's shoreline in combination with major lake level fluctuations (among other reasons) (see e.g. Sturmbauer et al., 2001). During periods of the most extreme low water stands, the lake was subdivided into three sub-basins (Salzburger et al., 2014). This previous separation of the lake in sub-basins is reflected today by many sister-species pairs showing a north versus south distribution (probably reflecting allopatric diversification in the sub-basins) or an east versus west distribution (probably reflecting dispersal along the paleo shore lines). However, the current taxonomy of Lake Tanganyika cichlids does not treat such cases consistently. In some cases, vicariant species-pairs were nominally described as two species (e.g. N. leleupi (Poll, 1956) from the West and N. longior (Staeck, 1980) from the East); in other cases, these were initially described as two species (e.g. Telmatochromis dhonti (Boulenger, 1919) from the South and T. caninus Poll, 1942 from the North) but later synonymized (Poll, 1986); while again in other cases only one species had been described (e.g. N. gracilis (Brichard, 1989) from the West with reports from a local variant at the eastern shore, N. sp. “gracilis tanzania”, see Table 2). Especially the East-West species pairs need revision, aiming at a more uniform taxonomic treatment of such sister-species pairs. Lake-wide sampling and phylogeographic studies (Rüber et al., 1999; Pauquet et al., 2018; Koblmüller et al., 2019) could serve as useful tool for future taxonomic revisions dealing with such difficult cases. Further, such studies can also contribute to the detection of yet unknown species. For example, a lake-wide genetic study of the tribe Eretmodini uncovered a distinct lineage within the genus Eretmodus (Rüber et al., 1999) which was later described as Eretmodus marksmithi Burgess, 2012.

Cases calling for revisions

Among Lake Tanganyika cichlids, several species are known to have been misplaced at the genus level. For example, Poll (1981) grouped two species, Gnathochromis permaxillaris (David, 1936) (type species of the genus Gnathochromis) and G. pfefferi (Boulenger, 1898) into the new genus Gnathochromis, based on morphological characteristics. Molecular work, however, placed G. pfefferi robustly within the Tropheini and G. permaxillaris within the Limnochromini (Salzburger et al., 2002; Takahashi, 2003). Yet, their generic name remains so far unchanged. The same applies to the genus Ctenochromis Pfeffer, 1893: Molecular data showed that C. horei (Günther, 1894) belongs to the Tropheini, while C. benthicola (Matthes, 1962) groups within the Cyphotilapiini (Muschick et al., 2012). In this case, none of the Tanganyikan species is the type species of the genus Ctenochromis. In both cases the generic misplacement affects only one or two species, respectively. However, within the Lamprologini the current genus assignment seems to disagree with the phylogenetic knowledge of the tribe for numerous taxa (see e.g. Colombo et al., 2016; Schelly et al., 2006). Those cases exemplify the need for a large-scale taxonomic revision of Lake Tanganyika cichlids.

Conclusions

We present a systematic overview of the taxonomic diversity of the lacustrine cichlid species flock from ancient Lake Tanganyika, East Africa. In particular, we provide an inventory of the valid cichlid fish species from Lake Tanganyika and list putatively undescribed species as well as local varieties. Based on this compilation, we estimate that Lake Tanganyika's cichlid species flock comprises at least 241 species, of which 208 (~86%) are nominally described and all but two (99.2%) are endemic to the basin (see Table 1, Table 2). To emphasize the demand for taxonomic revision, we highlighted some taxa at the species, genus and tribe level, needing further investigation.

Although Lake Tanganyika seems to be one of the most thoroughly examined aquatic ecosystems in tropical Africa, basic systematic work is pressing. Solid taxonomic knowledge is not only the basis for scientific study but also for nature conservation. As many other biodiversity hotspots, the unique ecosystem of Lake Tanganyika faces numerous anthropogenic threats. For example, the lake has become the focus of attention for future oil drilling projects (see Verheyen, 2016). A comprehensive understanding of the biological diversity of Lake Tanganyika is the basic prerequisite for any conservation measure, for example the delineation of small-scale protected areas (Sturmbauer, 2008). Although the IUCN Red List (International Union for Conservation of Nature's Red List of Threatened Species) accepts varieties and sub-population with reservations, valid species or subspecies are easier to assess in the system (IUCN Standards and Petitions Subcommittee, 2006).

Communicated by Björn Stelbrink