Chapter 3 Fishes - IUCN
Chapter 3 Fishes - IUCN
Chapter 3 Fishes - IUCN
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>Chapter</strong> 3.<br />
The status and distribution<br />
of freshwater fi shes<br />
Snoeks, J.¹, Harrison, I.J.² and<br />
Stiassny, M.L.J.³<br />
1 Zoology Department, Royal Museum for Central Africa, Leuvensesteenweg 13, B-3080 Tervuren, Belgium<br />
and Laboratory of Animal Diversity and Systematics, Katholieke Universiteit Leuven, Charles Deberiotstraat<br />
32 B-3000 Leuven, Belgium<br />
2 Conservation International, 2011 Crystal Drive, Suite 500, Arlington, VA 22202, USA<br />
3 Department of Ichthyology, American Museum of Natural History, Central Park West at 79th Street, New York,<br />
NY 10024, USA
3.1 Overview of the ichthyofauna of Africa 44<br />
3.1.1 A short introduction to the ichthyofaunal<br />
regions 44<br />
3.2 Conservation status 49<br />
3.3 Patterns of overall species richness 49<br />
3.3.1 All fi sh species 49<br />
3.3.2 Threatened species 55<br />
3.3.3 Restricted Range species 60<br />
3.3.4 Data Defi cient species 63<br />
3.3.5 Extinct species 65<br />
3.4 Major threats to species 66<br />
3.5 Research actions required 70<br />
3.6 Conservation recommendations 71<br />
Species in the spotlight – The Congo blind barb:<br />
Mbanza Ngungu’s albino cave fi sh 74<br />
Species in the spotlight – Tilapia in eastern Africa<br />
– a friend and foe 76<br />
Species in the spotlight – Forest remnants in western<br />
Africa – vanishing islands of sylvan fi shes 79<br />
Species in the spotlight – A unique species fl ock<br />
in Lake Tana – the Labeobarbus complex 82<br />
Species in the spotlight – The Twee River redfi n –<br />
a Critically Endangered minnow from South Africa 85<br />
Species in the spotlight – Cauldrons for fi sh<br />
biodiversity: western Africa’s crater lakes 87
CHAPTER 3 | FISH<br />
44<br />
3.1 Overview of the ichthyofauna of<br />
Africa<br />
Africa harbours a well-diversifi ed fi sh fauna, resulting from a<br />
long history of complex climatic and geological events that<br />
resulted in geographic isolation followed by speciation for<br />
some populations, or extinction for others (Roberts 1975;<br />
Lévêque 1997). While the African ichthyofauna shows many<br />
unique features compared to other continents, it shares<br />
affi nities with both South America and Asia, as a result of its<br />
connection to these landmasses as part of Gondwana.<br />
Africa has several archaic and phylogenetically isolated taxa<br />
(e.g., the bichirs, Polypteridae; lungfi shes, Protopteridae)<br />
(Lundberg et al. 2000). Lungfi shes, which include four<br />
African species in the genus Protopterus, are members of<br />
the most ancient group of bony fi shes that still have living<br />
representatives. Fossil lungfi shes date back some 410 million<br />
years (Le Cointre and Le Guyader 2001), although members<br />
of the extant genera are younger, originating in the Eocene,<br />
35-54 million years ago. Africa also has representative<br />
species groups that have undergone extensive recent<br />
adaptive radiation, if not even ‘explosive speciation’. The<br />
best-known case is the 500 or more species of cichlids<br />
(Cichlidae) in Lake Victoria that went through a major<br />
lineage diversifi cation about 100,000 years ago (Verheyen<br />
et al. 2003). While it is clear that the desiccation of Lake<br />
Victoria about 14,700 years ago had a large infl uence on this<br />
fauna, the evolutionary impact of a possibly completely dry<br />
Lake Victoria has been heavily debated (Stager et al. 2004;<br />
Verheyen et al. 2004). Lundberg et al. (2000) also provide<br />
other examples, such as the species fl ock of Labeobarbus in<br />
Lake Tana (see also Species in the spotlight – A unique<br />
species fl ock in Lake Tana – the Labeobarbus complex)<br />
and the diversity of sympatric mormyrids found in small<br />
rivers of the rainforests of western and central Africa. The<br />
levels of endemism are high for many parts of Africa, with<br />
some notable examples in the East African Great Lakes, the<br />
crater lakes (see Species in the Spotlight– Cauldrons<br />
for fi sh biodiversity: western Africa’s crater lakes) and<br />
the rivers of central Africa.<br />
The scientifi c study of African fresh and brackish water fi shes<br />
is more than a century old. A fi rst major step in compiling<br />
the existing information was set by George Boulenger, a<br />
Belgian ichthyologist, then working in the Natural History<br />
Museum in London. His important four volumes Catalogue<br />
of the African fi shes (Boulenger 1909-1916) provided the<br />
most authoritative account of 1,425 species. Even now, this<br />
catalogue is a major source of information for poorly known<br />
areas and poorly defi ned taxa. In a more recent effort to<br />
inventory all the fi shes occurring in African continental<br />
waters, the Check-List of Freshwater <strong>Fishes</strong> of Africa<br />
[CLOFFA] (Daget et al. 1984-1991) listed 2,908 species<br />
(Boden et al. 2004). CLOFFA represented a considerable<br />
increase in knowledge, and this knowledge has expanded<br />
further over the last two decades, as a result of several<br />
Polypterus endlicherii congicus (LC), a subspecies of bichir<br />
widespread throughout the Congo catchment and Lake<br />
Tanganyika. Bichirs are within the Polypteridae family, an archaic<br />
and phylogenetically isolated group of fi shes. © ULI SCHLIEWEN<br />
regional accounts. Lévêque et al. (1990, 1992) and Paugy<br />
et al. (2003a,b) published accounts for western Africa;<br />
Skelton (1993) published for southern Africa; and Stiassny<br />
et al. (2007a,b) published for the Lower Guinean region (see<br />
below). Many of these and other studies, as well as data from<br />
museum collections from around the world, are synthesized<br />
in FishBase (Froese and Pauly, 2010), highlighting that the<br />
total number of scientifi cally known African freshwater<br />
fi shes has risen dramatically since CLOFFA was compiled.<br />
For example, currently in FishBase, the number of fi shes in<br />
the Ethiopian or Afrotropical Zoological realm (that is, Africa<br />
excluding northern Africa, but including Madagascar and<br />
the southern part of the Arabian Peninsula) is about 3,200,<br />
almost all of them endemic to the realm. This endemism is<br />
not restricted to the species level; the majority of genera are<br />
endemic, as are about half of the families. In addition, it is<br />
likely that several hundred species are still to be described<br />
from the African continent, especially from the Great Lakes<br />
region, and the Congolian and Angolan river systems.<br />
3.1.1 A short introduction to the ichthyofaunal regions<br />
Attempts to subdivide Africa into ichthyofaunal provinces<br />
dates back more than a century. However, the basis for<br />
a modern synthesis was formulated by Roberts (1975),<br />
who based his work on Boulenger (1905); Pellegrin (1911,<br />
1921, 1933); Nichols, (1928); Blanc (1954); Poll (1957,<br />
1974); and Matthes (1964). Roberts (1975) recognised ten<br />
ichthyofaunal provinces (Figure 3.1):<br />
● Maghreb<br />
● Abyssinian (Ethiopian) Highlands<br />
● Nilo-Sudan<br />
● Upper Guinea<br />
● Lower Guinea<br />
● Zaire (Congo) (including lakes Kivu and Tanganyika)<br />
● East Coast<br />
● Zambezi<br />
● Quanza<br />
● Southern (including Cape of Good Hope)<br />
The Maghreb province in the north of Africa is quite distinct<br />
from other regions of continental Africa. It is relatively<br />
poor in species numbers, with dominance of Cyprinidae<br />
(Doadrio 1994). In biogeographic terms, this part of the<br />
continent has a closer affi nity with the Mediterranean part<br />
of the Palaearctic realm than with the remaining part of<br />
Africa (Balian et al. 2008b; Lévêque et al. 2008).
Figure 3.1. The major ichthyofaunal provinces of continental Africa, modifi ed from Stiassny et al. (2007a). (1)<br />
Maghreb, (2) Nilo-Sudan, (3) Abyssinian Highlands, (4) Upper Guinea, (5) Lower Guinea, (6) Congo (Zaire), (7)<br />
Quanza, (8) Zambezi, (9) East Coast, (10) Southern (including Cape of Good Hope).<br />
The Nilo-Sudan province is the largest, spanning the total<br />
width of the continent from Senegal to Mozambique. It<br />
includes two major river systems, the Nile (but excluding<br />
lakes Victoria and Edward and their affl uents) and the Niger,<br />
many West African coastal basins, and the endorheic Lake<br />
Chad system. It excludes a region spanning the coastal<br />
basins of the so-called Guinean ridge, from Guinea to<br />
the western part of Ivory Coast. This is the Upper Guinea<br />
province (see below), for which the exact boundaries are<br />
not well defi ned (Lévêque 1997). In a recent review of the<br />
fi shes of western Africa (Paugy et al. 2003a,b), including<br />
the Upper Guinea and the Nilo-Sudan provinces, but<br />
excluding the Nile system, 584 fresh and brackish water<br />
species were listed. A review of the Nile Basin fauna is long<br />
overdue, but counts of the River Nile (excluding the region<br />
of Lakes Victoria and Edward) include 128 species (Witte<br />
et al. 2009). However, several additional species that are<br />
endemic to Lakes Tana (Labeobarbus; see Species<br />
in the spotlight– A unique species fl ock in Lake Tana<br />
– the Labeobarbus complex), Albert and Turkana<br />
(haplochromines) should be added to this fi gure. The<br />
existing inventory of the Nile system is far from complete,<br />
especially in the poorly documented Sudanese part. In<br />
addition, the endemic cichlid fauna of Lake Albert contains<br />
many species that are still to be described (Snoeks pers.<br />
obs.).<br />
The Cross River forms the border of the Nilo-Sudan and<br />
the Lower Guinea province. While the river is considered<br />
to be part of the Lower Guinea province, its fauna includes<br />
elements of both provinces (Stiassny and Hopkins 2007).<br />
The Lower Guinea province included 577 fresh and brackish<br />
water species at the time of a recent review of its ichthyofauna<br />
(Stiassny et al. 2007a,b), more than half of which are endemic.<br />
The region spans the area from the Cross River southwards<br />
to just north of the Congo Basin. The Shiloango (Chiloango),<br />
with its lower reaches in Cabinda (Angola), is regarded as<br />
the southernmost large basin of this area. However, some<br />
smaller coastal basins with probably a mixed fauna occur in<br />
the region between the Shiloango and the Congo.<br />
CHAPTER 3 | FISH 45
CHAPTER 3 | FISH<br />
46<br />
A freshly caught specimen of the catfi sh (Euchilichthys<br />
royauxi) (LC) that lives in large rapids in the Congo River<br />
basin. © JOHN FRIEL<br />
The Congo Basin has the largest catchment area of any<br />
basin in Africa and globally is second only to the Amazon<br />
(Revenga and Kura 2003). As currently estimated, its<br />
fi sh fauna includes around 1,250 species. In the east it<br />
includes part of the Rift Valley region, including lakes Kivu<br />
and Tanganyika and the Malagarasi system. However, in<br />
ichthyofaunal terms, Lake Kivu belongs to the East Coast<br />
province (Snoeks et al. 1997). Lake Tanganyika is a quite<br />
distinct and noteworthy section of the Congo Basin. It is the<br />
oldest large lake in Africa which is refl ected in its distinctive<br />
ichthyofauna. More than 95% of its approximately 200<br />
cichlids are endemic to the lake, as are more than 60% of<br />
its non-cichlid species (Snoeks 2000; De Vos et al. 2001).<br />
Its major affl uent, the Malagarasi contains an ichthyofauna<br />
of mixed origin (De Vos et al. 2001). Even when disregarding<br />
the Lake Tanganyika endemics, an estimated 75% of the<br />
Congo species are endemic.<br />
The Quanza province is relatively small, including most<br />
of the coastal basins in Angola, south of the Congo and<br />
north of the Cunene. An estimate of the number of species<br />
is diffi cult to give, as this region is one of the least wellknown<br />
ichthyofaunal provinces in Africa. Poll (1967) listed<br />
109 freshwater fi sh species from this region, including<br />
several endemics.<br />
The Cunene, the Okavango Basin, the large Zambezi<br />
system and the Limpopo are the major components of<br />
the Zambezi ichthyofaunal province. On the eastern part,<br />
the southern border is delimitated by the St. Lucia Basin<br />
(Skelton 1994). Geographically, Lake Malawi is also part of<br />
this system. However, the lake’s basin and the Upper Shire<br />
The confl uence of the Inkisi River and the lower section of the Congo River, D. R. Congo. Only the Amazon and perhaps the<br />
Mekong have greater fi sh species richness. © ROBERT SCHELLY
The Chobe River, a major tributary of the Zambezi. The<br />
Zambezi ichthyofaunal province is relatively well known and<br />
contains many typical sub-Saharan African fi sh families.<br />
© TRACY FARRELL<br />
harbour a unique fauna dominated by some 800 or more<br />
endemic cichlids (Snoeks, 2004), and probably many more<br />
undescribed species (see section 3.3.1). The remaining<br />
part of the Zambezi province is relatively well known,<br />
with species numbers in this region characteristically<br />
decreasing from north to south (Skelton 2001).<br />
While the Zambezi province includes many of the typically<br />
sub-Saharan African fi sh families, this is less evidently the<br />
case for the southern or Cape ichthyofaunal province, which<br />
is relatively depauperate in families but high in species<br />
endemism. The border between the Cape and Zambezi<br />
provinces is not well defi ned. The Cape province includes<br />
the Orange system and all basins south of it. However,<br />
the Orange also includes typical Zambezi elements. Fortytwo<br />
species occur in the well-known Cape province, 36 of<br />
which are endemic (Skelton 2001).<br />
The East Coast ichthyofaunal province is situated to<br />
the north of the Zambezi province and is relatively poor<br />
in species and dominated by savanna. Skelton (1994)<br />
recorded 125 species from the various river systems, with<br />
an endemicity of about 60%. These systems include all<br />
coastal basins north of the Zambezi to the Tana system<br />
in northern Kenya. In this region species richness and<br />
endemism may well be underestimated, as many of these<br />
rivers are underexplored. The other main component of<br />
the province includes a series of lakes of various origins,<br />
Rukwa, Kivu, Edward, George, Kyoga, Victoria and its<br />
numerous smaller satellite lakes (Snoeks et al. 1997). All<br />
these lakes, except Rukwa, have elements of a regional<br />
super species fl ock, comprising more than 700 endemic<br />
species of haplochromine cichlids, and many undescribed<br />
species.<br />
Roberts’ (1975) work on ichthyofaunal provinces was<br />
continued by Greenwood (1983) and complemented<br />
by others (e.g., Skelton 1994; Lévêque 1997; Snoeks<br />
et al. 1997). While these provinces mostly refl ect past<br />
and current drainage patterns and are defi ned mainly<br />
The tigerfi sh, Hydrocynus vittatus (LC), a species popular among the sport fi shing community. This iconic species is generally<br />
common and abundant with a wide distribution across Africa, but is locally depleted by heavy fi shing pressure.<br />
© MELANIE L.J. STIASSNY<br />
CHAPTER 3 | FISH 47
CHAPTER 3 | FISH<br />
48<br />
Figure 3.2. A map of the freshwater ecoregions of Africa, from Thieme et al. (2005)<br />
on a characteristic combination of endemic taxa, their<br />
boundaries are not always straightforward, and transition<br />
zones often exist. Thieme et al. (2005) developed a slightly<br />
different system for defi ning the biogeographic regions of<br />
Africa, based on ‘freshwater ecoregions’ (See <strong>Chapter</strong> 1,<br />
Figure 1.2).<br />
Thieme et al. (2005) recognised 79 freshwater ecoregions<br />
on continental Africa (exclusive of Madagascar and<br />
offshore islands). Abell et al. (2008) subsequently revised<br />
this, recognising only 78 ecoregions (incorporating the<br />
Thysville caves into the Lower Congo ecoregion and<br />
renaming some of the ecoregions) (Figure 3.2). Ecoregions<br />
were defi ned by a combination of physical and biological<br />
characteristics, including the hydrological features of<br />
the region, the communities of aquatic species present,<br />
and associated ecological and evolutionary processes.<br />
Consequently, boundaries of ecoregions are not always<br />
exactly matched to river catchments; in some cases they<br />
may include partial catchments, or may aggregate subbasins<br />
that are components of quite different catchments.<br />
While the ecoregion approach provides very useful<br />
biological, ecological, and biogeographic information<br />
about a region, conservation planning and management<br />
for freshwater ecosystems are usually implemented for<br />
complete catchments or sub-catchments, rather than<br />
partial sub-catchments. For these reasons, the method of<br />
describing species distributions by sub-catchments has<br />
been adopted by <strong>IUCN</strong> for the freshwater fi shes included<br />
in the assessments of the status of freshwater species in<br />
Africa (e.g., Darwall et al. 2005, 2009; Smith et al. 2009;<br />
García et al. 2010a; Brooks et al. 2011).
3.2 Conservation status<br />
Of the 2,836 African freshwater fi sh species assessed<br />
at the scale of mainland continental Africa (that is, not<br />
including the four species ranked as ‘Not Applicable’<br />
(NA)), over half (57.4% are classifi ed as ‘Least Concern’<br />
(see Table 3.1, Figure 3.3). This may be partly explained<br />
by the large areas of Africa that are sparsely populated<br />
and where there is relatively little agricultural, industrial, or<br />
urban development that currently present a severe threat to<br />
fi shes in other areas (see below). Such undeveloped areas<br />
include large parts of the Congo Basin, Lower Guinea,<br />
and regions of southern Africa (Stiassny et al., 2007a,b;<br />
Tweddle et al. 2009; Stiassny et al., 2011).<br />
Over 500 species (18% of the classifi ed species) are ‘Data<br />
Defi cient’ (DD), with insuffi cient information about their<br />
taxonomy, ecology or distribution to assess whether they<br />
are threatened or not. This underscores the conclusion<br />
that a considerable amount of additional surveying and<br />
monitoring of African freshwaters is required to provide<br />
a more accurate assessment of the conservation status<br />
of species in these ecosystems, particularly in parts of<br />
central Africa and the Rift Valley lakes of eastern Africa<br />
where numbers of Data Defi cient species are greatest<br />
(see section 3.3.4 below). Nevertheless, even before such<br />
surveying and monitoring is implemented, it is possible<br />
to say that all existing evidence indicates that many<br />
freshwater fi shes face signifi cant threats. Six hundred<br />
and nineteen species (with an additional 16 sub-species)<br />
are classifi ed as threatened – representing 21.8% of all<br />
assessed species, or 26.6% of all species if one discounts<br />
the Data Defi cient species. Most of the threatened species<br />
are in the lowest threatened category, classifi ed as<br />
‘Vulnerable’ (57.2% of all threatened species), with another<br />
23.9% of threatened species listed as ‘Endangered,’ and<br />
18.9% as ‘Critically Endangered’. These fi gures represent<br />
large numbers of species (Table 3.1), further highlighting<br />
the severity of threats to African freshwater fi shes. Three<br />
species are reported as ‘Extinct’, although this is probably<br />
an underestimate of the true numbers (see section 3.3.5).<br />
Table 3.1. The number of African freshwater fi sh<br />
species in each <strong>IUCN</strong> Red List Category.<br />
<strong>IUCN</strong> Red List Category<br />
Number<br />
of species<br />
Number of<br />
endemic<br />
species<br />
Extinct 3 3<br />
Critically Endangered 117 117<br />
Endangered 148 148<br />
Vulnerable 354 353<br />
Near Threatened 75 73<br />
Data Defi cient 510 509<br />
Least Concern 1629 1585<br />
Total species 2836 2788<br />
Figure 3.3. The proportion (%) of freshwater fi sh<br />
species in each regional <strong>IUCN</strong> Red List Category in<br />
mainland continental Africa.<br />
3.3 Patterns of overall species richness<br />
3.3.1 All fi sh species<br />
Lévêque (1997) undertook a review of the numbers of<br />
fresh and brackish water fi shes of Africa according to<br />
taxonomy (i.e., families, genera, species). Although the<br />
numbers will have changed since then, his study provides<br />
a useful overview of taxonomic diversity. He recorded 76<br />
families in Africa, with the freshwater fauna dominated by<br />
ostariophysans (1,159 species), many of which are typically<br />
riverine; however, several families are represented by only<br />
a few species. The Cyprinidae (475 species) form the<br />
greatest proportion of ostariophysans, and Characiformes<br />
are also well represented by the Alestidae (109 species) and<br />
Distichodontidae (90 species). The Siluriformes (catfi shes)<br />
include numerous species of Mochokidae (176 species),<br />
Claroteidae (98 species), and Clariidae (74 species).<br />
Among the non-ostariophysan groups, Cichlidae is<br />
the most species rich family, with at least 870 species<br />
according to Lévêque (1997); most of these are represented<br />
by species endemic to the lakes of eastern Africa. Other<br />
families with large numbers of species include the former<br />
Cyprinodontidae, killifi sh (at least 243 species are currently<br />
classifi ed in Nothobranchiidae and Poeciliidae), and the<br />
Mormyridae (elephantfi shes), with 198 species.<br />
The geographic distribution of species shows some<br />
distinctive areas of high richness as well as areas of very<br />
low richness, or even absence of species (Figure 3.4). Not<br />
surprisingly, in this study there are no fi shes recorded from<br />
some of the driest parts of Africa, for example, much of the<br />
Sahara, parts of Ethiopia and Somalia, the Kalahari Desert<br />
of Botswana, and large parts of Namibia. In contrast, areas<br />
CHAPTER 3 | FISH 49
CHAPTER 3 | FISH<br />
50<br />
Figure 3.4. The distribution of freshwater fi sh species across mainland continental Africa. Species richness =<br />
number of species per river/lake sub-catchment.<br />
of greatest species richness include some of the large lakes<br />
of the Rift Valley of eastern Africa (where cichlids dominate<br />
the fauna), and the main channel of the Congo River. Of<br />
the large lakes, Lake Malawi (in the Zambesi ichthyofaunal<br />
province) has the greatest number of species for which<br />
assessments have been completed for the <strong>IUCN</strong> Red List<br />
(358 species have been assessed). Lake Malawi and its<br />
infl uents, Lake Malombe and the Upper Shire River that<br />
connects the two lakes, comprise the Malawi ecoregion<br />
(Thieme et al. 2005). This region includes an estimated 800<br />
species of fi shes, most of which are endemic (see section<br />
3.3.3). However, many are not yet formally described so<br />
are not assessed for the <strong>IUCN</strong> Red List. The majority<br />
of species present are cichlids, mainly represented by<br />
mouth-brooding haplochromine species. An estimated 67<br />
of these species (including two introduced species and<br />
10 that are yet to be formally described) are not cichlids<br />
(Snoeks 2004). The lake has some economically important<br />
cyprinids, including a sardine-like pelagic Engraulicypris<br />
sardella (usipa) (LC), a salmon-like Opsaridium microlepis<br />
(mpasa), which is Endangered, and a trout-like O.<br />
microcephalum (sanjika) (VU) (Thieme et al. 2005). Between<br />
nine and 12 species of mostly deep-water large catfi shes<br />
of the genus Bathyclarias are endemic to the lake (Snoeks<br />
2004), appearing to have originated from a widespread<br />
generalist species, Clarias gariepinus (LCRG), that is also<br />
present in the lake (Agnese and Teugels 2001). Several<br />
of the sub-catchments around the lake are also rich in<br />
species, with a total of 54 non-cichlid species found in the<br />
affl uent rivers (Snoeks 2004).
Haplochromis desfontainii (EN) is from Algeria and Tunisia,<br />
where it is found in warm, freshwater springs. It is a member<br />
of the Cichlidae family, the most speciose in Africa.<br />
© ANTON LAMBOJ<br />
The Lake Tanganyika Basin (part of the Congo Basin<br />
ichthyofaunal province) harbours an estimated 470 species<br />
of fi shes, 287 of which had been formally described from<br />
the lake itself at the time these numbers were reported by<br />
Thieme et al. (2005). The lake has high levels of endemism,<br />
especially for cichlids (see section 3.3.3). Around 300 of<br />
the 470 recorded species (64%) are cichlids, including<br />
species-rich lineages of substrate-brooding as well as<br />
mouth-brooding cichlids (Coulter 1991; De Vos and<br />
Lake Malawi ecoregion includes an estimated 800 species<br />
of fi shes, most of which are endemic and from the family<br />
Cichlidae. © FRANK DOUWES<br />
Mormyrops anguilloides (LC), a species of elephant fi sh<br />
widespread in Sub-Saharan Africa. It is a member of the<br />
Mormyridae family, one of the most species-rich families in<br />
Africa. © JOHN FRIEL<br />
Snoeks 1994; Snoeks 2000; Thieme et al. 2005). Thieme<br />
et al. (2005) note that the lake also has species fl ocks of<br />
catfi shes (Claroteidae and Mochokidae), snooks (Latidae)<br />
and spiny eels (Mastacembelidae). The lake also supports<br />
a unique community of pelagic fi shes including endemic<br />
clupeids (Limnothrissa miodon (LC) and Stolothrissa<br />
tanganicae (LC)) that are prey to several other species<br />
and which support an off-shore fi shery in the lake (Thieme<br />
et al. 2005). Affl uent drainages at the northern tip of the<br />
lake support up to 43 species, while the lower section<br />
of the Malagarasi on the eastern shore of the lake holds<br />
the highest number of species (71 species have been<br />
assessed) for any of the affl uents.<br />
Estimates for the total number of species in Lake Victoria<br />
(part of the East Coast province) are variable, although<br />
Thieme et al. (2005) note there may be more than 600<br />
endemic species. Most of these are cichlids, and several<br />
are Critically Endangered or Possibly Extinct (see section<br />
3.3.2 and 3.3.5). Many of the endemic species of cichlids<br />
are thought to have gone extinct since the 1980s (Harrison<br />
and Stiassny 1999). The sub-catchments adjacent to Lake<br />
Victoria hold between 21 and 51 species that have been<br />
assessed for the Red List, with the greatest numbers found<br />
in the Nzoia drainage to the north-east of the lake.<br />
More than 858 fi sh species have been assessed for the<br />
Congo Basin (i.e., that part of the Congo ichthyological<br />
province exclusive of the Rift Valley lakes Tanganyika)<br />
(Stiassny et al. 2011). This number is certainly an<br />
underestimate, and many of these regions are poorly<br />
explored or not explored at all; the recorded number is<br />
increasing as more species are described each year<br />
(Stiassny et al. 2011). For example, recent surveys in the<br />
lower section of the Congo River downstream of Malebo<br />
CHAPTER 3 | FISH 51
CHAPTER 3 | FISH<br />
52<br />
The Malebo Pool, one of the most species-rich areas currently known throughout the Congo catchment. © ROBERT SCHELLY<br />
Pool have more than doubled the number of species<br />
documented there, including the identifi cation of more<br />
than 10 new species in the last fi ve years (Stiassny et al.<br />
2011).<br />
The main courses of the major rivers of the Congo Basin<br />
have particularly high numbers of species, with more<br />
than 150 species reported for most reaches of the Congo<br />
River, as well as sections of the Lualaba, Kasai, Ubangi/<br />
Uele, and Sangha rivers. The middle section of the<br />
Congo, between Boyoma Falls and Malebo Pool, has the<br />
greatest species numbers, with several sections having<br />
more than 250 assessed species, while the Malebo Pool<br />
region itself has 316 assessed fi sh species. However, the<br />
smaller tributaries distributed throughout the Congo Basin<br />
have fewer species; fewer than 30 in many cases, and<br />
some sub-catchments had no species recorded in these<br />
biodiversity assessments. This apparent distribution of<br />
species is likely an artifact of more intensive sampling that<br />
has occurred in the larger channels of the Congo Basin<br />
compared to the smaller, more inaccessible, tributaries.<br />
Recent surveys of small river basins in the lower Congo<br />
region such as the Inkisi, Nsele, and Mpozo have, for<br />
example, found them to harbour many more species than<br />
previously documented (Thieme et al. 2008; Wamuini et al.<br />
2008; Monsembula pers. comm.; Schliewen pers. comm.),<br />
and similar observations are being made throughout<br />
other parts of the basin as inventories are undertaken.<br />
An excellent case in point is the Léfi ni River, from which<br />
virtually no species were known until recently when, after<br />
a thorough exploration of its lower reaches, it was found to<br />
harbour 140 species (Ibala-Zamba 2010).<br />
The equatorial location, large size and the relative longevity<br />
and climatic stability of the forested, moist tropical regions<br />
of central Africa contribute to the high levels of species<br />
This species of freshwater pufferfi sh, Tetraodon miurus (LC),<br />
captured in the Odzala National Park, D. R. Congo, is quite<br />
widespread throughout the Lower Congo River basin.<br />
© JOHN FRIEL
Rapids on the Dja River fl owing to the Congo basin. The Dja<br />
River headwaters were once captured by rivers of Lower<br />
Guinea, possibly the reason why the two regions share some<br />
ichthyological fauna. © TIMO MORITZ<br />
richness in this area (Kamdem Toham et al. 2006; Thieme<br />
et al. 2008). In addition, the region has a complex mosaic<br />
of habitats, contributing to 19 freshwater ecoregions, often<br />
with distinct hydrographic barriers between the habitats<br />
(for example, waterfalls and rapids); all of these factors<br />
appear to promote high species diversifi cation (Thieme<br />
et al. 2005; Brummett et al. 2009; Markert et al. 2010;<br />
Stiassny et al. 2010).<br />
Although the species richness observed in the Congo<br />
Basin and the east African Rift Valley lakes exceeds that<br />
observed in any other part of Africa, there are several<br />
other regions that have a relatively high species richness<br />
distributed over large areas. This is particularly noticeable<br />
over almost all of the Lower Guinean ichthyological<br />
province and large parts of western Africa (covering the<br />
western part of the Nilo-Sudan, and the Upper Guinea<br />
ichthyological provinces).<br />
The Lower Guinea province is adjacent to the Congo<br />
Basin, and the two regions share some fauna, perhaps as<br />
a consequence of historic capture of the headwaters of<br />
Congo Basin rivers by Lower Guinean rivers (for example,<br />
capture of the Dja headwaters by the Nyong, Ntem, Ivindo)<br />
(Thys van den Audenaerde 1966; Stiassny et al. 2011).<br />
More than 550 species have been reported from Lower<br />
Guinea (Stiassny et al. 2007a,b). The most species-rich<br />
drainages are the lower part of the Sanaga (Cameroon),<br />
the Ogowe (Gabon), the upper Ngounie (Gabon/Republic<br />
of Congo), and the lower Kouilou systems (Republic of<br />
Congo), each having over 100 recorded (and assessed)<br />
species. Most sub-catchments of the Lower Guinean<br />
province have between 50 and 100 assessed species.<br />
These include several coastal rivers that are relatively short<br />
(i.e., 60km or less) and are disproportionately rich in fi sh<br />
species relative to their small size. Lower Guinea has been<br />
a focus for ichthyological survey over the last 20 years.<br />
This has promoted extensive taxonomic and revisionary<br />
work, including the description of many new species and<br />
Protopterus annectens (LC), the African lungfi sh. Lungfi sh<br />
are adapted to survive periods of drought by burrowing to<br />
the bottom of mud in drying pools and aestivating there for<br />
up to eight months. © TIMO MORITZ<br />
the production of a guide to the freshwater fi sh fauna of<br />
the region (Stiassny et al. 2007a,b).<br />
Similarly, much of western Africa has been well studied,<br />
especially over the latter part of the 20th century and early<br />
21st century, with the production of taxonomic revisions<br />
and faunal guides (e.g., Paugy et al. 2003a,b). It is probably<br />
one of the better-known large areas after southern Africa.<br />
Western Africa includes 17 freshwater ecoregions,<br />
distributed through the western parts of the Nilo-Sudan<br />
province and the Upper Guinea province. According to<br />
Paugy et al. (2003a), there are 584 species of fresh and<br />
brackish water fi shes distributed through western Africa;<br />
521 of these species (the freshwater component) were<br />
assessed for the Red List (Laleye and Entsua-Mensah<br />
2009).<br />
While several river basins in western Africa have relatively<br />
high species numbers, the region is less uniformly rich in<br />
species than the Lower Guinea province of central Africa.<br />
The Niger River, which fl ows more or less from west to<br />
east across a large part of western Africa, and is Africa’s<br />
third longest river, has a patchy species density pattern.<br />
Those parts of western Africa that are richest in species<br />
tend to be the coastal basins, moist forests and woodland<br />
savanna, whereas those that have fewer species are<br />
found in the Sahel, where conditions are drier and rivers<br />
are smaller (with the exception of the Niger) or may fl ow<br />
only seasonally. <strong>Fishes</strong> found in these drier regions often<br />
show adaptations to periods of drought. For example, the<br />
lungfi sh, Protopterus annectens (LC) is an air breather (and<br />
must take lungfuls of air occasionally in order to survive)<br />
and can burrow into mud at the bottom of drying pools<br />
and survive there, aestivating, usually for up to seven or<br />
eight months; this may be extended experimentally to<br />
up to four years in P. aethiopicus (Helfmann et al. 1997).<br />
There are also species of killifi sh (e.g., Pronothobranchius<br />
kiyawensis (NT)) that have drought resistant eggs (Laleye<br />
and Entsua-Mensah 2009).<br />
CHAPTER 3 | FISH 53
CHAPTER 3 | FISH<br />
54<br />
The turquoise killifi sh, Nothobranchius furzeri (LC), from the Bahini National Park in Mozambique. This beautiful killifi sh is<br />
typically found in seasonal pans, which they often share with lungfi shes. They lay drought-resistant eggs, and are apparently<br />
one of the shortest lived killifi shes (four to fi ve months). © SAIAB/ROGER BILLS<br />
The greatest numbers of species in western Africa are<br />
found in the Niger Delta ecoregion (152 species have<br />
been assessed). The delta and surrounding areas are<br />
also some of the most heavily impacted areas in Africa<br />
(see section 3.3.2). To the west, 107 species have been<br />
assessed in the Ogun River basin in the region of the<br />
Lagos Lagoon, including six threatened species (see<br />
section 3.3.2). Several catchments and sub-catchments<br />
of western Africa have more than 70 species, particularly<br />
those in the Upper Guinea province (covering parts of<br />
southern Guinea, Sierra Leone, and Liberia), the Upper<br />
Niger and Inner Niger Delta ecoregions (in Guinea and<br />
Mali), and some coastal catchments from Ivory Coast to<br />
south-western Nigeria. The Volta ecoregion (including<br />
Lake Volta) has between 160 and 185 species (Laleye and<br />
Entsua-Mensah 2009). According to recent biodiversity<br />
assessments, 105 fi sh species are present in the lake<br />
itself; however, this number may be misleading, since it<br />
probably includes species found in drainages close to the<br />
lake in riverine and marginal habitats. Construction of the<br />
Akosombo and Kpong dams has signifi cantly affected the<br />
ecology of the region (most noticeably by the formation<br />
of Lake Volta) and has contributed to the decline of some<br />
species (Thieme et al. 2005; and see section 3.3.5).<br />
The northern parts of the Nilo-Sudan province in western<br />
Africa, which include the middle reaches of the Niger<br />
River, are situated in the Sahel, where species richness is<br />
muted (see above). An exception to this is the endorheic<br />
Lake Chad (with 69 species) and the region covering the<br />
Yedseram and lower Chari river basins (with 72 species<br />
recorded), both of which are major affl uents of Lake<br />
Chad. Many of the species in the region are adapted to<br />
patterns of seasonal fl ooding in the lake and around the<br />
lake margins.<br />
Other parts of Africa have lower species numbers<br />
compared to western Africa, Lower Guinea, the Congo,<br />
and the Rift Valley lakes discussed above. While the Nile<br />
River is the longest in the world (Revenga and Kura 2003),<br />
fewer than 30 species are recorded and assessed for most<br />
of its length. Many affl uents of the main channel, in both<br />
the Upper and Lower Nile ecoregions, and throughout<br />
much of the Ethiopian Highlands province, have fewer<br />
than fi ve species recorded. This is probably a refl ection of<br />
the aridity within much of the Nile’s catchment area, where<br />
affl uents tend to be small, and many fl ow intermittently<br />
and are unable to support large numbers of species.<br />
Nevertheless, part of the reason for low recorded species<br />
numbers is the limited exploration of large parts of the<br />
basin within Sudan (covering some of the Lower Nile and<br />
all of the Upper Nile ecoregions), including the vast Sudd<br />
swamps in southern Sudan.<br />
There are some exceptions to the low species numbers<br />
recorded in the Nile Basin. The greatest numbers of<br />
species (where 40 to 50 species have been assessed) are<br />
found upstream from Khartoum in the Blue Nile system,<br />
and in the wetlands around Gambela National Park (in<br />
the westernmost part of Ethiopia) that drain to the White<br />
Nile. Forty seven species from Lake Nasser, formed by the<br />
Aswan Dam, have been recorded and assessed. Despite<br />
this elevated number of species, the overall impact of the<br />
dam has been detrimental to the freshwater fi sh fauna<br />
of the Nile system (just as with Lake Volta in western<br />
Africa; see above) (see section 3.3.2), with many species<br />
apparently extirpated from the former parts of their range<br />
below the dam.<br />
Most of the East Coast province just south of the Ethiopian<br />
Highlands, and the eastern part of the Nilo-Sudan province,<br />
have low species richness. Lake Victoria and its satellite<br />
lakes and affl uent rivers (discussed above) are the most<br />
noteworthy exceptions to this. Twenty nine species are<br />
recorded and assessed for the Tana River basin in Kenya,<br />
39 in the Ruvu and Rufi ji river basins in Tanzania, and 25
in the Ruvuma River on the border between Tanzania and<br />
Mozambique. Otherwise, most of the east coast basins<br />
have fewer than 20 species, providing a striking contrast<br />
with the species rich coastal basins of Lower Guinea on<br />
the other side of the continent, and probably refl ecting a<br />
combination of limited survey and the well-documented<br />
episodes of aridity experienced by eastern Africa.<br />
Southern Africa encompasses the Quanza, Zambezi, and<br />
Cape ichthyofaunal provinces; it includes 22 freshwater<br />
ecoregions and some very diverse habitat types. The<br />
general pattern is one of declining species richness<br />
towards the west and south. For example, most of the subcatchments<br />
of South Africa have fewer than 10 species.<br />
Highest species richness is found in parts of the Zambezi<br />
basin upstream from Lake Kariba (more than 50 species<br />
are recorded in many sub-catchments, and 80 species<br />
are present immediately upriver from the lake itself) and<br />
in the lowest parts of the basin; in parts of the Okavango<br />
basin; some higher parts of the Limpopo basin; the Buzi<br />
and Save basins; the Incomati-Pongola system; and some<br />
smaller coastal basins in Mozambique. However, some of<br />
these densities appear, in part, to refl ect the intensity of<br />
collection efforts (Tweddle et al. 2009).<br />
3.3.2 Threatened species<br />
The distribution of threatened freshwater fi sh species<br />
(Figure 3.5) is largely focused in a band that runs along the<br />
coast of western Africa and the Lower Guinea province<br />
from Senegal to D. R. Congo, throughout the Zambezi/<br />
Okavango basins in the northern part of southern Africa,<br />
through the river basins and lakes of the Rift Valley of<br />
eastern Africa, also including some of the coastal basins<br />
of eastern Africa, and some basins in the eastern and<br />
southern parts of South Africa. There are a few pockets of<br />
threatened species along the Uele River in central Africa, in<br />
the region of Lake Tana in north-eastern Africa, and in the<br />
Atlantic and Mediterranean Northwest Africa freshwater<br />
ecoregions in the Maghreb region of northern Africa<br />
The absence of threatened species throughout most of<br />
the arid areas of northern Africa, and parts of the Horn<br />
of Africa (comprising Somalia and eastern Ethiopia), as<br />
well some parts of southern Africa (especially Namibia<br />
and Botswana) is unsurprising; fi shes are totally absent<br />
from many of these regions (see above). Some other<br />
sub-catchments may have only a very small number of<br />
species (e.g., fewer than three species) but in these cases<br />
The Lugenda River is located in northern Mozambique, where it fl ows from Lake Amaramba and forms the largest tributary of<br />
the Ruvuma River. The Lugenda River Valley’s rich wildlife has led to development of the area as a destination for ecotourism.<br />
Within the river itself, species such as Barbus atkinsoni, Labeo cylindricus and Oreochromis placidus, together with around 40<br />
other species of fi sh, sustain an important local fi shery. © SAIAB/ROGER BILLS<br />
CHAPTER 3 | FISH 55
CHAPTER 3 | FISH<br />
56<br />
Figure 3.5. The distribution of threatened freshwater fi sh species across mainland continental Africa. Species<br />
richness = number of species per river/lake sub-catchment.<br />
Aphanius saourensis (CR), the Sahara<br />
aphanius, is a species of killifi sh<br />
(Cyprinodontidae family) endemic to<br />
Algeria. Many killifi sh species survive<br />
periods of drought by having droughtresistant<br />
eggs. © HEIKO KAERST
Figure 3.6. The proportion of freshwater fi sh species that are threatened within each sub-catchment across<br />
mainland continental Africa. Species richness = proportion of species per river/lake sub-catchment that are<br />
threatened.<br />
between 44% and 75% of the species are threatened; this<br />
is the case for some sub-catchments in the Maghreb and<br />
in the Etosha and the Karstveld Sink Holes ecoregions in<br />
Namibia (Figure 3.6).<br />
One Critically Endangered species, the Sahara aphanius<br />
(Aphanius saourensi), is present in the Sahara freshwater<br />
ecoregion. This species is endemic to the Oued Saoura<br />
Basin, but has disappeared from several parts of the<br />
basin and is now restricted to a single population found<br />
near Mazzer in the Sahara desert (García et al. 2010b).<br />
Excessive groundwater extraction for agriculture, pollution<br />
of remaining wetlands, and introduction and proliferation of<br />
mosquitofi sh (Gambusia holbrooki) are the main reasons for<br />
this decline. There are two species of fi shes from northern<br />
Africa that are considered Endangered. Haplochromis<br />
desfontainii (EN) and Pseudophoxinus punicus (EN), native<br />
to Tunisia and Algeria, are threatened by groundwater<br />
extraction, dams, water pollution and drought, which<br />
widely affect the area (García et al. 2010b).<br />
Most of the Nile Basin lacks globally threatened species,<br />
with the exception of the upper part of the basin (adjacent<br />
to Lake Victoria), and in the vicinity of Lake Tana. High<br />
proportions of threatened species (44% to 75% of the<br />
species assessed) are found in Lake Victoria and several<br />
of the adjoining sub-catchments. Between 26% and 43%<br />
of assessed species are threatened in the region of Lake<br />
CHAPTER 3 | FISH 57
CHAPTER 3 | FISH<br />
58<br />
A Lake Malawi cichlid from the group of rock-dwelling<br />
species collectively known as ‘mbuna’. These species are<br />
endemic to the lake, and many have extremely restricted<br />
ranges. © SARAH DEPPER<br />
Tana, with seven threatened species in Lake Tana itself<br />
(see Species in the spotlight– A unique species fl ock<br />
in Lake Tana – the Labeobarbus complex). Although<br />
globally threatened species are not recorded throughout<br />
the rest of the Nile Basin, it is important to note that several<br />
species in the basin have undergone serious declines<br />
in the northern African parts of their range. García et al.<br />
(2010b) note that at least 80% of the 24 northern African<br />
freshwater fi shes listed as Regionally Extinct (to northern<br />
Africa) were previously found in the Nile Basin in Egypt,<br />
and the construction of the Aswan Dam was a major cause<br />
of these extirpations.<br />
The low numbers of threatened species recorded for much<br />
of central Africa, including the species rich Congo Basin,<br />
may be partly attributed to the lack of human development<br />
in many parts of this region. However, several parts of<br />
central Africa are also insuffi ciently surveyed to accurately<br />
assess threats to species found there (see section 3.3.4<br />
on Data Defi cient species). Moreover, in some tributaries<br />
of the Sangha River, and tributaries of the Kwango River<br />
draining to the Kasai, the ratio of threatened species to the<br />
total species richness is still relatively high (26% or more).<br />
The highest number of threatened species occurs in<br />
Lake Malawi, where there are 105 recorded threatened<br />
species (28% of the total species assessed for this lake).<br />
A number of these assessments are, however, based on<br />
the ecological characteristics of many cichlid species,<br />
such as their highly restricted ranges and low numbers<br />
of offspring, which make them particularly vulnerable to<br />
extinction. Given the more recent requirement to also<br />
document evidence of current or imminent threats to a<br />
species for it to be assessed as threatened, it is possible<br />
some may be downgraded to a lower Red List category<br />
when next re-assessed. The rock-dwelling cichlids, often<br />
called mbuna, have particularly restricted distributions<br />
(intra-lacustrine endemism). These cichlids grow slowly<br />
and produce small numbers of offspring, are extremely<br />
vulnerable to habitat degradation and exploitation, and<br />
recover slowly from population declines (Ribbink 2001;<br />
Thieme et al. 2005). The potamodromous species that<br />
migrate from the lake into affl uent rivers to spawn are<br />
also threatened by fi sheries operations at the river mouths<br />
where they congregate during migration, and by degraded<br />
spawning habitats within the rivers (Tweddle 1996).<br />
Lake Victoria has the next highest number of threatened<br />
species (81 species; 44% of the assessed species in the<br />
lake), resulting from a combination of well-documented<br />
threats, including: the introduction of the piscivorous<br />
predator, Nile perch (Lates niloticus), and the water<br />
hyacinth (Eichornia crassipes) which has reduced light and<br />
oxygen levels in the lake’s waters; overfi shing and use of<br />
fi sh poisons; and habitat deterioration and eutrophication<br />
resulting from increasing lakeside agriculture, urbanisation,<br />
and deforestation (for further discussion and extensive<br />
references see Harrison and Stiassny 1999; Kaufman<br />
1992; Witte et al. 1992a,b; Kaufman and Ochumba<br />
1993; Seehausen and Witte 1995; Oijen and Witte 1996;<br />
Seehausen 1996; Seehausen et al. 1997a,b; Kaufman<br />
et al. 1997; Witte et al., 2007) (also see <strong>Chapter</strong> 1, Box<br />
1). Witte et al (1992b) initially estimated that as many as<br />
200 species of haplochromine cichlid in Lake Victoria had<br />
disappeared or were threatened with extinction within the<br />
lake. Harrison and Stiassny (1999) recognised that the lake<br />
was undergoing catastrophic ecological and limnological<br />
changes that represent a serious threat to the endemic<br />
cichlids; nevertheless, they believed it was premature to<br />
suggest that many of these species were actually extinct,<br />
because Witte et al (1992b) were using data limited to an<br />
11-year period for a small part of the lake (Mwanza Gulf),<br />
which could not be extrapolated to the whole lake and used<br />
as a measure of extinction. Harrison and Stiassny (1999:<br />
table 9) listed 48 species of cichlids from Lake Victoria<br />
that might be extinct but could not be confi rmed as such<br />
because of complications in their taxonomy (in most cases,<br />
the species had not been scientifi cally described). They<br />
listed another 54 species (Harrison and Stiassny 1999:<br />
table 10) that might be extinct but could not be confi rmed<br />
as such because of inadequate surveying and sampling,<br />
and 30 species (Harrison and Stiassny 1999: table 11),<br />
which could not be classifi ed as probably or possibly<br />
extinct, due to a lack of data. Harrison and Stiassny’s<br />
caution in classifying the Lake Victoria cichlids as extinct<br />
has been supported by evidence of a resurgence in several<br />
of the species, with greater resurgence of zooplanktivores<br />
compared to detritivores (Witte et al. 2007). Even with this<br />
resurgence, the threats to many of the species in Lake<br />
Victoria are still quite evident, and it is pragmatic to record
them in categories of high threat (52 species are Critically<br />
Endangered, and many of these are also noted in <strong>IUCN</strong>’s<br />
database as Possibly Extinct). Nevertheless, many fi sh<br />
species in the lake remain Data Defi cient (see section 3.3.4),<br />
and more extensive surveying and sampling are required<br />
throughout Lake Victoria to fully assess the conservation<br />
status of the cichlid species present. Because Lake Victoria<br />
is Africa’s largest lake by area, this represents a signifi cant<br />
challenge. Moreover, there would be the requirement for a<br />
larger number of highly trained taxonomists than currently<br />
exist in Africa, or elsewhere, to identify the 600 or more of<br />
species amongst the many thousands of specimens that<br />
would be collected. In light of this taxonomic impediment,<br />
an ecological classifi cation of Lake Victoria’s cichlids into<br />
trophic guilds may offer a pragmatic, short-term solution<br />
(Witte et al. 2007).<br />
There are 12 threatened species in Lake Tanganyika (5%<br />
of the total number of 245 species recorded in the Red List<br />
assessments). The overall number of threatened species<br />
is lower than in Lake Malawi partly because there are<br />
fewer species present in Lake Tanganyika; the proportion<br />
of threatened species (relative to the total number) is lower<br />
compared to Lakes Victoria and Malawi because the threats<br />
are generally more localized (Cohen et al. 1995) (especially<br />
compared to Lake Victoria); and the Tanganyikan species<br />
tend to have a wider distribution, extending into areas<br />
where there are fewer threats.<br />
Outside the area of the large lakes of the African Rift<br />
Valley, the regions with high numbers of threatened<br />
species occur in and around the rapids in the Lower<br />
Congo and some coastal basins in western Africa and<br />
Lower Guinea. The Lower Congo has up to 24 threatened<br />
species just upstream of Inga, and has one Critically<br />
Endangered species of cichlid, Teleogramma brichardi,<br />
apparently restricted to the Kinsuka rapids near Kinshasa.<br />
This species is increasingly threatened by the impacts<br />
of urbanization at Kinshasa and Brazzaville (Stiassny et<br />
al. 2011). However, further collections are necessary to<br />
establish the precise distribution of this species. Several<br />
other Endangered species are found, especially in the<br />
vicinity of Malebo Pool.<br />
At least 13 threatened species are found in the delta region<br />
of the Niger River, including two Critically Endangered<br />
species that are threatened by the impacts oil exploration<br />
in the delta, the distichodontid Neolebias powelli and the<br />
killifi sh Fundulopanchax powelli. Six threatened species<br />
are recorded nearby, in the species rich lower Ogun River<br />
at Lagos lagoon. These species are threatened mainly<br />
by deforestation (e.g., Brycinus brevis, assessed as<br />
Vulnerable), as well as agricultural and urban development<br />
(e.g., the mormyrid Marcusenius brucii, assessed as<br />
Vulnerable); however, the small red-eyed tetra, an alestid,<br />
Arnoldichthys spilopterus (assessed as Vulnerable) is<br />
threatened by an extensive harvesting for the aquarium<br />
Teleogramma brichardi (CR), from Kinsuka rapids near<br />
Kinshasa, D. R. Congo. © MELANIE L.J. STIASSNY<br />
fi sh trade. Just north of the Niger Delta, in a tributary of<br />
the Benue River on the Bauchi plateau, the cyprinid Garra<br />
trewavasae is Critically Endangered due to the impacts of<br />
tin mining.<br />
At least 10 threatened species are found in the coastal<br />
drainages of Sierra Leone and Liberia; these include some<br />
Critically Endangered species (e.g., Labeo currie, Barbus<br />
carcharinoides, Epiplatys ruhkopfi , Tilapia cessiana and T.<br />
coffea), and several Endangered species, especially in the<br />
vicinity of the St. Paul and Lofa rivers. These species are<br />
threatened by habitat degradation caused by deforestation<br />
and mining. In the Konkouré River in Guinea, the catfi sh<br />
Synodontis dekimpei is Critically Endangered for the same<br />
reasons. In the Fouta-Djalon ecoregion of Guinea, the<br />
killifi sh Scriptaphyosemion cauveti, a Critically Endangered<br />
species from a tributary to the Kolenté River, is threatened<br />
by expansion of the nearby city of Kindia.<br />
Twenty six threatened species are recorded from the<br />
Western Equatorial Crater Lakes freshwater ecoregion<br />
and the river drainages nearby, at the border of south-west<br />
Cameroon and Nigeria (see Species in the spotlight –<br />
Cauldrons for fi sh biodiversity: western Africa’s crater<br />
lakes). Many of these species are Endangered or Critically<br />
Endangered, and the majority are cichlids endemic to<br />
crater lakes, although there are also several killifi shes,<br />
Neolebias powelli (CR), a small riverine pelagic distichodontid<br />
characiform, is endemic to a very localised part of the Lower<br />
Niger Delta, where it is threatened by oil exploration within<br />
the delta. © TIMO MORITIZ<br />
CHAPTER 3 | FISH 59
CHAPTER 3 | FISH<br />
60<br />
Denticeps clupeoides (VU), from the Iguidi River, south-east<br />
Benin (see Species in the Spotlight: Forest remnants in<br />
western Africa – vanishing islands of sylvan fi shes).<br />
© TIMO MORITZ<br />
and some cyprinids and catfi shes. Within Lower Guinea,<br />
high numbers of threatened species (12 to 14 species) are<br />
found in sections of the Ivindo, Bouniandjé and Nouna (a<br />
tributary to the upper Ivindo) systems.<br />
In eastern and southern Africa, the number of threatened<br />
species is low for most basins (excluding the Rift Valley<br />
lakes). The greatest numbers in eastern Africa are found in<br />
the small Ruvu River, a coastal basin near Dar es Salaam<br />
that harbours nine threatened species. Although other<br />
basins in eastern Africa have fewer numbers of threatened<br />
species, some include one or more Critically Endangered<br />
species, which in most cases are are cichlids. For example,<br />
Oreochromis pangani, is a Critically Endangered species<br />
from the Pangani basin, which has been impacted by<br />
several different threats. Outbreaks of disease reduced the<br />
population in the late 1960s; subsequently, overfi shing and<br />
fi shing with illegal gear, as well as siltation and pollution,<br />
have continued to threaten populations. The clearance of<br />
macrophytes also removed important refuges and feeding<br />
areas for the fi sh. Orthochromis uvinzae is restricted to the<br />
middle Malagarasi River drainage to Lake Tanganyika, in<br />
Tanzania, and is impacted by habitat loss. Oreochromis<br />
chungruensis is endemic to Lake Chungruru, a crater lake<br />
north of Lake Malawi, and is impacted by siltation and<br />
dropping water level.<br />
In southern Africa, the Olifants River in the southwest<br />
part of South Africa has the greatest numbers of<br />
threatened species, with seven species (70% to 75% of<br />
all assessed species in the Olifants basin), including two<br />
Critically Endangered species, Barbus erubescens (see<br />
Species in the spotlight– The Twee River redfi n –<br />
a Critically Endangered minnow from South Africa)<br />
and an undescribed species of Pseudobarbus. Both<br />
are threatened by competition with, and predation by,<br />
introduced species, as well as deterioration in habitat and<br />
water abstraction caused by intensive farming. Slightly to<br />
the south, in the Tradou catchment of the Breede River<br />
system, Pseudobarbus burchelli is similarly Critically<br />
Endangered due to introduced species and pollution. A<br />
number of other species are also assessed as Endangered<br />
in the southern part of South Africa and in Lesotho (e.g.,<br />
Pseudobarbus asper and the Maloti minnow, Pseudobarbus<br />
quathlambae), where they face similar threats to the abovementioned<br />
species. Tweddle et al. (2009) gives a short<br />
account of the Maloti minnow conservation project. South<br />
Africa and Mozambique also harbour some undescribed<br />
but Critically Endangered species of Pseudobarbus,<br />
Kneria and Barbus. Two Critically Endangered species<br />
are found in the Karstveld Sinkholes ecoregion of central<br />
Namibia: one is Tilapia guinasana, which occurs naturally<br />
only in Lake Guinas. where it is threatened by groundwater<br />
extraction, as well as competition and predation from, as<br />
well as possible hybridization with, introduced tilapiines;<br />
and the other is the cave catfi sh, Clarias cavernicola,<br />
known only from a single tiny lake (18m by 2.5m) in the<br />
Aigamas Cave, near the town of Otavi, which is threatened<br />
by over abstraction of water and might also be impacted<br />
by collections made for the aquarist trade.<br />
There are three Critically Endangered species in the Zambezi<br />
River basin. An undescribed species of Barbus (Barbus sp. nov.<br />
Banhine) is known from four neighbouring sites at the southeastern<br />
edge of the Banhine National Park in Mozambique,<br />
in the Zambezian Lowveld freshwater ecoregion. In the upper<br />
Zambezi fl oodplain ecoregion, Neolebias lozii is restricted<br />
to the Sianda River that has been canalised, probably to<br />
aid drainage for agriculture. Unlike most other Critically<br />
Endangered species with restricted distributions, the cichlid<br />
Oreochromis mortimeri is widely distributed in the Middle<br />
Zambezi-Luangwa ecoregion and parts of the Zambezian<br />
Highveld ecoregion. This species is threatened mainly by the<br />
widespread introduction of O. niloticus, which is displacing it<br />
throughout much of its range.<br />
3.3.3 Restricted Range species<br />
Restricted range species (identifi ed as those species with<br />
distribution ranges of less than 50,000km²) are found in<br />
several African sub-catchments, but mainly in Upper and<br />
Lower Guinea, and some parts of the Congo Basin, and<br />
in eastern and southern Africa (particularly in the Cape<br />
Province of South Africa) (Figure 3.7). Most sub-catchments<br />
have only one to three restricted range species recorded<br />
(although this is likely to be an underestimate in some<br />
areas; see below). The largest of the Rift Valley lakes have<br />
many more restricted range species than are found in any<br />
other part of the continent. The high numbers of restricted<br />
range species found in the Rift Valley lakes are largely<br />
due to the endemic species-rich fl ocks of cichlids found<br />
in these lakes. For example, 99% of cichlid species in<br />
Lake Malawi are endemic, and more than 95% of the Lake<br />
Tanganyika cichlids are endemic (Snoeks 2000). Lake Kivu,<br />
to the north of Lake Tanganyika, harbours 15 restricted<br />
range species, while the Western Equatorial Crater Lakes<br />
ecoregion in Cameroon harbours 12. The restricted range
Figure 3.7. The distribution of restricted range freshwater fi sh species across mainland continental Africa.<br />
Species richness = number of species per river/lake sub-catchment.<br />
species in the crater lakes are also mainly cichlids. Many<br />
of those lakes with restricted range species, which also<br />
hold Critically Endangered and Endangered species, are<br />
potential candidates for designation as Key Biodiversity<br />
Areas, or possibly Alliance for Zero Extinction sites (see<br />
chapter 8, section 8.3).<br />
Lake Turkana, to the north-east of Lake Victoria, has nine<br />
restricted range species. Unlike the other large lakes of<br />
the Rift Valley, the fi sh fauna of Lake Turkana is composed<br />
mainly of nilotic riverine species rather than cichlids. Lake<br />
Tana in Ethiopia has eight restricted range species, mostly<br />
represented by the species fl ock of endemic cyprinids of<br />
the genus Labeobarbus (see Species in the spotlight –<br />
A unique species fl ock in Lake Tana – the Labeobarbus<br />
complex).<br />
Lake Tanganyika<br />
contains many<br />
restricted range<br />
cichlid species,<br />
95% of which are<br />
endemic to the<br />
lake. © JOHN FRIEL<br />
CHAPTER 3 | FISH 61
CHAPTER 3 | FISH<br />
62<br />
A selection of the rocky shore cichlid species endemic to Lake Tanganyika. © SASKIA MARIJNISSEN<br />
A number of rivers also have notably higher numbers of<br />
restricted range species. In the Upper Guinea province,<br />
the relatively small coastal basins of the Lofa, St. Paul, St.<br />
John, and Cess rivers have between fi ve and 12 restricted<br />
range species, the greatest number being in the Cess.<br />
Up to 17 restricted range species are found in the vicinity<br />
of Inga, in the lower part of the Lower Congo Rapids<br />
ecoregion. These species, some of which are assessed as<br />
Endangered, most likely show restricted ranges because<br />
they are adapted to the fast currents, low light intensity,<br />
The largemouth yellowfi sh, Labeobarbus kimberleyensis (NT),<br />
is endemic to the Orange River system in South Africa, where<br />
it is reasonably common, especially in large deeper pools in<br />
the middle and lower Vaal and Orange rivers, respectively. It<br />
is promoted as a fl agship angling species, with most anglers<br />
practising catch and release. © SAIAB/ROGER BILLS<br />
and high turbidity of the rapids. They are also likely to<br />
be impacted by development of the hydropower dam<br />
complex at Inga (see section 3.3.2). The current state of<br />
knowledge of the taxonomy and biogeography of species<br />
in the Lower Congo is incomplete, but recent research has<br />
identifi ed many more species than were previously known<br />
(Lowenstein et al. 2011; Stiassny et al. 2011). It is probable<br />
that more restricted range species will be found in the<br />
Lower Congo rapids region as surveys there continue.<br />
The Upper Congo rapids ecoregion also has marginally<br />
higher numbers of restricted range species (seven species)<br />
compared to surrounding areas, probably for the same<br />
reason as for the Lower Congo rapids (see above). Other<br />
parts of the Congo Basin with high numbers of restricted<br />
range species include the Ubangi River in the region of Bangui<br />
(six species), the Lubi basin draining to the Sankuru (seven<br />
species), the upper part of the Lowa Basin near Lake Kivu,<br />
and in several parts of the Lufi ra, Luvira, and Luapula basins<br />
in the Upper Lualaba and Bangweulu-Mweru ecoregions<br />
(all with fi ve restricted range species). In the Lower Guinea<br />
province, the greatest numbers of restricted range species<br />
(fi ve to eight species) are found in the Ivindo and upper parts<br />
of the Ogowe, and in the lower Sanaga River.<br />
The rivers of eastern and southern Africa have fewer<br />
restricted range species, possibly a consequence of the<br />
generally lower numbers of species found there. The<br />
sub-catchments with the greatest numbers of restricted<br />
range species are: the upper part of the Tana River, near
The Lower Congo rapids, home to many restricted specialist<br />
species. © ROBERT SCHELLY<br />
Nairobi (fi ve restricted range species); the Ruvu River near<br />
Dar es Salaam (eight restricted range species); and the<br />
lower part of the Malagarasi River near its outfl ow into<br />
Lake Tanganyika (seven restricted range species). The<br />
only catchment in southern Africa with moderately high<br />
numbers of restricted range species is the Olifant River in<br />
the Western Cape that holds fi ve restricted range species,<br />
representing around 75% of the total number of species<br />
recorded in the catchment.<br />
3.3.4 Data Defi cient species<br />
Species assessed as Data Defi cient (DD) (Figure 3.8) are<br />
those for which the taxonomy remains uncertain, or for<br />
which there is insuffi cient information to make a reliable<br />
Figure 3.8. The distribution of Data Defi cient freshwater fi sh species across mainland continental Africa. Species<br />
richness = number of species per river/lake sub-catchment.<br />
CHAPTER 3 | FISH 63
CHAPTER 3 | FISH<br />
64<br />
assessment of how they are impacted by threats to the<br />
freshwaters in their sub-catchment. The most common<br />
reason for this is the absence of reliable information about<br />
the total distribution of the species. Many species are known<br />
from only one or a few specimens from a single collection.<br />
Stiassny (1999) took a random selection of catfi shes and<br />
mormyrids included in the Check List of Freshwater <strong>Fishes</strong><br />
of Africa (CLOFFA, Daget et al. 1984, 1986, 1991), and<br />
found that, for the different genera and families selected,<br />
25% to 80% were known only from the type series or the<br />
type locality from where the species was fi rst described.<br />
Some species may be so poorly represented by collected<br />
specimens that their taxonomy is unresolved. There are<br />
many species of Barbus, for example, where taxonomic<br />
complexity precludes assigning the conservation status<br />
as anything but Data Defi cient. Barbus eutaenia is a<br />
good case in point. As currently defi ned, B. eutaenia is<br />
widespread across much of southern and central Africa,<br />
but it probably represents a complex of species that<br />
have yet to be diagnosed and described (Tweddle et al.<br />
2004). This is true for numerous other poorly diagnosed,<br />
putatively ‘widespread’ species that, on closer taxonomic<br />
scrutiny and with broad geographic sampling, will likely be<br />
revealed to represent species complexes.<br />
A persistent problem presented by older, historical<br />
collections is that they frequently contain ambiguous<br />
locality information, such that the presence or absence of<br />
a species in a particular river system cannot be determined<br />
with certainty. In such cases it may be impossible to reliably<br />
map the distribution of these species. Even allowing for<br />
this underrepresentation, the map showing distribution<br />
of Data Defi cient fi sh species clearly indicates that data<br />
defi ciency is a signifi cant problem for assessments of<br />
African freshwater fi shes.<br />
The greatest numbers of Data Defi cient species are found<br />
in the Rift Valley lakes of eastern Africa. Lake Victoria has<br />
the highest proportion of Data Defi cient species for any<br />
of these lakes, with around 35% of the assessed species<br />
being classifi ed as Data Defi cient. The high numbers of<br />
Data Defi cient species found in these lakes results from<br />
a combination of factors. The lakes are rich in overall<br />
species numbers (see section 3.3.1), extensive areas<br />
have not been well surveyed, and data on the complete<br />
distribution and ecology of many of these species is still<br />
lacking. Another important factor is that the taxonomy of<br />
the endemic lake cichlids is notoriously diffi cult, and many<br />
taxonomic problems abound.<br />
Five or more Data Defi cient species are found in several<br />
parts of the Congo Basin, with higher numbers particularly<br />
around Malebo Pool and sub-catchments immediately<br />
downstream in the Lower Congo (13 to19 Data Defi cient<br />
species), the Upper Congo rapids ecoregion (15 Data<br />
Defi cient species), and the Ubangi River adjacent to the<br />
town of Bangui (11 Data Defi cient species). The high<br />
numbers of Data Defi cient species recorded from the<br />
Congo Basin tend to be in areas that have been relatively<br />
well explored. As with the Rift Valley lakes of eastern<br />
Africa, relatively large numbers of species have been<br />
collected from these regions, but data on their full ranges<br />
and ecology may be lacking. This is also true, to a lesser<br />
extent, for the Nile system, where Lake Tana and some<br />
parts of the Blue and the White Nile system in Sudan<br />
contain more than fi ve Data Defi cient species.<br />
Data Defi cient species are found in sub-catchments<br />
through most of the Lower Guinea ichthyofaunal province<br />
of western central Africa, and in several sub-catchments<br />
of western Africa, but rarely in large numbers. There are a<br />
few exceptions, however. In the Northern Gulf of Guinea<br />
drainages freshwater ecoregion, at the border of the<br />
Nilo-Sudan and Lower Guinea ichthyolofaunal provinces,<br />
the headwaters of the Cross in Cameroon contain nine<br />
Data Defi cient species. This region is biogeographically<br />
interesting, since it includes a mix of species from the two<br />
ichthyofaunal provinces (Reid 1989), and there is evidently<br />
a need to learn more about the distribution and ecology of<br />
many of these species. For example, Teugels et al. (1992)<br />
found that the number of species in the Cross River had<br />
previously been underestimated by as much as 73%. This<br />
was a particular surprise because, prior to this, the Cross<br />
River was thought to be one of the better-surveyed rivers<br />
of western central Africa. Other parts of the Lower Guinea<br />
province that are noteworthy for Data Defi cient species are<br />
the middle reaches of the Nyong River (six Data Defi cient<br />
species) and the Kribi River (fi ve Data Defi cient species)<br />
in Cameroon, the lower reaches of the Ivindo River (fi ve<br />
Data Defi cient species) in Gabon, and a large part of the<br />
Niari-Kouilou system (fi ve Data Defi cient species) in the<br />
Republic of Congo (Brazzaville). In western Africa, the<br />
middle part of the Konkouré Basin in Guinea and most<br />
of the Rokel Basin in Sierra Leone each have fi ve Data<br />
Defi cient species.<br />
Data Defi cient species are also found in several subcatchments<br />
of the Quanza and Zambezian Headwaters<br />
freshwater ecoregions. But, as with western Africa and<br />
Lower Guinea, the numbers of Data Defi cient species<br />
are usually low – the exceptions being a couple of subcatchments<br />
with fi ve Data Defi cient species, and one<br />
tributary to the middle section of the Quanza that has<br />
10 Data Defi cient species. The entire Quanza ecoregion<br />
(which roughly corresponds with the Quanza ichthyofaunal<br />
province described by Roberts (1975)) is one of the least<br />
explored and poorly known areas in Africa.<br />
In the East Coast province there is a small concentration<br />
of Data Defi cient species in the headwaters of the Pangani<br />
Basin at the border between Tanzania and Kenya (seven<br />
species), and in the Galana basin (in Kenya). However, the<br />
Tana system in Kenya covers a larger area and has more<br />
Data Defi cient species (fi ve to nine, with up to 32% of the
Rivers in northern Africa typically have a hydrological regime that is unpredictable and which may experience periods of intense<br />
fl ooding. Such conditions present a considerable challenge to the freshwater species that live in these habitats. The region is<br />
home to an endemic group of ’Maghreb barbs‘, which are increasingly threatened by loss of habitat. © JEAN-PIERRE BOUDOT<br />
assessed species being Data Defi cient in some parts of<br />
the basin). This illustrates that, although most rivers of the<br />
East Coast province are characterized by relatively low<br />
species numbers, considerably more research is required<br />
for several of these species.<br />
The catchments in the Atlantic Northwest and<br />
Mediterranean Northwest freshwater ecoregions have<br />
consistently low numbers of Data Defi cient species;<br />
however, in all these catchments, the Data Defi cient species<br />
represent 25% or more of the species present (and in some<br />
cases up to 100%). Similar patterns are seen in the Horn<br />
of Africa, and in the south-western Cape of Africa, where<br />
there are numerous sub-catchments with only one or two<br />
Data Defi cient species, but these represent the majority of<br />
the assessed species in those catchments. There are also<br />
areas where there is a complete lack of data for any fi sh<br />
species. A lack of species is unsurprising in the Sahara,<br />
and several other arid areas such as the Kalahari and<br />
Namib ecoregions in southern Africa, and the Shebelle-<br />
Juba ecoregion in eastern Africa (as discussed in section<br />
3.3.1). However, there are also sub-catchments in parts<br />
of the Congo main basin that have no assessed species<br />
(Stiassny et al. 2011), and some of these catchments are<br />
in undisturbed parts of the Congo forest where one would<br />
reasonably expect to fi nd large numbers of fi sh species.<br />
A similar situation exists for the Sudd wetland in southern<br />
Sudan. Lack of data in these cases is undoubtedly due a<br />
lack of surveys in these regions, indicating an urgent need<br />
to undertake targeted surveys before rare or undescribed<br />
species are extirpated (see <strong>Chapter</strong> 8).<br />
Some species have not been formally, scientifi cally<br />
described, but are nonetheless recognized by taxonomists<br />
as being valid. The reason they are not formally described<br />
is usually because taxonomists have not had the time or<br />
resources to publish the scientifi c descriptions. This is<br />
evidence of an urgent need for more support and capacity<br />
building for freshwater taxonomists in Africa (Stiassny<br />
2002; Lowenstein et al. 2011). Without a full scientifi c<br />
description and account of the distribution of the species<br />
it is not possible to ascertain their conservation status,<br />
and they must remain categorized as Data Defi cient.<br />
Numerous examples exist in the literature, and some of<br />
these are included in the analyses presented here but,<br />
according to the Red List guidelines, they are not usually<br />
added to the <strong>IUCN</strong> Red List unless they are thought likely<br />
to be threatened.<br />
3.3.5 Extinct species<br />
Three species are classifi ed as Extinct (Aplocheilichthys<br />
sp. nov. ‘Naivasha’; Barbus microbarbis; Salmo pallaryi);<br />
all three had restricted distributions and their apparent<br />
extinctions are attributed, at least in part, to introductions<br />
of alien species. Aplocheilichthys sp. nov. ‘Naivasha’ is a<br />
poeciliid of indeterminate taxonomy that has been reported<br />
as A. antinorii but, according to Seegers et al. (2003),<br />
is quite distinct. It has probably been extinct since the<br />
1970s or 1980s, following competition or predation from<br />
introduced species. The cyprinid Barbus microbarbis was<br />
known from its type locality, Lake Luhondo (=Ruhondo)<br />
in Rwanda, though it possibly also inhabited the small<br />
CHAPTER 3 | FISH 65
CHAPTER 3 | FISH<br />
66<br />
streams fl owing into the lake. Only one specimen was ever<br />
caught in 1934, despite intensive sampling in the region<br />
(De Vos et al. 1990; Harrison and Stiassny 1999). If it is a<br />
valid species it is almost surely extinct, possibly as a result<br />
of the introduction of species of Tilapia and Haplochromis<br />
(De Vos et al. 1990). However, the sole type specimen was<br />
suspected to be a hybrid between a Varicorhinus and a<br />
Barbus species (Banister 1973). This hypothesis could<br />
be correct because, in other regions, putative hybrids<br />
between both genera have been identifi ed (Wamuini 2010).<br />
The salmonid, Salmo pallaryi, was restricted to Aguelman<br />
de Sidi Ali, a high altitude lake in the Atlas Mountains of<br />
northern Morocco, and known from at least 19 specimens<br />
(Delling and Doadrio 2005). The species apparently went<br />
extinct around 1938, probably due to the introduction<br />
of common carp in 1934, although Delling and Doadrio<br />
(2005) note that an unnamed population of trout (‘truite<br />
verte’) from Lake Isli, to the south-west of Aguelman de<br />
Sidi Ali, might be conspecifi c with S. pallaryi.<br />
Harrison and Stiassny (1999) and Helfman (2007) have<br />
discussed the possible evidence of extinction for several<br />
other species that are not classifi ed as such according to<br />
the recent <strong>IUCN</strong> Red List assessments.<br />
The schilbeid catfi sh Irvineia voltae is known only from<br />
the lower Volta Basin and is currently categorized as<br />
Endangered in the Red List. However, there have not<br />
been confi rmed reports of this species since the original<br />
collection before 1943. Harrison and Stiassny (1999)<br />
note that intensive but unsuccessful attempts were<br />
made to collect this species between 1961 and 1988,<br />
and it is possible that the species became extinct due<br />
to modifi cation of the river fl ow after construction of the<br />
Akosombo dams on the Volta River in the mid 1960s. At<br />
least prior to 1995, local fi shermen knew the species and<br />
had a name for it, but could not confi rm that they had seen<br />
it (DeVos 1995). More concerted surveying is necessary<br />
to resolve whether the species has disappeared. Besides<br />
the impacts from the dams, habitat quality is declining due<br />
to water pollution from agriculture, and possibly also from<br />
inadequately treated human waste. The species may also<br />
be affected by aquatic weeds.<br />
Harrison and Stiassny (1999) thought the cichlid Stomatepia<br />
mongo, endemic to Lake Barombi Mbo, Cameroon, might<br />
be extinct, because fi shermen had noted its absence<br />
(Reid 1991). However, the species is still extant because it<br />
was exported in 2007 to the USA for aquarium purposes<br />
(GCCA Forum 2007). Nevertheless, the fi shes of the<br />
crater lake, Barombi Mbo, face several serious threats.<br />
These include sedimentation and pollution from slashand-burn<br />
agriculture and oil plantations, deforestation,<br />
water abstraction for the neighbouring town of Kumba,<br />
commercial development of the region for tourism, and<br />
occasional fi sh kills caused by sudden releases of carbon<br />
dioxide contained under pressure in deep waters and<br />
sediments (similar to the event that occurred at Lake Nyos<br />
in 1986, killing more than 5,000 people and livestock)<br />
(Stiassny et al. 2011). Stomatepia mongo, endemic to this<br />
lake, is currently considered to be Critically Endangered.<br />
It is often noted that the largest extinction event in recent<br />
historical times (since 1500 AD) may have been in Lake<br />
Victoria, with the decline of endemic cichlid fi shes in the<br />
lake since the 1980s. There are indeed many species of<br />
cichlids endemic to Lake Victoria that are possibly extinct,<br />
but many of these cannot be defi nitively categorized as<br />
Extinct because there has been insuffi cient sampling<br />
or their taxonomy is not suffi ciently well described. As<br />
noted in section 3.3.2, many of the species in the lake are<br />
categorized as Data Defi cient, and there is an urgent need<br />
for more research and surveying of Lake Victoria fi sh fauna,<br />
in order to fully understand the geographic and taxonomic<br />
scope of this undisputed decline in cichlid species.<br />
Given the abundant threats to species in many parts of<br />
Africa (as indicated by the large numbers of threatened<br />
species recorded from these regions: see section 3.3.2) and<br />
high numbers of Data Defi cient species found throughout<br />
the continent (see section 3.3.4), it is reasonable to<br />
expect that further research and sampling throughout the<br />
continent might reveal evidence of extinction for several of<br />
these DD species.<br />
3.4 Major threats to species<br />
Deforestation, habitat loss and sedimentation<br />
Habitat modifi cation, through deforestation and associated<br />
increased sedimentation, is one of the most widespread<br />
threats to freshwater fi shes in Africa. The effect of this,<br />
even on a micro scale, has been demonstrated recently in<br />
the Léfi ni River, where within a relatively small stretch of<br />
shoreline, species composition was found to differ clearly<br />
between tree-covered and open areas (Ibala-Zamba<br />
2010). Loss of forest cover deprives many species of<br />
fi shes of shelter from predators, and changes the water<br />
temperature and hydrological regime of rivers (Brummett et<br />
al. 2009). Excessive sunlight and higher temperatures may<br />
then promote algal blooms and eutrophication. The low<br />
dissolved mineral and nutrient concentration of many of<br />
the rivers in the forested parts of Upper and Lower Guinea,<br />
and the Congo Basin, results in food webs dependent on<br />
allochthonous materials from the forest. The removal of<br />
riparian forest can affect peak fl ow fl ooding events, which<br />
impacts the freshwater species present and the human<br />
communities within the catchments (Bradshaw et al. 2007,<br />
2009; Brummett et al. 2009; Farrell et al. 2010). Thus,<br />
deforestation can signifi cantly affect the ecohydrology<br />
of river systems (and lakes, such as the crater lakes of<br />
western Africa (Stiassny et al. 2011)). Deforestation results<br />
in signifi cant increases in sedimentation as delicate forest<br />
soils, which are easily eroded, are exposed and washed
into streams, rivers, and lakes. The sediment covers<br />
submerged surfaces, reducing suitable habitat for breeding<br />
and feeding of many fi sh populations. The increased<br />
turbidity can clog the gills of fi shes and suffocate their<br />
eggs (Roberts 1993), as well as reducing the light levels<br />
so that submerged plants cannot photosynthesize and so<br />
die, exacerbating eutrophication of the waters.<br />
Expansion and intensifi cation of logging and agriculture<br />
are common causes of deforestation. Threats from<br />
deforestation are particularly strong and widespread in<br />
the Upper and Lower Guinea provinces, and in the Congo<br />
Basin, given that these encompass the last remaining<br />
extensively forested regions on the continent. Smith et<br />
al. (2009) note that in western Africa (which includes the<br />
Upper Guinea and Nilo-Sudan ichthyofaunal provinces),<br />
deforestation is especially prevalent along the banks of<br />
the Volta, Niger, and Senegal rivers. Some examples of<br />
fi shes in the Upper Guinea region that are impacted by<br />
deforestation are discussed in section 3.3.2. The central<br />
African region has already lost an estimated 46% of its<br />
rainforest to logging and conversion to agriculture, and<br />
continues to lose forested watershed at an average rate of<br />
7% per year (Revenga et al. 1998). The Kasai, Sanga, and<br />
Upper Congo freshwater ecoregions are some of the more<br />
seriously impacted areas in terms of ongoing deforestation,<br />
while the Lower Congo has been almost entirely deforested<br />
for many decades (Stiassny et al. 2011). Deforestation is a<br />
major threat in many parts of the Lower Guinea province,<br />
and much of this is associated with logging (e.g., in the<br />
Nyong (Cameroon), the Ogowe/Ivindo system (Gabon),<br />
and Kouilou/Niari systems (Republic of Congo)). However,<br />
as well as slash-and-burn agriculture, charcoal production,<br />
banana, rubber, and oil palm plantations are other major<br />
drivers of deforestation and associated sedimentation<br />
through Lower Guinea (especially in Cameroon), and in<br />
some parts of the Congo Basin (Brummett et al. 2009).<br />
Extremely biodiverse rainforest along the Cameroon<br />
coast, roughly from the Ndian to the Kribi (Kienké) rivers,<br />
has been converted to oil palm, as have parts of the Upper<br />
Congo freshwater ecoregion, and there are plans for<br />
development of oil palm plantations around Lake Tumba<br />
(Brummett et al. 2011). Deforestation has also taken place<br />
to allow for the development of eucalyptus plantations<br />
along the coast of the Republic of Congo and Cabinda.<br />
Eucalyptus is mainly used for building materials and for<br />
fi rewood and charcoal. Deforestation specifi cally for the<br />
production of fi rewood and charcoal is a more serious<br />
problem for parts of western and central Africa, especially<br />
near areas of urban development, where remnant riparian<br />
vegetation is particularly under threat (Smith et al. 2009;<br />
Brummett et al. 2011).<br />
Deforestation for logging also opens up large tracts of<br />
forest for further exploitation, for example, by mining and<br />
agriculture ventures, bush meat hunting, and settlement.<br />
Mining has signifi cantly contributed to further habitat<br />
Mining, such as observed here in the East Nimba Forest<br />
Reserve in Liberia, has signifi cantly contributed to habitat<br />
loss and high levels of sedimentation in many regions across<br />
Africa. © K.-D.B. DIJKSTRA<br />
loss and high levels of sedimentation in many regions.<br />
Small-scale alluvial mining, larger commercial mines, and<br />
extraction of sand or clay have impacted rivers basins in<br />
parts of southern, western, and central Africa (Darwall et<br />
al. 2009; Smith et al. 2009; Brooks et al. 2011). Mining<br />
is especially common in the Congo, Sanga, and Kasai<br />
basins. Land conversion for agriculture impacts riparian<br />
forests, as well as many other habitats. In eastern Africa,<br />
the high number of threatened fi shes in the Malagarasi<br />
Basin is due to encroachment of agriculture into the<br />
wetlands (Darwall et al. 2005). Similarly, the high number<br />
of threatened species in the Ruvu River, draining to the<br />
coast of Tanzania, probably results from many regionally<br />
endemic species being found in areas at high risk of<br />
habitat modifi cation.<br />
Livestock are frequently placed on the cleared land, causing<br />
overgrazing, and these poorly managed processes of land<br />
clearing and grazing result in accelerated soil erosion and<br />
increased sedimentation. The Ubangi River in central Africa<br />
is so severely impacted by sedimentation from mining and<br />
agriculture in the region of Mpoko that the reduced river<br />
depth prevents shipping for four or fi ve months in most<br />
years (Brummett et al. 2009).<br />
Not surprisingly, loss of riparian habitat and deterioration<br />
of freshwater ecosystems are greatest in areas of high<br />
human settlement. For example, several species of fi shes<br />
are impacted by habitat loss in the heavily populated lower<br />
Ogun Basin in Nigeria, as well as in the northern parts of<br />
Lower Guinea, and close to the large cities of Kinshasa<br />
and Kisangani in the Congo Basin (Stiassny et al. 2011). In<br />
several parts of the Congo Basin (particularly the eastern<br />
part of the basin and in the vicinity of the Sangha; see<br />
Brummett et al. 2011) and Upper Guinea, war and civil<br />
unrest have displaced local communities into previously<br />
undisturbed regions of forest, with these communities<br />
often settling along waterways (Thieme et al. 2005), with<br />
consequent loss of forest cover.<br />
CHAPTER 3 | FISH 67
CHAPTER 3 | FISH<br />
68<br />
Pollution<br />
Water pollution is a problem in many parts of Africa. The<br />
impacts of pesticides and fertilizers from agriculture on<br />
freshwater systems have been reported in all regional<br />
studies (Darwall et al. 2005, 2009; Smith et al. 2009; García<br />
et al. 2010a; Brooks et al. 2011). The effects of pollution<br />
are usually also coupled with impacts from increased<br />
sedimentation caused by soil erosion, and frequently<br />
result in eutrophication of the lakes and rivers (e.g., the<br />
Malagarasi Basin, which is impacted by agriculture<br />
(see above)). Pollution from mining is a serious threat in<br />
the regions where small-scale or commercial mining is<br />
prevalent (see above). The use of pesticides in vector control<br />
programs for diseases like malaria, schistosomiasis, and<br />
trypanosomiasis (e.g., in parts of western Africa (Smith et<br />
al. 2009)) may also impact fi shes, with low doses possibly<br />
compromising their physiology and behaviour. The use<br />
of such pesticides as a method of fi shing is an additional<br />
problem in central Africa.<br />
Organic pollution from human and domestic waste<br />
threatens fi shes in all areas where there are sizeable<br />
settlements. Pollution from oil exploration, factories or<br />
other urban industries, cars in the cities, and from boat<br />
traffi c on rivers impacts the freshwater systems, especially<br />
those close to large cities, such as Kinshasa and Lagos.<br />
Pollution from oil exploration, and associated loss of<br />
habitat, specifi cally threatens several restricted range<br />
species in the Niger Delta, and may pose a threat to<br />
species in coastal freshwater systems of Gabon, Cabinda,<br />
the Republic of Congo and Angola.<br />
Dams<br />
Some 1,207 dams have been constructed on small and<br />
large rivers of Africa; at least 135 of these are classifi ed as<br />
large dams (> 500,000m³) (FAO 2010) (see also <strong>Chapter</strong><br />
1, Figure 1.3). The greatest concentrations of dams,<br />
and some of the largest, are in the Maghreb province of<br />
northern Africa, much of western Africa, and Zimbabwe in<br />
southern Africa. Dams prevent the longitudinal migration<br />
of fi sh, and create lake-like conditions upstream that are<br />
uninhabitable for many of the riverine fi shes that were<br />
originally present. These lake reservoirs also attract<br />
the attention of anglers for the introduction of exotic<br />
lacustrine species that may become invasive in the lake<br />
environment, especially in southern Africa. Downstream<br />
fl ow and sediment load may be changed to such an<br />
extent that the habitats immediately below the dam are<br />
also unsuitable for habitation by previously native fi shes.<br />
The overall impact of dams may be severe, as noted<br />
for the Aswan Dam (see section 3.3.2). The potential<br />
for future development of hydropower projects in Africa<br />
is also large, including plans for some very large dams,<br />
such as the 39,000MW ‘Grand Inga’ dam proposed for<br />
development on the lower section of the Congo River<br />
by 2025 (see Brummett et al. 2011) (but see <strong>Chapter</strong> 1,<br />
section 1.2.2.1).<br />
Dams, such as the Akosombo Dam on the Volta River in<br />
Ghana, will have a signifi cant impact on freshwater species.<br />
These impacts need to be evaluated before construction,<br />
and operation procedures, for new dams are approved.<br />
© WILLIAM DARWALL<br />
River channelization and water abstraction<br />
The geomorphology and fl ow of many rivers is affected<br />
by channelization (often for irrigation or inter-basin<br />
transfer of water) and abstraction of water to supply<br />
agriculture, and industrial and domestic consumption.<br />
In South Africa, inter-basin transfer schemes of water<br />
have facilitated the spread of introduced species (Darwall<br />
et al. 2009), and may also allow the migration of native<br />
species to new basins beyond their normal range, with the<br />
possibility of interbreeding between populations that are<br />
normally isolated. This may result in detrimental genetic<br />
homogenization of the populations.<br />
Water abstraction is a threat particularly in arid areas that<br />
have human settlements, such as northern and southern<br />
Africa (see section 3.3.2). In western Africa, water<br />
abstraction (coupled with climate change), has severely<br />
impacted the wetlands of Lake Chad, greatly reducing the<br />
surface area of the lake from 25,000km² in the early 1960s<br />
to around 1,350 km² in 2001 (García et al. 2010a). When<br />
water removal signifi cantly disturbs the environmental<br />
fl ow of a stream or river, this may pose a signifi cant threat<br />
to any endemic species whose range is restricted to the<br />
The geomorphology and fl ow of many rivers is affected by<br />
channelization, as is the case here on the Kou stream in<br />
Burkina Faso. © TIMO MORITZ
The high diversity of fi sh communities, such as seen here on<br />
the Malagarasi, Tanzania, are often impacted when dams are<br />
constructed. © JOHN FRIEL<br />
area where the water removal is occurring. Small dams<br />
or weirs, river channelization, and water abstraction may<br />
all operate together in some places, where the weirs are<br />
installed to divert water and create deeper pools from<br />
which it is easier to extract the water by pump.<br />
Overfi shing<br />
Overfi shing is a major threat to many species in the Great<br />
Lakes of the African Rift Valley, such as in Lake Malawi<br />
(see section 3.3.2), particularly where fi sheries are focused<br />
along the migration routes of species as they move from<br />
the lake into river mouths to spawn. Smith et al. (2009)<br />
also note that many freshwater bodies in western Africa<br />
are overfi shed, particularly the Volta. Those authors report<br />
that there has been a disappearance of larger species<br />
in some western African rivers such as the Oueme, as a<br />
result of sequentially fi shing down the food web. Allan et<br />
al. (2005) note that this decline in fi sh size is accepted in<br />
parts of Africa, due to some regional preference for small<br />
fi sh in the cuisine. García et al. (2010b) note that Lates<br />
niloticus, Anguilla anguilla, Barbus bynnii, Hydrocynus<br />
forskahlii, and Alestes dentex are unsustainably harvested<br />
in northern Africa. Threats from poorly managed fi sheries<br />
(overharvesting or the use of small mesh, unselective fi shing<br />
gear, fi sh poisons and explosives) have also been reported<br />
for the Congo Basin, particularly in areas such as Malebo<br />
Pool, Lake Tumba, and Mai N’dombe. Some fi sh species<br />
are also the focus of commercial and artisanal fi sheries for<br />
the aquarium trade, for example, some killifi shes in Lower<br />
Guinea (Stiassny et al. 2011) and Arnoldichthys spilopterus<br />
(see section 3.3.2).<br />
Invasive species<br />
Invasive species are a problem in many parts of Africa,<br />
and their success has been aided by habitat modifi cation<br />
and the development of dams (see above). The impact<br />
on haplochromine cichlids and the introduced Nile perch<br />
(Lates niloticus) to Lake Victoria in East Africa has been<br />
extensively documented (see section 3.3.2 and 3.3.5). In<br />
eastern Africa, introduced tilapiine and haplochromine<br />
Arnoldichthys spilopterus (VU), the Niger tetra, is endemic to Nigeria, where it is restricted to the lower Ogun and lower Niger<br />
rivers. © TIMO MORITIZ<br />
CHAPTER 3 | FISH 69
CHAPTER 3 | FISH<br />
70<br />
Drought is a serious problem for many parts of Africa, with<br />
many lakes drying out, such as the Mare Bali water hole in<br />
Pedjari National Park, Benin, pictured here. © TIMO MORITZ<br />
cichlids are themselves a threat to a cyprinid, Varicorhinus<br />
ruandae, in Lake Luhondo (a small lake in the northern part<br />
of Rwanda), through competition and predation (De Vos et<br />
al. 1990).<br />
Invasive species pose the greatest recorded threat to<br />
fi shes in southern Africa, mainly coming from introduced<br />
European or North American species (e.g., Micropterus<br />
dolomieu; Oncorhynchus mykiss and Salmo trutta; also see<br />
Species in the Spotlight– Tilapia in eastern Africa –<br />
a friend and foe), and from invasive species of the cichlid<br />
genus Oreochromis. Introduced mosquitofi sh (Gambusia)<br />
have had very signifi cant impacts on native species in<br />
northern Africa (see section 3.3.2). The water hyacinth<br />
(Eichhornia crassipes) is one of the most widespread<br />
invasive species in Africa, presenting a considerable threat<br />
to freshwater ecosystems in most of sub-Saharan Africa.<br />
The economic impacts of the water hyacinth are estimated<br />
at USD 20-50 million every year in seven African countries,<br />
and may be as much as USD 100 million annually across<br />
all of Africa (Chenje and Mohamed-Katerere 2006) (see<br />
Species in the spotlight – Water hyacinth, a threat<br />
to the freshwater biodiversity).<br />
Climate change and extreme events<br />
Fish faunas already weakened by many of the threats<br />
noted above are especially susceptible to the impacts of<br />
natural disasters, such as drought, and to climate change.<br />
According to García et al. (2010b), natural disasters are<br />
the second most serious cause of decline for almost two<br />
thirds of the freshwater fi sh in northern Africa. Drought is a<br />
serious problem in parts of northern and southern Africa,<br />
where many once permanent streams have become<br />
seasonal or have dried completely. Lake Chad, in western<br />
Africa, was reduced to 5.4% of its surface area between<br />
the 1960s and 2001 (see above), and it is estimated that<br />
50% of this reduction was caused by changes in climate<br />
patterns (Pietersen and Beekman 2006) (see also <strong>Chapter</strong><br />
1, Figure 1.5). Climate change is also expected to impact<br />
the forested regions of Africa, although the precise nature<br />
of these impacts is unclear (Schiermeier 2008; Thieme<br />
et al. 2010; Brummett et al. 2011). It is expected that,<br />
by the 2050s, more than 80% of Africa’s freshwater fi sh<br />
species may experience hydrologic conditions that are<br />
substantially different from the existing conditions. A more<br />
detailed account of the impacts of climate change are<br />
given in <strong>Chapter</strong> 8, this volume.<br />
In conclusion, there are many types of threat that are<br />
impacting the freshwater fi sh fauna of Africa. In most<br />
cases, decline in population size and distribution of any<br />
species is a product of a combination of these factors,<br />
rather than the result of any single threat. Harrison and<br />
Stiassny (1999) discuss how several combined threats<br />
resulted in the decline of many of the cichlid species of<br />
Lake Victoria in the 1980s (see also <strong>Chapter</strong> 1, Box 1.1).<br />
A more recent example is provided by the calamitous<br />
decline of the European eel (Anguilla anguilla) (Critically<br />
Endangered) throughout its range. In northern Africa,<br />
this decline may be attributed to the combined effects of<br />
overfi shing of silver eels in coastal waters, the impacts of<br />
parasitic pathologies, pollution, construction of dams and<br />
water abstraction, and gravel extraction from river beds<br />
(García et al. 2010b). Similarly, the decline of Aphanias<br />
saourensis to Critically Endangered status in northern<br />
Africa is attributable to a combination of factors including<br />
groundwater abstraction, pollution, and introduction of<br />
invasive species (see section 3.3.2).<br />
3.5 Research actions required<br />
This study highlights specifi c patterns of the distribution<br />
and conservation status of fi shes throughout the continent,<br />
and has identifi ed some conservation recommendations<br />
(see section 3.6) and priorities. However, the study has<br />
also confi rmed the opinion of Lundberg et al. (2000) that<br />
signifi cant gaps in our knowledge of this fauna still exist.<br />
Considerable additional research is required to provide<br />
basic baseline data for several potentially biodiverse<br />
regions to support conservation management. The lack<br />
of knowledge of even the most basic distributional and<br />
ecological data, and the need for further research on the<br />
fi shes of large parts of Africa (in particular, the Congo<br />
Basin), is well exemplifi ed by the increased knowledge<br />
resulting from the recent study on the ichthyofauna of the<br />
Léfi ni River (as reported above), a tributary draining from<br />
the west into the middle Congo (Ibala-Zamba 2010).<br />
Where information is available on species’ presence and<br />
distributions within a catchment, there is, however, often<br />
very little additional information about the ecology of each<br />
particular species. In the absence of such information it is<br />
very diffi cult to accurately assess the conservation status of<br />
the species, and to make suitable conservation decisions.<br />
A major challenge is to accumulate data that can help<br />
reduce the total number of species that are classifi ed as
Data Defi cient (514 species, 18% of all assessed species)<br />
(see <strong>Chapter</strong> 8 for discussion of prioritising fi eld work to fi ll<br />
these knowledge gaps).<br />
The impact of climate change on freshwater ecosystems<br />
of Africa is an important concern, but one which may be<br />
mitigated through a better understanding of the resilience<br />
of species to change, and through proper management<br />
of freshwater resources (see chapter 8, section 8.6.1).<br />
However, this also requires further research on the diversity<br />
and ecology of the species present, the environmental<br />
fl ows required to support this biodiversity, and additional<br />
meteorological and hydrological data. These features<br />
of climate change must then be considered alongside<br />
other threats (for example, dam construction, overfi shing,<br />
or deforestation) in order to predict the overall impact.<br />
Spatial modelling is useful for all these studies, although<br />
interpretation of the results is often diffi cult for aquatic<br />
species. The diffi culties lie mainly in the complexities of<br />
modelling characteristics of underwater habitats, and<br />
in identifying routes of dispersal within catchments,<br />
and barriers to this dispersal both within and between<br />
catchments.<br />
The best instrument for evaluating changes in the<br />
ichthyofauna of Africa is long-term, standardized<br />
monitoring. This will detect shifts in species composition,<br />
as well as changes in biomass and local incidences of<br />
fi shing down the food web. Standardized monitoring<br />
should also be implemented as a routine step before any<br />
large infrastructure is developed on or along the waterway.<br />
In many regions, however, political and socio-economic<br />
instability, logistical problems, and lack of fi nances and<br />
taxonomic expertise combine to hamper even the most<br />
basic studies, including monitoring programmes. If these<br />
problems of infrastructure, training and fi nance can be<br />
remediated, then hopefully the recommended studies of<br />
species and population diversity and ecology to monitor<br />
the health of the fauna can be initiated.<br />
3.6 Conservation recommendations<br />
The fi ndings of this assessment confi rm that the freshwater<br />
fi shes of Africa are signifi cantly threatened in many parts<br />
of the continent, and it is reasonable to assume that even<br />
greater stress will be exerted on this fauna in the future.<br />
Nevertheless, it may be diffi cult for policy makers to set<br />
conservation recommendations as priorities when there<br />
are many other urgent issues associated with ensuring that<br />
people are guaranteed an acceptable standard of living. The<br />
many actions required to meet basic human requirements<br />
may appear to be at odds with the objectives of freshwater<br />
biodiversity conservation. There is, however, a need to<br />
The importance of inland fi sheries to local economies can been seen by this thriving fi sh market on the banks of the Congo<br />
River, at Mbandaka, D. R. Congo. © R. SCHELLY<br />
CHAPTER 3 | FISH 71
CHAPTER 3 | FISH<br />
72<br />
The banded distichodus, Distichodus sexfasciatus (LC), is widespread throughout central Africa. A beautiful species such as<br />
this is collected for the aquarium trade – it is also an important food fi sh. © SAIAB/ROGER BILLS<br />
protect and sustainably manage freshwater ecosystems to<br />
deliver the many ecosystem services that are also essential<br />
to people. At the most basic level, effectively functioning<br />
aquatic ecosystems with healthy fi sh populations that<br />
can be sustainably exploited are to the benefi t of all. But<br />
recent studies, such as that by Vörösmarty et al. (2010),<br />
have shown that the need for sustainable management of<br />
freshwater resources goes far beyond ensuring biodiversity<br />
conservation and food security through reliable fi sheries.<br />
Their study has shown that the provision of adequate<br />
human water security in wealthy nations (such as in parts<br />
of Europe and North America) has only been possible by<br />
massive fi nancial investment in water technology to offset<br />
the impacts of threats, but this investment is not possible<br />
in less wealthy nations, where biodiversity and human<br />
water security remain vulnerable. In much of Africa, the<br />
sustainable management of freshwater resources and<br />
biodiversity offers a cost effective and environmentally<br />
sustainable alternative.<br />
Habitat loss or modifi cation ranks among the primary<br />
threats for extinction of freshwater fi shes not only in Africa<br />
(see section 3.4) but also worldwide (Harrison and Stiassny<br />
1999) (see <strong>Chapter</strong> 1, this volume). Adequate protection<br />
and management of freshwater and riparian habitats is,<br />
therefore, a key recommendation for the conservation of<br />
freshwater fi shes. Numerous nominal parks and reserves<br />
exist in Africa (e.g., see Stiassny et al. 2011) but, in practice,<br />
many are focused on terrestrial habitats rather than the<br />
freshwater ones that exist within them or along their<br />
borders, where they are especially vulnerable (Abell et al.<br />
2007; Allan et al. 2010). For example, it is not uncommon<br />
that fi shing rights are still exerted within protected areas<br />
or parks. Development of parks and protected areas<br />
that specifi cally address the conservation challenges for<br />
rivers, lakes and wetlands will be important for the future<br />
conservation of Africa’s freshwater fi shes. For example,<br />
in central Africa the Ministry of Environment for D. R.<br />
Congo and the Institut Congolais Pour la Conservation<br />
de la Nature (ICCN) initiated a country-wide biodiversity<br />
assessment to identify priority areas for conservation,<br />
and have identifi ed 30 wetland priority areas (Thieme et<br />
al. 2008).<br />
Lake Malawi National Park was established in 1980,<br />
especially aimed at protecting part of Lake Malawi’s<br />
unique fi sh fauna; nevertheless, the scope of the park is<br />
quite limited, and more initiatives are required to reduce<br />
the high levels of threat, such as from overfi shing. To this<br />
end, special programmes have been implemented, such<br />
as the ‘chambo restoration strategic plan’ to protect one of<br />
the more important taxa (Oreochromis) in the lake fi shery.
Allan et al. (2010) discuss the need for a new conceptual<br />
framework for developing and managing protected areas<br />
that accounts for the need to conserve ecosystems while<br />
also allowing for the diverse requirements of people. This<br />
concept is based on Abell et al.’s (2007) recommendation<br />
for a multiple-use zoning framework, where focal areas<br />
for freshwater conservation are embedded in critical<br />
management zones, and these, in turn, are embedded in<br />
catchment management zones. Various other programs,<br />
ranging from large-scale management of catchments to<br />
species-specifi c or site-specifi c programmes, have been<br />
shown to be important for conservation of freshwater<br />
biodiversity and provision of ecosystem services. García et<br />
al. (2010b) noted that Integrated River Basin Management<br />
(IRBM) is a key conservation action required to stop<br />
species decline in northern Africa. Cross (2009) described<br />
the catchment-scale actions of the Pangani River Basin<br />
Flow Assessment Initiative (FA), co-ordinated by the <strong>IUCN</strong>-<br />
Pangani Basin Water Offi ce (PBWO). Tweddle et al. (2009)<br />
describe some successful projects and conservation<br />
recommendations in southern Africa directed at diverse<br />
scales, from whole landscapes to local sites and individual<br />
species. These include eradication of invasive species and<br />
limitation of the use of alien invasive species in aquaculture<br />
programmes. Such actions, at both the site and catchment<br />
scales, are covered in more detail in <strong>Chapter</strong> 9.<br />
Another recommendation is that reliable fi sheries catch<br />
statistics are maintained and made available. In many<br />
areas, little information is available on species composition<br />
and catch quantities. This is directly related to the lack<br />
of inventories and identifi cation keys in many areas<br />
and a chronic lack of taxonomically trained personnel.<br />
Trained local staff would then assimilate and translate<br />
the knowledge of local fi shermen and make it available<br />
to resource managers and scientists to inform decisions<br />
and policy.<br />
On a more general note, ‘Payment for Ecosystem Services’<br />
(PES) programs are mechanisms where the benefi ciaries of<br />
freshwater ecosystem services pay for the supply of these<br />
services (Forsland et al. 2009). For example, downstream<br />
communities that receive clean water for domestic and<br />
agricultural use pay the upstream communities to conserve<br />
and manage the habitats so they continue to supply clean<br />
and plentiful water. Conservation stewardship agreements<br />
are another form of PES scheme, where trust funds are set<br />
up to supply local communities with fi nancial incentives<br />
(for example, money for new jobs, or healthcare) in return<br />
for agreement that they manage their freshwater resources<br />
sustainably.<br />
Many of the conservation recommendations discussed<br />
above will be impossible if not backed by adequate<br />
development of policy and enforcement of regulations<br />
and laws (Smith 2010). However, many governmental<br />
bodies lack the fi nancial and logistical means to enforce<br />
existing laws and rules, and therefore mechanisms must<br />
be put in place to assist in this process. The Convention<br />
of Biological Diversity (ratifi ed by several African<br />
countries) and the Ramsar Convention (Landenbergue<br />
and Peck 2010) may help with the development of policy<br />
and laws. The Convention was supported recently by the<br />
agreement at the 10th Conference of the Parties to the<br />
Convention on Biological Diversity (CBD-COP10, held at<br />
Nagoya in November 2010) that 17% of terrestrial and<br />
inland water areas globally should be protected, and that<br />
’By 2020 the extinction of known threatened species<br />
has been prevented and their conservation status,<br />
particularly of those most in decline, has been improved<br />
and sustained.’<br />
In conclusion, our objective for the future must be to<br />
effectively conserve and manage African freshwater fi sh<br />
biodiversity, at the same time as supporting the livelihoods<br />
and economies of the people who are dependent on these<br />
resources and are the critical stakeholders in ensuring<br />
sustainable management practices. This can only be<br />
achieved by multi-disciplinary approaches to scientifi c<br />
research, development of tools for the application of<br />
that research to conservation and management, and the<br />
implementation of policy that supports recommendations<br />
made as a consequence of all of the above (Farrell 2010;<br />
Smith 2010).<br />
CHAPTER 3 | FISH 73
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
74<br />
Species in the spotlight<br />
The Congo blind barb: Mbanza-<br />
Ngungu’s albino cave fi sh<br />
The enigmatic, Congo blind<br />
barb, Caecobarbus geertsii,<br />
was scientifi cally described<br />
by Boulenger (1921), based<br />
on four specimens collected in 1920,<br />
from the ‘Grottes de Thysville’ in<br />
the Lower Congo region (Roberts<br />
and Stewart 1976) of D. R. Congo.<br />
It was the fi rst African cave fi sh to<br />
be discovered. The species is locally<br />
referred to as ‘Nzonzi a mpofo’ in<br />
Kikongo (the local Ndibu dialect)<br />
which literally means ‘blind barb’.<br />
Although the eyes are not visible,<br />
they are present. They are deeply<br />
embedded in the head, lack a<br />
lens, and have only a rudimentary<br />
retina and optical nerve (Gerard<br />
1936). Nevertheless, Thinès (1953),<br />
contrary to Petit and Besnard<br />
(1937), notes that the species moves<br />
away from light, demonstrating a<br />
typical photonegative reaction due<br />
to the existence of extra-ocular<br />
photosensitivity.<br />
The species also lacks<br />
pigmentation (Boulenger 1921;<br />
Heuts 1951) and is considered a<br />
true albino, as placing live animals<br />
under light for more than one month<br />
does not result in development of<br />
pigment (Gerard 1936). However,<br />
Poll (1953) reported the presence of<br />
melanophores in a specimen kept<br />
for seven months in an aquarium.<br />
The lateral vein creates a vivid red<br />
Vreven, E.¹, Kimbembi ma Ibaka, A.² and Wamuini Lunkayilakio, S²<br />
A live specimen of Caecobarbus geertsii from the cave ‘Grotte de Lukatu’, D. R. Congo. © ROYAL MUSEUM FOR CENTRAL AFRICA<br />
band along the lateral line. Below<br />
the operculum the gills are visible as<br />
a purplish region, and the intestinal<br />
region is visible through the<br />
abdomen (Petit and Besnard 1937).<br />
Heuts (1951) estimated longevity<br />
at nine to 14 years; Proudlove and<br />
Romero (2001) stated the lifespan<br />
may exceed 15 years, but this needs<br />
to be confi rmed. The species reaches<br />
a maximum size of 80 to 120mm<br />
total length, based on the largest<br />
specimen housed at the Royal<br />
Museum for Central Africa.<br />
Following explorations of<br />
several caves in 1949, Heuts (1951)<br />
and Heuts and Leleup (1954)<br />
recorded C. geertsii from seven<br />
caves around Mbanza-Ngungu<br />
(formerly Thysville), situated on the<br />
western slope and the top of the<br />
Thysville mountain ridge (Monts<br />
de Cristal: 750 to 850m elevation).<br />
One population was reported as<br />
extirpated by the exploitation of<br />
limestone between 1930 and 1935<br />
(Leleup 1956; see also Heuts and<br />
Leleup 1954). Indeed, a visit to the<br />
cave site in 2005 found it to have<br />
completely disappeared following<br />
excavation of the slope.<br />
The presence of C. geertsii in at<br />
least four of the other caves reported<br />
by Heuts and Leleup (1954) has<br />
been confi rmed by recent surveys by<br />
Kimbembi (2007) and the authors.<br />
1 Royal Museum for Central Africa, Vertebrate Section, Ichthyology, Leuvensesteenweg 13, B-3080 Tervuren, Belgium<br />
2 Institut Supérieur Pédagogique de Mbanza-Ngungu, Bas-Congo, Laboratoire de Biologie, Democratic Republic of Congo<br />
Statistical population surveys<br />
have been impossible because the<br />
subterranean habitat is extensive<br />
and diffi cult to sample (Heuts<br />
1951); however, a gross population<br />
estimate for the seven caves<br />
reported by Heuts and Leleup (1954)<br />
would be about 7,000 individuals<br />
(based on information supplied by<br />
those authors). Kimbembi (2007)<br />
discovered seven more caves with at<br />
least small populations of C. geertsii,<br />
although no population estimations<br />
have been made for these.<br />
Heuts (1951) and Heuts and<br />
Leleup (1954) previously considered<br />
C. geertsii to be present in only two<br />
upper tributaries of the Kwilu Basin<br />
(an affl uent of the Lower Congo),<br />
namely the Fuma and the Kokosi.<br />
One of the new caves that Kimbembi<br />
(2007) identifi ed as holding C.<br />
geertsii is on the Tobo River, another<br />
affl uent of the Kwilu Basin. Lévêque<br />
and Daget (1984) and Banister (1986)<br />
also reported the species from the<br />
Inkisi Basin, but at the time had no<br />
evidence for this. However, inferred<br />
from mapping of the new cave<br />
localities identifi ed by Kimbembi<br />
(2007), the species’ presence in<br />
the Inkisi River basin seems to be<br />
confi rmed by two of them – one on<br />
the Tubulu River and another one<br />
on the Uombe or possibly the Kela<br />
River, a tributary to the Uombe. The
presence of C. geertsii in D. R. Congo,<br />
as reported by Lévêque and Daget<br />
(1984), is incorrect. Thus, the entire<br />
distribution area of the species is<br />
about 120km 2 . Heuts (1951) noted<br />
important differences between the<br />
different populations of C. geertsii<br />
in the Kwilu basin. Populations<br />
present in affl uents of the Kokosi<br />
River have an opercular guanine<br />
spot which may cover one third of<br />
the operculum (in addition to a few<br />
other guanine spots and marks).<br />
This spot is absent in all other<br />
populations (affl uents of the Fuma<br />
River). Furthermore, within one<br />
cave the population has a serrated<br />
dorsal spine, which was not found in<br />
all other populations examined by<br />
Heuts (1951).<br />
Traditionally, caves are sacred<br />
in the area (Laman 1962) and, as a<br />
result, access to most of the caves is<br />
restricted still today. A law, passed<br />
on 21 April 1937, protected C. geertsii<br />
from all hunting and fi shing, except<br />
for scientifi c purposes (Frenchkop<br />
1941, 1947, 1953; Duren 1943). The<br />
species was added to the CITES<br />
Annex II (on 6 June 1981), resulting<br />
in an international trade restriction<br />
which means that the species cannot<br />
be traded without appropriate<br />
export and import permits. C. geertsii<br />
is still the only African freshwater<br />
fi sh species on the CITES list. The<br />
<strong>IUCN</strong> Red List status of C. geertsii<br />
is Vulnerable (VU), due to a limited<br />
geographic range and a decline in<br />
the area and quality of its habitat<br />
(Moelants 2009).<br />
Caecobarbus geertsii was found in<br />
only seven of the 45 caves explored<br />
by Heuts and Leleup in 1949. This<br />
indicates, according to Heuts<br />
(1951) and Heuts and Leleup (1954),<br />
that caves must have a specifi c<br />
combination of ecological conditions<br />
if they are to be populated by C.<br />
geertsii, and they summarised the<br />
following conditions:<br />
1. high calcium bicarbonate<br />
concentrations in the water; and<br />
2. a distinct periodicity of the<br />
subterranean river fl ow regime<br />
through the caves.<br />
Due to this periodic inundation<br />
of the caves inhabited by C.<br />
geertsii, other typical cave animals,<br />
such as terrestrial insects, are<br />
absent. Therefore, C. geertsii is<br />
entirely dependent on an external,<br />
exogenous, food supply to the caves<br />
during the rainy season with, as a<br />
result, important fl uctuations in<br />
food resources between seasons.<br />
Moelants (2009) states that<br />
the species may feed on small<br />
crustaceans living in the caves,<br />
but this needs to be confi rmed.<br />
Consequently, growth is extremely<br />
slow, and all further available data<br />
suggest a very low reproduction rate,<br />
justifying protection measurements.<br />
A visit to the Kambu cave by the<br />
authors in August 2009 failed to fi nd<br />
the species, although its presence<br />
had been reported by Kimbembi<br />
(2007). However, several individuals<br />
of at least one species of Clarias (±<br />
200mm standard length) were found<br />
in the different isolated pools. This<br />
observation suggests predation of<br />
C. geertsii by species of Clarias, as<br />
previously proposed by Heuts and<br />
Leleup (1954) and by Leleup (1956).<br />
Caecobarbus geertsii has, in the<br />
past, been traded as an aquarium<br />
fi sh, with large numbers having<br />
been exported to industrialized<br />
nations. Collection pressure should<br />
have been reduced through listing<br />
under CITES; however, a CITES<br />
certifi cate was issued to import<br />
1,500 individuals to the Unites<br />
States (Proudlove and Romero<br />
2001). Three other primary threats<br />
to the species were identifi ed by<br />
Brown and Abell (2005): changes<br />
in hydrology of the small rivers<br />
feeding the caves; increasing<br />
human population; and associated<br />
deforestation (Kamdem Toham<br />
et al. 2006). Since 2003, with the<br />
attenuation of the political situation<br />
in D. R. Congo and the rehabilitation<br />
of the Matadi-Kinshasa road, there<br />
has been a signifi cant infl ux of<br />
rural people towards Mbanza-<br />
Ngungu. Consequently, land use has<br />
increased around Mbanza-Ngungu<br />
for buildings as well as agriculture.<br />
One cave is now used as a quarry,<br />
with consequential loss of the<br />
Caecobarbus population (Leleup<br />
1956; Poll 1956; and see above), and<br />
others are at risk of collapse due<br />
to human disturbance (Kimbembi<br />
2007; Moelants 2009). Agriculture is<br />
practiced preferentially in the valleys<br />
near to the caves but may also occur<br />
on the hillside slopes surrounding<br />
and covering the caves, leading to<br />
increased erosion and landslides. In<br />
the past, these areas were covered<br />
with lowland rainforest and<br />
secondary grassland (White 1986),<br />
limiting erosion. Further research<br />
and conservation initiatives in the<br />
fi eld are necessary if this unique<br />
species of fi sh is to survive.<br />
Land use around the entrance of the ‘Grotte de Lukatu’, with subsequent landslides<br />
visible (9 March 2007). The entrance to the cave is directly below the largest trees<br />
in the middle of the photograph. © ROYAL MUSEUM FOR CENTRAL AFRICA<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
75
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
76<br />
Species in the spotlight<br />
Tilapia in eastern Africa – a friend and foe<br />
Tilapia form the basis for<br />
much of the aquaculture<br />
industry that is important<br />
to so many people across<br />
Africa. Its success as a commercially<br />
fi shed and cultured species is<br />
attributed to several characteristics:<br />
its ability to establish and occupy<br />
a wide variety of habitats; its wide<br />
food spectrum from various trophic<br />
levels (Moriarty 1973; Moriarty<br />
and Moriarty 1973; Getachew 1987;<br />
Khallaf and Aln-Na-Ei 1987); high<br />
growth rate; large maximum size;<br />
and high fecundity (Ogutu-Ohwayo<br />
1990). All of these factors accord O.<br />
niloticus with great competitiveness<br />
over other tilapia, which can<br />
become a problem where they have<br />
been introduced, or escaped, to<br />
areas outside of their native range.<br />
Aquaculture is also one of the most<br />
common sources of invasive species<br />
in many parts of the world, and the<br />
famous Nile tilapia (Oreochromis<br />
niloticus niloticus), in particular, is<br />
recognised as a signifi cant threat<br />
to other native fi sh species. The<br />
popularity of tilapia in Africa is<br />
indicated by their high market<br />
value and, consequently, the high<br />
fi shing pressure in most lakes and<br />
rivers (Abban et al. 2004; Gréboval<br />
et al. 1994).<br />
The Nile tilapia<br />
Eastern Africa is endowed with<br />
six sub-species of Nile tilapia: O.<br />
niloticus niloticus (Linnaeus, 1758),<br />
originally from the White Nile<br />
Basin but now widely introduced<br />
elsewhere; O. niloticus eduardianus<br />
(Boulenger, 1912) in Lakes Edward,<br />
Kivu, Albert and George; O.<br />
niloticus vulcani (Trewavas, 1933) in<br />
Lake Turkana; O. niloticus sugutae<br />
Trewavas, 1983 in the Suguta<br />
The Nile tilapia, Oreochromis niloticus,<br />
(LC), a highly favoured species for<br />
aquaculture. © LUC DE VOS<br />
river basin; O. niloticus baringoensis<br />
Trewavas, 1983 in Lake Baringo;<br />
and one other recently discovered<br />
(Nyingi et al. 2009), but still<br />
undescribed subspecies from the<br />
Lake Bogoria Hotel spring near<br />
the Loboi swamp, between Lake<br />
Baringo and Bogoria in the Kenyan<br />
Rift Valley.<br />
Oreochromis niloticus was<br />
introduced to Lake Victoria for<br />
the purpose of improving tilapia<br />
fi sheries in several phases between<br />
1954 and 1962, due to decreasing<br />
stocks of native tilapia species<br />
O. esculentus and O. variabilis.<br />
Oreochromis niloticus rapidly<br />
colonized the entire lake and by<br />
the end of the 1960s was well<br />
established in inshore habitats<br />
(Mann 1970; Ogutu-Ohwayo 1990;<br />
Twongo 1995). It is thought that<br />
the introduction of O. niloticus<br />
caused the disappearance of the<br />
two native tilapia species (O.<br />
variabilis and O. esculentus) from<br />
the main part of the lake – O.<br />
esculentus having once represented<br />
the bulk of the fi sheries in the lake.<br />
It was initially hypothesised that<br />
hybridization with subspecies of<br />
O. niloticus was the main driver of<br />
the decline of O. variabilis and O.<br />
esculentus, because O. niloticus is well<br />
known for its ability to hybridize<br />
Nyingi, D.W.¹ and Agnèse, J.-F.² , ³<br />
with other tilapiines (Welcomme<br />
1988; Mwanja and Kaufman 1995;<br />
Rognon and Guyomard 2003;<br />
Nyingi and Agnèse 2007). However,<br />
the competitive superiority of<br />
O. niloticus subspecies over the<br />
two former native species was<br />
demonstrated to be the most likely<br />
contribution for their extinction<br />
(Balirwa 1992; Agnèse et al. 1999).<br />
Tilapia and aquaculture<br />
The greatest limitation to<br />
development of aquaculture in<br />
eastern Africa has been fi nancial,<br />
with all new activities in the<br />
sector initiated and dependent on<br />
foreign fi nancing. In Kenya, the<br />
government has stepped up efforts<br />
to promote aquaculture under the<br />
Economic Stimulus Programme.<br />
The government’s intention has<br />
been to highlight fi sh farming as<br />
a viable economic activity in the<br />
country by raising the income of<br />
farmers and other stakeholders in<br />
the fi shing industry. The project,<br />
worth 1,120 million Kenya shillings<br />
(EUR 10.67 million) was launched<br />
by the Ministry of Fisheries<br />
Development to construct 200<br />
fi sh ponds in 140 constituencies<br />
by June 2013. According to<br />
existing plans, each constituency is<br />
geared to receive 8 million Kenya<br />
shillings (EUR 70,000) for ponds.<br />
In Kenya, the Sagana Fish farm,<br />
under the Fisheries Department,<br />
provides fi ngerlings for warm<br />
water freshwater species. So far,<br />
the centre has been effi cient in<br />
provision of seed fi sh to farmers<br />
and in research and production<br />
of suitable feed. Despite these<br />
advances, considerable investment<br />
is still needed to ensure the<br />
provision of suitable species for<br />
1 National Museums of Kenya, P.O. Box 40658 Nairobi, 00100, Kenya<br />
2 Département Biologie Intégrative, CNRS UMR 5554, Université de Montpellier II CC 63 Place Eugène Bataillon F- 34095 Montpellier Cedex 5, France<br />
3 Institut de Recherche pour le Développement, 213 rue La Fayette 75180 Paris Cedex 10, France
the various regions, ensuring<br />
development of the industry.<br />
With the government supporting<br />
new initiatives, the greatest<br />
challenge is to identify a suitable<br />
species that will ensure high<br />
yield, while also safeguarding<br />
native species from the impacts of<br />
introduced aquaculture species.<br />
Unfortunately, in Africa the search<br />
for suitable species for aquaculture<br />
has often disregarded potential<br />
impacts on the native species.<br />
The most important culture<br />
species are still mainly taken<br />
from the wild, and populations<br />
are often translocated to basins<br />
far beyond their native range,<br />
potentially bringing closely related<br />
but formerly isolated species or<br />
populations into contact with<br />
each other. Where there has been<br />
inadequate research and planning,<br />
an introduced cultured species<br />
may directly compete with native<br />
species, or may hybridize with<br />
them, as noted above for O. niloticus<br />
when it was introduced to Lake<br />
Victoria. Unfortunately, O. niloticus<br />
has, in many cases, been the species<br />
of choice for aquaculture, therefore<br />
leading to further problems of<br />
competition and hybridisation.<br />
Oreochromis leucosticus was<br />
originally known from drainages<br />
near the border of Uganda and<br />
the D. R. Congo, specifi cally<br />
Lakes Edward, and Albert, and<br />
associated affl uents. However, it<br />
was introduced to Lake Naivasha<br />
in Kenya in 1957 (Harper et al.<br />
1990). About 150km away from<br />
Lake Naivasha is Lake Baringo, in<br />
the Kenyan Rift Valley, home to the<br />
endemic subspecies of Nile tilapia,<br />
O. niloticus baringoensis. Nyingi<br />
and Agnèse (2007) note that O.<br />
niloticus baringoensis share genetic<br />
characteristics of O. leucosticus,<br />
suggesting that O. leucosticus might<br />
have been introduced also to Lake<br />
Baringo, with some subsequent<br />
transfer of genetic material through<br />
hybridization with O. niloticus<br />
baringoensis. Even though impacts<br />
of the possible introduction of<br />
O. leucosticus are still unknown,<br />
introductions of tilapiines continue<br />
to be made within the region,<br />
either intentionally or accidentally<br />
through escape from culture ponds.<br />
Such issues are a clear indication<br />
of a failure of well-defi ned policies,<br />
or implementation of the existing<br />
regulations, for the management<br />
of natural fi sheries resources in<br />
Kenya. Through lack of awareness,<br />
and desperation to increase<br />
yield, fi sh farmers are breeding<br />
alien species of tilapia that could<br />
naturally hybridize in a similar<br />
manner – as seems to have occurred<br />
in Lake Baringo. Consequently,<br />
native species may be lost in several<br />
parts of eastern Africa, as already<br />
observed in Lake Victoria.<br />
As noted above, a new subspecies<br />
of Oreochromis was recently<br />
discovered from the Lake Bogoria<br />
Hotel spring near the Loboi swamp.<br />
This population was formerly<br />
Fish farms, such as this one in Malawi, represent an important source of food and income for people throughout Africa.<br />
However the traits that make species such as Oreochromis niloticus suitable for aquaculture mean that they pose a signifi cant<br />
threat to local species should they escape. © RANDALL BRUMMETT<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
77
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
78<br />
thought to have been introduced,<br />
but genetic and morphological<br />
analysis demonstrated its<br />
originality (Nyingi 2007; Nyingi<br />
and Agnèse 2007). The main body of<br />
the Loboi swamp acts as a physical<br />
and chemical barrier between the<br />
warm water springs (where the<br />
new sub-species is found) that<br />
fl ow into the swamp, and the<br />
Loboi River, which drains from it<br />
to Lake Baringo. The swamp has a<br />
signifi cantly low dissolved oxygen<br />
level (around 4% saturated dissolved<br />
oxygen, compared to around 60%<br />
in the springs and groundwater<br />
discharges), which is a consequence<br />
of high oxygen consumption during<br />
aerobic decomposition of detritus<br />
from macrophytes in the swamp<br />
(Ashley et al. 2004).<br />
The new apparent sub-species<br />
from the springs draining into the<br />
Loboi swamp offers interesting<br />
new possibilities for aquaculture<br />
development, if managed properly.<br />
The sub-species inhabits high<br />
temperatures (approximately 36°C)<br />
and may have developed hypoxic<br />
resistance mechanisms as dissolved<br />
oxygen levels may also be low. This<br />
sub-species may also have developed<br />
special mechanisms to regulate its<br />
sex-ratio, since sex determination<br />
is known to be infl uenced by<br />
high temperatures (Baroiller and<br />
D’Cotta 2001; Tessema et al. 2006).<br />
Therefore, the new sub-species<br />
may be a model for the study of sex<br />
determination in Oreochromis.<br />
However, the population from<br />
the Loboi swamp and associated<br />
rivers is under threat from human<br />
encroachment. The Loboi swamp<br />
itself has receded by around 60%<br />
over the last 30 years due to water<br />
abstraction for irrigation since 1970<br />
(Ashley et al. 2004; Owen et al. 2004).<br />
In addition, periodic avulsions<br />
have caused changes in the course<br />
of rivers in this region. The most<br />
recent was during the El Niñoinduced<br />
heavy rains of 1997, which<br />
caused changes in the courses of the<br />
Loboi and Sandai Rivers. The Sandai<br />
River now partly fl ows into Lake<br />
Baringo and partly to Lake Bogoria.<br />
...many challenges still lie<br />
ahead, and it will be critical<br />
to reinforce policy and<br />
management action with<br />
programmes of public awareness<br />
and education.<br />
Similarly, the Loboi River, which<br />
used to feed Lake Baringo, has<br />
changed its course and now fl ows<br />
to Lake Bogoria. These changes<br />
of fl ow were also due to intensive<br />
agricultural encroachment by<br />
local farmers leading to weakening<br />
of the river banks (Harper et<br />
al. 2003; Owen et al. 2004). This<br />
situation is not unique to the<br />
Loboi swamp but is common in<br />
almost all lakes and river systems in<br />
Kenya. The National Environment<br />
Management Authority in Kenya<br />
has been actively involved in<br />
ensuring rehabilitation of the<br />
Nairobi River, which had been<br />
greatly impacted due to solid waste<br />
disposal, sewage, run-off from car<br />
washes, and other human activities<br />
within the city and suburbs of<br />
Nairobi (Nzioka 2009). The success<br />
of this project is a clear indication<br />
that the National Environment<br />
Management Authority is able to<br />
protect hydrological systems in<br />
Kenya. There is, however, a need to<br />
replicate these successes elsewhere.<br />
Management of tilapia<br />
fi sheries<br />
A signifi cant challenge has existed<br />
where freshwater resources are<br />
shared by different countries. For<br />
example, fi sheries management<br />
of Lake Victoria was highly<br />
compromised in the early 1960s<br />
following independence of the<br />
countries bordering the lake<br />
(Kenya, Uganda and Tanzania),<br />
when they adopted different<br />
fi shing regulations based on the<br />
stocks targeted for exploitation<br />
(Marten 1979). These different<br />
regulations and priorities for<br />
exploitation have made it diffi cult<br />
to manage the lake as a complete<br />
ecosystem (Ntiba et al. 2001; Njiru<br />
et al. 2005). Ironically, this lack<br />
of management has contributed<br />
to declines in the introduced O.<br />
niloticus, which was previously<br />
responsible for the decline in the<br />
native sub-species (see above).<br />
Stock analyses for O. niloticus<br />
surveys of 1998 to 2000 and 2004<br />
to 2005 show that artisanal catches<br />
were dominated by immature<br />
fi sh, most being below the legally<br />
allowed total length of 30cm (Njiru<br />
et al. 2007). The paucity of mature<br />
individuals observed in commercial<br />
catches (Njiru et al. 2005) may<br />
be partly due to the increased<br />
numbers of introduced Nile<br />
perch (Lates niloticus) (Lubovich<br />
2009), but is also probably due<br />
to overexploitation. In the past,<br />
this overexploitation has been<br />
possible because of the laxity and<br />
weakness in enforcement of the<br />
Fisheries Act of 1991, which is<br />
highly explicit on the manner in<br />
which fi shing activities should be<br />
conducted. Signifi cant efforts are<br />
being made to address the challenge<br />
of providing a comprehensive,<br />
consistent set of policies and<br />
programs for sustainable<br />
management of the lake’s fi shery<br />
resources. For example, in March<br />
2007, Kenya, Tanzania, and Uganda<br />
adopted a Regional Plan of Action<br />
for the Management of Fishing<br />
Activity; this plan called on the<br />
respective governments to review<br />
their national policies and develop<br />
a harmonized fi shing framework<br />
(LVFO 2007; Lubovich 2009).<br />
Nevertheless, many challenges still<br />
lie ahead, and it will be critical to<br />
reinforce policy and management<br />
action with programmes of public<br />
awareness and education.
Species in the spotlight<br />
Forest remnants in western Africa –<br />
vanishing islands of sylvan fi shes<br />
A<br />
signifi cant part of western<br />
Africa is covered by<br />
differing types of savanna<br />
that are drained by a few<br />
large rivers, like the Niger, Volta<br />
and the Senegal. The vegetation<br />
refl ects climatic conditions<br />
including a cycle of dry and wet<br />
seasons. Closer to the coast, partly<br />
bordered by the Guinean highlands<br />
(from the highlands of the southern<br />
Fouta Djallon in south-eastern<br />
Guinea, through northern Sierra<br />
Leone and Liberia, to northwestern<br />
Côte d’Ivoire), the climate<br />
is more humid, allowing different<br />
types of forest to grow. These<br />
forests are inhabited by animals<br />
closely resembling or even identical<br />
to those of the central African<br />
forests. Thus, many sylvan (forest<br />
dwelling) fi sh species and speciesgroups<br />
fi nd their most westerly<br />
distributions within the western<br />
Africa coastal forests. A number of<br />
these westerly sub-populations may<br />
be discrete sub-species, or separate<br />
species within species complexes,<br />
showing distinct colour morphs, or<br />
other unique features.<br />
The high number of unconnected<br />
coastal rivers in western Africa is<br />
thought to have promoted these<br />
speciation processes, which have<br />
led to noticeably high levels of<br />
endemism, such as for characids,<br />
barbs and cichlids. Furthermore,<br />
several remarkable fi shes, which<br />
are sometimes called ‘relict‘ species,<br />
occur in the Guinean regions.<br />
These species belong to phyletically<br />
old groups previously represented<br />
by more numerous and widespread<br />
species but, following evolutionary<br />
events, now represented by only<br />
a few, often locally restricted<br />
species. Examples include the<br />
fourspine leaffi sh (Afrononandus<br />
sheljuzhkoi) and the African leaffi sh<br />
(Polycentropsis abbreviata), with<br />
their closest relatives in Asia and<br />
South America, and the enigmatic<br />
denticle herring (Denticeps<br />
clupeoides), which is the only<br />
extant representative of the family<br />
Denticipetidae, the sister group of<br />
all other clupeomorphs.<br />
The climatic conditions of the<br />
Guinean region not only provide<br />
good conditions for forest<br />
ecosystems, but also support a more<br />
diverse and reliable agriculture<br />
compared with the Sahelo-Sudan<br />
region. This promotes better<br />
livelihood opportunities which, in<br />
turn, lead to increased population<br />
densities and a greater demand for<br />
land. With increased demands for<br />
agricultural land, deforestation<br />
continues, leaving only forest<br />
fragments in some areas.<br />
1 Deutsches Meeresmuseum, Museum für Meereskunde und Fischerei – Aquarium, Stralsund, Germany<br />
Moritz, T. ¹<br />
Lokoli swamp forest in southern Benin. This small forest fragment serves as one of<br />
the last refuges for many forest dwelling species. © T. MORITZ<br />
Lokoli forest – a refuge<br />
An exemplary forest remnant<br />
is the Lokoli swamp forest in<br />
southern Benin. This small,<br />
(approximately 500ha) piece of<br />
forest is permanently fl ooded by a<br />
network of channels from the Hlan<br />
River, an affl uent of the Ouémé<br />
River. It is approximately 20km<br />
east of Bohicon and 100km north<br />
of Cotonou and can only be crossed<br />
by boat. The forest is densely<br />
vegetated with high tree density,<br />
and the tree cover is usually closed<br />
above the channels. Most channels<br />
are less than a metre wide and are<br />
only navigable using small dugout<br />
canoes. Water depth varies by less<br />
than one metre within a year, and is<br />
usually around 1 to 2.5 metres. The<br />
water has a dark brown colouration<br />
due to leaf litter decomposition, a<br />
moderate acidic pH of 6 to 7 and a<br />
temperature of around 26°C. The<br />
channel substrate is predominantly<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
79
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
80<br />
Barboides britzi, one of Africa’s smallest freshwater fi sh, is a newly described species endemic to Lokoli forest in Benin.<br />
© T. MORITZ<br />
sand, with small patches of gravel<br />
where the current is stronger; most<br />
places have a mud and leaf litter<br />
layer of variable depth.<br />
The Lokoli forest serves as one of<br />
the last refuges for forest dwelling<br />
animals in the Dahomey Gap,<br />
including pangolins, fl ying squirrels<br />
and the red-bellied monkey<br />
(Cercopithecus erythrogaster),<br />
which is endemic to Benin. While<br />
herpetological surveys have shown<br />
relatively few exclusive forest<br />
species of reptiles and amphibians<br />
(Rödel et al. 2007; Ullenbruch et<br />
al. 2010), the situation for fi shes<br />
is very different. Despite a direct<br />
connection to the main channel<br />
of the Ouémé River, fi shes of the<br />
Lokoli are, to a high degree, typical<br />
forest species, otherwise known<br />
from the coastal forested rivers of<br />
the Niger Delta and the connected<br />
network of lagoons parallel to the<br />
coast. The reedfi sh (Erpetoichthys<br />
calabaricus), butterfl y fi sh (Pantodon<br />
buchholzi), and elephant nose fi sh<br />
(Gnathonemus petersii) in Lokoli<br />
are at the most western points of<br />
their ranges (Montchowui et al.<br />
2007; pers. obs.). The cyprinid genus<br />
Barboides, consisting of two of the<br />
smallest African freshwater species,<br />
is also at the most westerly point<br />
of its range, with B. britzi endemic<br />
to the Lokoli forest itself. This<br />
miniature fi sh becomes sexually<br />
mature at a smaller size than any<br />
other freshwater fi sh in Africa, at<br />
12.6mm standard length (Conway<br />
and Moritz 2006). It is likely that<br />
more fi sh species, especially of<br />
smaller body size, await discovery<br />
in such unusual habitats. For<br />
example, the bottom dwelling<br />
distichodontid Nannocharax signifer<br />
was only recently described, from a<br />
small affl uent of the Lokoli forest<br />
(Moritz 2009).<br />
Impacts to this small forest<br />
fragment are signifi cant, with<br />
extensive clearance for agriculture<br />
along the forest margins. Within<br />
the forest itself, despite religious<br />
taboos prescribing at least some<br />
regulations for hunting, the bush<br />
meat trade remains an important<br />
source of income and bush meat<br />
is openly sold along the main road<br />
leading to Cotonou. Palm wine and<br />
secondary products are produced<br />
by cutting off the tops of the<br />
palm Raphia hookeri, with evident<br />
impacts to the plants themselves.<br />
As a result, the abundance of this<br />
formerly dominant palm has been<br />
signifi cantly reduced through<br />
over-harvesting. The forest fl ora<br />
has been further impacted through<br />
introduction of alien species such<br />
as the taro (Colocasia esculenta), an<br />
introduced plant valued for its root<br />
tubers, which is widely planted,<br />
even within clearings in the swamp<br />
forest.<br />
Forested coastal rivers, although<br />
more spacious than some of the<br />
Guinean forested rivers, face similar<br />
threats. In addition to pollution,<br />
which is heavily impacting certain<br />
areas, the primary problem is, once<br />
more, habitat degradation due to<br />
expanding agriculture. The Iguidi<br />
River at the border of Benin and<br />
Nigeria provides a good example.<br />
The course of this small coastal<br />
river is clearly visible on aerial or
satellite images, due to its bordering<br />
gallery forest. The forest stands out<br />
in stark contrast to neighbouring,<br />
and continuously expanding, fi elds.<br />
The Iguidi River fl ows in a northsouth<br />
direction, starting out as a<br />
small forested stream that develops<br />
into a swamp. As is typical for a<br />
forest stream, the water is brown to<br />
dark brown in colour, although the<br />
pH is not especially acidic, at 6.5 to<br />
7.5; water temperature is commonly<br />
26 to 29°C; and conductivity is<br />
low at 50 to 65µS (Moritz 2010).<br />
Despite the river’s low salt content,<br />
fi shes characteristic of brackish<br />
environments are also present,<br />
such as the freshwater pipefi sh of<br />
the genus Enneacampus, and the<br />
sleeper goby (Eleotris daganensis).<br />
The majority of fi shes from the<br />
Iguidi are, however, typically<br />
freshwater, forest-dwelling<br />
species such as the dotted catfi sh<br />
(Parauchenoglanis monkei), the<br />
small distichodontid, Neolebias<br />
ansorgii, and the cryptic mormyrid<br />
(Isichthys henryi). This small river<br />
represents an outpost of the Lower<br />
Guinean forest, and holds the most<br />
westerly distributions of several<br />
Lower Guinean species, such as the<br />
aforementioned Neolebias ansorgii,<br />
the Niger tetra (Arnoldichthys<br />
spilopterus), and the catfi sh<br />
Schilbe brevianalis (Moritz 2010).<br />
Furthermore, the Iguidi River is the<br />
type locality for the rare, miniature<br />
Barbus sylvaticus, the even smaller<br />
Barboides gracilis, and the denticle<br />
herring (Denticeps clupeoides), all of<br />
which are assessed as Vulnerable or<br />
Endangered.<br />
In conclusion, at fi rst glance,<br />
small forest fragments seem to<br />
be of minor importance for the<br />
conservation of forest dwelling<br />
species – often being too small<br />
to sustain endemic species, or<br />
even too small to harbour a<br />
discrete population of a sylvan<br />
species. Many inhabitants of<br />
forest remnants are, therefore,<br />
non-specialist or even savanna<br />
species. A closer view of the fi shes,<br />
however, reveals a quite different<br />
picture. Forest remnants, such as<br />
the Lokoli, can sustain a number<br />
of small endemic species. What<br />
is more important, however, is<br />
the complexity of biodiversity<br />
that is found and that needs to be<br />
conserved. Forest fragments and<br />
remnants of gallery forests are focal<br />
points of habitat complexity, edge<br />
effects and ecological interactions<br />
– and as outposts for species<br />
distributions, they may be of high<br />
importance for maintaining genetic<br />
variability within a species and in<br />
ongoing evolutionary processes.<br />
Therefore, despite their small size,<br />
forest fragments deserve greater<br />
focus within conservation plans;<br />
their inclusion will help to ensure<br />
preservation of biodiversity in all<br />
its forms.<br />
The freshwater butterfl yfi sh, Pantodon buchholzi (LC), a widespread species in<br />
Africa, reaches the most westerly point of its range in Lokoli forest, Benin. This<br />
species is capable of jumping out of the water to search for insects or to escape<br />
from predators. It is not a glider, but a ballistic jumper, with tremendous jumping<br />
power. © T. MORITZ<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
81
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
82<br />
Species in the spotlight<br />
A unique species fl ock in Lake Tana –<br />
the Labeobarbus complex<br />
Lake Tana, in Ethiopia, and<br />
the rivers that drain into<br />
it, are home to a unique,<br />
endemic species fl ock<br />
belonging to the cyprinid genus<br />
Labeobarbus. The lake, which has<br />
a surface area of 3,150km 2 , is the<br />
largest in Ethiopia. It is situated<br />
in the north-western highlands<br />
at an altitude of approximately<br />
1,800m. It was formed during the<br />
early Pleistocene when a 50kmlong<br />
basalt fl ow blocked the course<br />
of the Blue Nile near its source<br />
(Mohr 1962; Chorowicz et al. 1998).<br />
Today, several rivers drain into<br />
Lake Tana, which itself forms the<br />
headwaters of the Blue Nile – the<br />
only river fl owing out of the lake,<br />
contributing more than 80% of the<br />
total volume of the Nile River at<br />
Khartoum, Sudan.<br />
The wetlands and fl oodplains that<br />
surround most of the lake form the<br />
largest wetland area in Ethiopia and<br />
are an integral part of the complex<br />
Tana ecosystem. The wetlands to<br />
the east of the lake serve as breeding<br />
grounds for Oreochromis niloticus<br />
(Nile tilapia) and Clarias gariepinus<br />
(North African catfi sh), both of<br />
which are important for the lake<br />
fi sheries (Vijverberg et al. 2009).<br />
There are 28 species of fi sh in<br />
Lake Tana, of which 20 are endemic<br />
to the lake and its catchments<br />
(Vijverberg et al. 2009). The fi sh<br />
fauna includes representatives of<br />
the genera Oreochromis, Clarias,<br />
Labeobarbus (i.e., the ‘large African<br />
barbs’), Barbus (i.e., the ‘small Barbus<br />
group’; see De Weirdt and Teugels<br />
2007), Garra, Varicorhinus and<br />
Nemacheilus. The population of O.<br />
niloticus in Lake Tana was described<br />
as a separate sub-species, Oreochromis<br />
1 Department of Biology, Addis Ababa University, Ethiopia<br />
niloticus tana. Two exotic species,<br />
Gambusia holbrooki and Esox lucius,<br />
were reported to have been brought<br />
from Italy during the late 1930s and<br />
introduced into the lake (Tedla and<br />
Meskel 1981); there is, however, no<br />
trace of these fi shes from the lake in<br />
recent times.<br />
The Labeobarbus species fl ock<br />
The Cyprinidae are the most<br />
species-rich family in the lake,<br />
represented by four genera, Barbus,<br />
Garra, Labeobarbus and Varicorhinus.<br />
Within the Labeobarbus is a unique<br />
complex of 17 species (Getahun<br />
and Dejen in prep.). It is thought<br />
that the lake is able to support such<br />
a large number of closely related<br />
species because, when it fi rst formed,<br />
it offered several new habitats<br />
that may have promoted adaptive<br />
radiation among the original<br />
colonising species, and it has since<br />
remained isolated due to the Tissisat<br />
Falls, located 30km downstream<br />
from the outfl ow of the lake. Most<br />
interesting is the speed of evolution<br />
for so many new species, as historical<br />
evidence suggests the lake dried<br />
out completely as recently as 16,000<br />
Getahun, A.¹<br />
Labeobarbus macrophtalmus is a benthopelagic species that forms an important<br />
component of the Lake Tana fi shery. © LEO NAGELKERKE<br />
years ago (Lamb et al. 2007), meaning<br />
the evolution of the Labeobarbus<br />
species complex may have taken<br />
fewer than 15,000 years (Vijverberg<br />
et al. 2009).<br />
Eight of the Labeobarbus species<br />
are piscivores, and most of them<br />
periodically migrate into the rivers<br />
for spawning. L. intermedius and L.<br />
tsanensis are abundant in the inshore<br />
habitats and are the predominant<br />
species at the river mouths. L.<br />
tsanensis and L. brevicephalus are the<br />
dominant species offshore.<br />
Spawning behaviour<br />
Limited surveys around Lake Tana<br />
indicate that the Ribb, Megech and<br />
Dirma Rivers and their tributaries<br />
provide ideal breeding grounds<br />
for these species in the northern<br />
and eastern parts of the lake. Five<br />
species were found to migrate from<br />
Lake Tana up both the Megech and<br />
Dirma rivers to spawn (Anteneh<br />
2005), although slightly greater<br />
numbers migrate up the Megech,<br />
which has more tributaries with<br />
gravel beds, and a slightly higher<br />
dissolved oxygen content. Three<br />
categories of spawning behaviour are
observed (Anteneh 2005), obligate<br />
river spawners, lake spawners and<br />
generalists (spawning in both the<br />
lake and its tributary rivers).<br />
At least seven species spawn in<br />
the headwaters of the main rivers<br />
draining to the lake. As yet, there is<br />
no evidence of river-specifi city, but<br />
this cannot be discounted. After a<br />
brief pre-spawning aggregation at<br />
the river mouths, the adults migrate<br />
upstream in July and August, at<br />
the onset of the rainy season. Final<br />
maturation and spawning occur in<br />
the tributaries of the major rivers, or<br />
possibly in gravel reaches in the main<br />
channels. After spawning, the adults<br />
return to the lake for feeding until<br />
the next cycle of breeding. Highly<br />
oxygenated water and gravel beds are<br />
important for development of the<br />
eggs and larvae. Deposition of eggs<br />
in gravel beds prevents them from<br />
being washed away, and clear water<br />
is required to ensure they are free of<br />
sediments that might obstruct the<br />
diffusion of oxygen.<br />
The juveniles start to return to<br />
the lake in September and October<br />
as fl ows reduce, where they feed<br />
and grow to sexual maturity. There<br />
is good evidence that, during their<br />
return to the lake, the juveniles may<br />
remain in the pools of the main river<br />
segments for an extended period,<br />
probably until the next rainy season,<br />
at which time they will be carried<br />
into the lake.<br />
The lake fi sheries<br />
The lake fi shery is clearly very<br />
important to the local population,<br />
employing more than 3,000 people<br />
in fi shing, marketing, and processing<br />
(Anteneh 2005). Traditionally, the<br />
main fi shery has been a subsistence<br />
reed boat fi shery targeting a range<br />
of species, sometimes including<br />
the Labeobarbus species. This was<br />
conducted throughout the lake<br />
until the 1980s; since then it has<br />
been replaced in many areas by<br />
other methods. The fi shery remains<br />
important in the more remote areas<br />
of the lake, with the catch being<br />
sold at small markets or used for<br />
household consumption. It mainly<br />
employs gillnets, and the main target<br />
species is Nile tilapia (O. niloticus).<br />
However, the reed boat (tankwa)<br />
fi shermen also use hooks and lines,<br />
and traps, as well as spears to catch<br />
catfi sh.<br />
In 1986, motorised boats and<br />
nylon gill nets were introduced<br />
as part of the Lake Tana Fisheries<br />
Resource Development Program<br />
(LTFRDP) (Anteneh 2005). Data<br />
collected from all commercial<br />
fi sheries recognizes only four<br />
species groups: Labeobarbus spp.,<br />
African catfi sh (C. gariepinus),<br />
Nile tilapia (O. niloticus) and beso<br />
(Varocorhinus beso). This fi shery<br />
mainly supplies larger markets,<br />
using 100m long gillnets. There are<br />
around 25 motorised fi shing boats,<br />
most of which land their catch in<br />
Bahir Dar, the main town on the<br />
shore of Lake Tana. The fi shery<br />
is, however, expanding to all 10<br />
Woredas (districts) bordering the<br />
lake, including the Gorgora area (on<br />
the northern shore).<br />
Total annual catches increased<br />
from 39 tonnes in 1987 to 360 tonnes<br />
in 1997 (Wudneh 1998). However,<br />
the catch per unit effort for the<br />
commercial gill net fi shery targeting<br />
Labeobarbus species dropped by<br />
more than 50% over the period<br />
1991 to 2001 (de Graff et al. 2004).<br />
The same authors have reported a<br />
75% decline (in biomass) and 80%<br />
(in number) of landed fi sh of the<br />
species of Labeobarbus (L. acutirostris,<br />
L. brevicephalus, L. intermedius, L.<br />
macrophthalmus, L. platydorsus and<br />
L. tsanensis) in the southern gulf<br />
of Lake Tana. The most plausible<br />
explanation for the decline is<br />
recruitment overfi shing by the<br />
commercial gillnet fi shery (de Graff<br />
et al. 2004), and poisoning of the<br />
spawning stock in rivers using the<br />
crushed seeds of ‘birbira’ (Milletia<br />
ferruginea) (Nagelkerke and Sibbing<br />
1996; Ameha 2004).<br />
The commercial gill net fi shery<br />
for species of Labeobarbus is<br />
highly seasonal and mainly targets<br />
spawning aggregations, as more<br />
than 50% of the annual catch is<br />
obtained in the river mouths during<br />
August and September. There is<br />
also a chase and trap fi shery based<br />
in the southern part of the lake, and<br />
longlines, cast nets and traps are<br />
occasionally used but contribute<br />
little to the total fi sh catch.<br />
Threats to Lake Tana and its<br />
Labeobarbus species<br />
Overfi shing<br />
Although a fi shery policy has been<br />
developed both at federal and<br />
regional levels, it is not effectively<br />
Fishermen cast their lines from papyrus boats ("tankwas") on Lake Tana in northern<br />
Ethiopia. Behind them lies the source of the Blue Nile. Near Bahir Dar, Ethiopia.<br />
© A. DAVEY<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
83
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
84<br />
implemented. Lakes and rivers<br />
are, unoffi cially, considered to be<br />
resources that are freely available<br />
to everyone. There are still many<br />
illegal, unregistered fi shermen<br />
exploiting the fi sh resources, and<br />
there is little regulation of fi shing<br />
gears. As reported above, this has<br />
led to overfi shing of Labeobarbus in<br />
some parts of the lake, especially in<br />
the south around the town of<br />
Bahir Dar.<br />
Habitat disturbance<br />
As seasonal fl ooding recedes, many<br />
people use the shores of the lake for<br />
‘fl oodplain recession agriculture’.<br />
Human encroachment on the<br />
wetlands increases every year,<br />
with the subsequent depletion of<br />
emergent macrophytes through<br />
harvesting and burning, while<br />
there is an expansion of submerged<br />
macrophyte stands in other areas.<br />
Over the last 15 years,<br />
deforestation has become very<br />
widespread, facilitating conditions<br />
for soil erosion, resulting in<br />
sediments draining into the lake<br />
and smothering upstream spawning<br />
areas. The soil loss rate from areas<br />
around the lake is between 31 and 50<br />
tonnes per hectare per year (Teshale<br />
et al. 2001; Teshale 2003). These<br />
huge deposits of sediment into the<br />
lake have led to a reduction in the<br />
lake’s area, a drop in water levels,<br />
and a loss of water holding capacity.<br />
This reduction in the water level<br />
has resulted in fragmentation of the<br />
available aquatic habitat, especially<br />
around shores. Some of the exposed<br />
land is now used for cultivation and<br />
excavation of sand.<br />
Water pollution<br />
Run-off from small-scale agriculture<br />
around the lake is bringing<br />
agricultural fertilizers, pesticides<br />
(including DDT), and herbicides<br />
into the lake. The use of these<br />
agricultural products by farmers<br />
is still relatively limited; however,<br />
a lack of effective regulation on<br />
their use presents a potential threat<br />
to water quality in the lake. Other<br />
chemicals, such as ‘birbira’ (Milletia<br />
Many of the Labeobarbus species migrate up the rivers fl owing into Lake Tana to<br />
spawn in gravel beds, such as seen here in the Gumara River. © LEO NAGELKERKE<br />
ferruginia) seed powder (used as an<br />
ichthyocide; see above), may also<br />
pollute the lake and kill the aquatic<br />
fauna, including Labeobarbus species.<br />
Domestic waste water from the<br />
town of Bahir Dar is, in most cases,<br />
discharged directly into Lake Tana –<br />
the development of an appropriate<br />
sewage system could solve or<br />
mitigate these pollution threats.<br />
Water abstraction and<br />
impoundment<br />
Water abstraction occurs at some<br />
points around the lake as a result of<br />
privately run, small-scale irrigation<br />
projects. However, because the<br />
Lake Tana and Beles sub-catchment<br />
is considered a growth corridor<br />
by the federal and regional<br />
governments, there are several other<br />
dam and irrigation projects under<br />
consideration or being implemented.<br />
These include the Tana Beles interbasin<br />
water transfer project, and the<br />
Koga, Ribb, Megech, Gilgel Abay<br />
and Gumara dams and irrigation<br />
projects. Some of these are intended<br />
to impound the lake’s tributaries to<br />
store water; some to pump water<br />
through tunnels from the lake<br />
to a hydropower facility before<br />
discharging the water into the Beles<br />
River; and some to pump water<br />
directly from the lake for irrigation<br />
purposes. These projects may lower<br />
the water level and quality in Lake<br />
Tana and its tributaries, with<br />
subsequent impacts to biodiversity.<br />
As reported above, many species<br />
of Labeobarbus undergo spawning<br />
migrations that, without effective<br />
measures to allow passage past<br />
newly constructed dams, may be<br />
blocked, potentially leading to the<br />
extinction of this unique fl ock of<br />
cyprinids. Environmental impact<br />
assessment (EIA) studies have<br />
been conducted for many of these<br />
projects, so it is hoped that the<br />
recommended mitigation measures<br />
and the management plans<br />
suggested will be strictly followed<br />
and implemented.<br />
Lack of information and<br />
institutional capacity<br />
Comprehensive scientifi c studies<br />
on the biology, behaviour, and<br />
ecology of the different species of<br />
Labeobarbus are still lacking. This<br />
makes it diffi cult to recommend<br />
mitigation measures in some of<br />
the EIA studies and follow up<br />
with implementation. In addition,<br />
the implementing agencies for<br />
EIAs still lack the strength and<br />
capacity to enforce and implement<br />
any recommendations made. The<br />
development of a Lake Tana subbasin<br />
authority is an option for<br />
solving this problem. Concerted<br />
action by all stakeholders is<br />
required if the unique fi sh fauna of<br />
this lake is to be conserved for the<br />
future.
The Twee redfi n, Barbus<br />
erubescens, a Critically Endangered<br />
fi sh from the Twee River, South<br />
Africa, where it is threatened by<br />
alien fi sh species. © D. IMPSON<br />
Species in the spotlight<br />
The Twee River redfi n – a Critically<br />
Endangered minnow from South Africa<br />
The Twee River redfi n<br />
(Barbus erubescens Skelton)<br />
was described in 1974,<br />
following an investigation<br />
that included extensive fi eld<br />
observations. The species is named<br />
for the bright reddish breeding<br />
dress assumed by spawning<br />
males, with females being less<br />
intensely coloured. The common<br />
name indicates that the species’<br />
distribution is restricted to one<br />
tributary system of the Olifants<br />
River in the Cedarberg Mountains<br />
of the Western Cape, South Africa.<br />
This tributary system includes the<br />
Twee and some of its affl uents, the<br />
Heks, Suurvlei and Middeldeur<br />
rivers.<br />
At the time of discovery, only<br />
one other fi sh species was known<br />
to be indigenous to the Twee River,<br />
and both species were isolated by<br />
a vertical waterfall of about 10m,<br />
located close to the confl uence<br />
of the Twee and Leeu rivers. This<br />
other indigenous fi sh is a species of<br />
South African Galaxias, formerly<br />
named as the Cape galaxias<br />
(Galaxias zebratus); however, more<br />
recently it has become evident<br />
that a number of populations of G.<br />
zebratus might represent distinct<br />
species. The population of Galaxias<br />
1 South African Institute for Aquatic Biodiversity, Private Bag 1015, Grahamstown 6140, South Africa<br />
Skelton, P.H.¹<br />
in the Twee River is one of these<br />
distinct populations. The Cape<br />
galaxias is currently assessed in the<br />
<strong>IUCN</strong> Red List as Data Defi cient,<br />
due to the taxonomic confusion<br />
associated with the species<br />
complex. Below the falls several<br />
other indigenous freshwater fi sh<br />
species are found, most of them<br />
endemic to the Olifants system.<br />
One of these species, Barbus calidus,<br />
is the sister species of B. erubescens<br />
(i.e., it is the phylogenetically<br />
most closely related species to<br />
B. erubescens). Barbus calidus, the<br />
Clanwilliam redfi n, itself classifi ed<br />
as Vulnerable, due to threats<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
85
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
86<br />
from invasive species, and habitat<br />
degradation caused by agriculture,<br />
is discussed below.<br />
Much has been learnt about the<br />
Twee River redfi n since its original<br />
description. In common with a<br />
disproportionately large number<br />
(80%) of barbine minnows from<br />
the temperate reaches of southern<br />
Africa, the Twee River redfi n is<br />
tetraploid (that is, it has four sets<br />
of each chromosome), with 100<br />
chromosomes in total. Its most<br />
distinctive external character is<br />
the high number of branched anal<br />
fi n branched rays – six or, more<br />
usually, seven – more than any<br />
other African barbine species. It has<br />
several other distinctive features,<br />
such as small scattered nuptial<br />
tubercles on both sexes, two pairs<br />
of well developed mouth barbels,<br />
and an unbranched ray in the dorsal<br />
fi n that shows either incipient or<br />
vestigial serrations.<br />
The species’ breeding behaviour<br />
features males congregating and<br />
forming a dense, swarming, nuptial<br />
school against rock surfaces to<br />
which individual breeding females<br />
are attracted and enticed to spawn<br />
over cobbles or rock crevices with<br />
several pursuant males. This occurs<br />
in spring or early summer (October<br />
to December) when streams are<br />
swollen by frontal rains. The<br />
species is a ‘broadcast spawner’<br />
(releasing the gametes into the<br />
water) and does not practice any<br />
form of parental care. It can live for<br />
up to fi ve or six years. The species<br />
feeds on drifting insects and other<br />
invertebrates or from rocks and<br />
other benthic surfaces.<br />
Conservation concerns<br />
When fi rst discovered, the species<br />
was common and widespread in<br />
the tributary system – with larger<br />
adults occupying open water<br />
habitats in pools and runs, and<br />
juveniles shoaling along marginal<br />
zones. Since the 1970s, the<br />
population has declined markedly<br />
and is absent from large sections of<br />
its former range. The reasons for<br />
this decline are several, including<br />
The Twee River in the Cedarburg Mountains, the Western Cape, South Africa.<br />
© SAIAB/P. SKELTON<br />
likely impacts from agricultural<br />
developments (riparian fruit<br />
orchards) impacting both water<br />
quality and quantity, and alien<br />
invasive fi sh species. The fi rst<br />
alien fi sh species to be recorded<br />
was a South African anabantid,<br />
the Cape kurper (Sandelia capensis)<br />
which, although not a large fi sh,<br />
is widespread throughout most<br />
of the tributary and an avid<br />
predator on small fi shes and<br />
invertebrates. The Clanwilliam<br />
yellowfi sh (Labeobarbus capensis),<br />
a large cyprinid of the Olifants<br />
River system, was introduced to<br />
the Twee River above the barrier<br />
waterfall by Nature Conservation<br />
authorities seeking to conserve<br />
that species in the face of threats<br />
from other introduced species! The<br />
Clanwilliam yellowfi sh is found<br />
mainly in the downstream reaches<br />
of the Twee and, although its<br />
precise impact is not known, it is<br />
a predator and grows much larger<br />
that the Twee River redfi n. Bluegill<br />
sunfi sh (Lepomis macrochirus),<br />
a North American centrarchid<br />
species, and another predator on<br />
small fi shes and invertebrates, have<br />
also invaded the system. Rainbow<br />
trout (Oncorhynchus mykiss) have<br />
been recorded from the Twee River<br />
but are not common.<br />
The Twee River has been<br />
extensively surveyed on several<br />
occasions to determine the<br />
conservation status of the redfi n<br />
and the Galaxias species. The<br />
decline in their populations is of<br />
great concern, as the tributary<br />
system is restricted in size and<br />
subject to increasing agricultural<br />
pressures as well as the invading<br />
alien species. There are few natural<br />
sanctuary reaches and, unless<br />
determined action to remove the<br />
alien species is taken, the fate of<br />
the threatened indigenous species<br />
might be sealed forever. Two things<br />
are essential for conservation<br />
action – political will by the<br />
authorities to do what they must<br />
in the face of contrary perceptions<br />
by the public (who, for example,<br />
may support introductions of<br />
species for fi shing), and a properly<br />
informed public, especially the<br />
local landowning public. If those<br />
elements are in place, the survival<br />
of these and other indigenous<br />
species in South Africa might<br />
be secured.
Species in the spotlight<br />
Cauldrons for fi sh biodiversity:<br />
western Africa’s crater lakes<br />
Globally, crater lakes<br />
are comparatively<br />
rare, usually small and<br />
specialised freshwater<br />
habitats formed in geological<br />
depressions, such as the Ojos del<br />
Salado in the Andes mountains,<br />
bordering Argentina and Chile<br />
– probably the highest altitude<br />
permanent lake of any description<br />
(68º32′W, 27º07’S, elevation 6,390m,<br />
diameter 100m, depth perhaps<br />
5 to 10m). Crater lakes are well<br />
represented in tropical Africa,<br />
especially in the Guinean rainforest<br />
zone of Cameroon, where there<br />
may be 36 or more. The entire<br />
region is a celebrated ‘biodiversity<br />
hotspot’ for both lacustrine and<br />
riverine fi shes (Reid 1989; 1996;<br />
Teugels et al. 1992; Schliewen 2005;<br />
Stiassny et al. 2007). Contemporary<br />
general studies on the world’s<br />
crater lakes address important<br />
topics such as: lake formation;<br />
physical, chemical, geological,<br />
geographical and biological<br />
evolution; paleoecology; historical<br />
biotic colonisation; and recent<br />
ecology – including the assessment<br />
of conservation status and threats<br />
to the survival of the contained<br />
habitats and species. The potential<br />
for (and impacts from) human use<br />
is studied, including water supply,<br />
agriculture, fi sheries and also<br />
recreation and ecotourism – such<br />
lakes often being scenic locations.<br />
Crater lakes everywhere may<br />
contain a substantial number<br />
of endemic fi shes and other<br />
aquatic and amphibious taxa.<br />
Among African fi shes endemic to<br />
craters, small phyletic and trophic<br />
assemblages of species and genera<br />
representing the family Cichlidae<br />
have attracted much international<br />
scientifi c attention. Crater lake<br />
cichlids, their taxonomy, phylogeny<br />
and ecology were documented early<br />
on in Cameroon, notably in Lake<br />
Barombi Mbo (Trewavas 1962;<br />
Trewavas et al. 1972; see below); and<br />
they continue to be discovered – for<br />
example, the recently documented<br />
‘fl ock’ of eight new species of Tilapia<br />
from Lake Bermin or Beme (5°9’N,<br />
9°38’E; diameter around 700m, depth<br />
around 16m, and age probably far<br />
less than 1 million years) (Stiassny<br />
et al. 2002; Schliewen 2005). Such<br />
Cameroonian assemblages are often<br />
regarded as small-scale tilapiine<br />
counterparts to the better known<br />
large haplochromine and other<br />
cichlid ‘species fl ocks’ of the East<br />
African Great Lakes (Klett and<br />
Meyer 2002; Salzburger and Meyer<br />
2004).<br />
Formation<br />
Whatever the location, all craters<br />
on earth are formed either by<br />
impact of extraterrestrial bodies or<br />
1 North of England Zoological Society, Caughall Road, Upton, Chester CH2 1LH, UK<br />
McGregor Reid, G.¹ and Gibson, C.¹<br />
Stomatepia mongo, a Critically Endangered cichlid endemic to Lake Brombi Mbo,<br />
Cameroon. © OLIVER LUCANUS/BELOWWATER.COM<br />
by vulcanism (Decker and Decker<br />
1997; Sigurösson 1999). They are<br />
often visible in photographic, radar<br />
and other imagery taken from space<br />
(Hamilton 2001).<br />
Impact crater lakes<br />
The impact of a meteorite, asteroid<br />
or comet creates a depression. This<br />
can be a simple bowl (depth to<br />
diameter ratio typically 1:5 to 1:7)<br />
or a larger, shallower, more complex<br />
depression (depth to diameter<br />
ratio 1:10 to 1:20) sometimes<br />
incorporating a central island or<br />
islands. Such islands are caused<br />
by a gravitational collapse of the<br />
rim and a rebound of material<br />
to the centre, analogous to the<br />
splash effect seen when raindrops<br />
hit water. An island may itself<br />
incorporate a hollow that later<br />
forms a ’lake within a lake’, as in<br />
Lake Taal, Philippines (Reid pers.<br />
obs.). In geological terms, impact<br />
depressions occur frequently but<br />
are often temporary, and only<br />
some 120 are currently known<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
87
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
88<br />
Lake Barombi Mbo, Cameroon. This lake is considered to be the oldest radiocarbon-dated crater lake in Africa. © U. SCHLIEWEN<br />
worldwide, most commonly<br />
from North America, Europe<br />
and Australia. Their occasional<br />
occurrence in Africa is therefore of<br />
considerable scientifi c interest. It is<br />
postulated that multiple terrestrial<br />
impacts, particularly large ones, are<br />
of importance in both geological<br />
and biological terms and are likely<br />
associated with periodic species<br />
extinction events on land and in<br />
the marine environment occurring<br />
since at least the Cretaceous period<br />
(around 60 million years ago). The<br />
nature, persistence and effects<br />
of impact depressions depend on<br />
the ‘target’ substrate, the velocity<br />
of the impactor, its composition<br />
and identifying ‘signature’ – the<br />
physical and chemical outputs,<br />
such as meteorite shards, shock<br />
metamorphism, ‘rock melt’ and<br />
silica rich glasses. All of this may<br />
become biotically signifi cant at<br />
some later stage of lake evolution.<br />
Other factors determining<br />
nature and persistence include<br />
the location, scale and form of<br />
the depression, and subsequent<br />
chemical, geological, geographical<br />
and biological processes including<br />
any underlying volcanic activity,<br />
erosion, deposition of sediments<br />
and ecological colonisation.<br />
Aorounga, in the Sahara Desert<br />
of northern Chad, contains a rare<br />
western Africa example of a large,<br />
ancient, much eroded impact<br />
crater (19°6’N, 19°15’E; diameter<br />
17km; age around 200 million<br />
years ago (Hamilton 2001)) which<br />
supports isolated temporary<br />
pools in rainy periods. Across the<br />
Sahelian region such pools may<br />
contain a remarkable density<br />
of life, albeit briefl y, including<br />
anacostracan crustaceans (‘fairy<br />
shrimps’) emerging from eggs<br />
resting in the sand since previous<br />
inundations of water; and anuran<br />
(frog and toad) tadpoles which<br />
appear ‘as if from nowhere’ (Reid<br />
pers. obs.). However, the craters<br />
are usually dry and contribute a<br />
fi ne diatomaceous lake substrate<br />
to dust storms generated within<br />
the Bodélé Depression and which,<br />
in winter, amount to an average of<br />
1,200,000 tonnes of dust per day<br />
carried for hundreds or thousands<br />
of kilometres (Todd et al. 2007).<br />
The Arounga crater is one of a local<br />
series, which may have been part<br />
of the more permanent and far<br />
more extensive ‘Mega Lake Chad’<br />
dating from the Pleistocene to<br />
Holocene periods (around 2 million<br />
years ago to 10,000 years ago) and<br />
persisting to some extent until a<br />
few thousand years ago. Lake Chad<br />
is now only 5% of its volume in<br />
the 1960s, mainly due to excessive<br />
human abstraction demands. The<br />
Mega Chad has been crucial in<br />
determining much of the large-scale<br />
aquatic and terrestrial patterns in<br />
historical and recent biogeography<br />
for western Africa and the Nilo-<br />
Sudan ichthyological province<br />
(Reid 1996).<br />
Lake Bosumtwi, Ghana is a better<br />
known, but still scarce, example of<br />
a comparatively young, permanent<br />
impact crater lake (06°32’ N,<br />
01°25’W; rim diameter 10.5km;<br />
maximum depth 75m; age 1.3 ± 0.2<br />
million years). The largest single<br />
natural lake in sub-Saharan western<br />
Africa, it lies over crystalline<br />
bedrock of the West African<br />
Shield and research indicates that<br />
sediments associated with Lake<br />
Bosumtwi have spread to the Ivory<br />
Coast and to oceanic deposits,<br />
nearby in the Gulf of Guinea<br />
(Hamilton 2001; Embassy of the<br />
Federal Republic of Germany 2011).<br />
Volcanic crater lakes.<br />
Craters formed through vulcanism,<br />
and their associated lakes, are<br />
sometimes divided into two<br />
classes: calderas which are deep<br />
inverted cones; and maars which<br />
are shallower with a low profi le.<br />
However, these distinctions are<br />
not always obvious, and the nature<br />
of the volcanic activity can be<br />
complex (Decker and Decker<br />
1997). The rocky rim is often<br />
created in a gaseous explosion<br />
when hot volcanic lava or magma<br />
in a subterranean chamber makes<br />
contact with groundwater.
By contrast, Lake Barombi<br />
Mbo is small (see above) and<br />
estimated to be biologically<br />
mature since about 25,000 to<br />
33,000 years ago; it is considered<br />
to be the oldest radiocarbondated<br />
crater lake in Africa<br />
Subsidence of materials creates a<br />
depression within the rim that may<br />
later fi ll with water. A diatreme<br />
often persists under the lake bed,<br />
that is, a pipe-like vertical volcanic<br />
vent that is fi lled with broken and<br />
cemented rock created by a single<br />
explosion. Such diatremes may<br />
remain active. Lake Nyos (around<br />
322km north-west of Yaoundé,<br />
Cameroon, close to the border with<br />
Nigeria) is an example of a simple<br />
maar lake, but a comparatively<br />
deep one (6°26′17″N, 010°17′56″E;<br />
1,091m above sea level; 2km long by<br />
1.2km wide; and 208m maximum<br />
depth). Lake Barombi Mbo in<br />
south-west Cameroon is formed in<br />
a caldera, albeit a fairly small one<br />
(4°39’46’’N, 9°23’52’’E; 303m above<br />
sea level; 2.15km wide; and around<br />
110m maximum depth) (Schliewen<br />
2005; Lebamba et al. 2010).<br />
Lake development<br />
Whether formed by impact or<br />
vulcanism, craters that persist<br />
anywhere may periodically or<br />
permanently fi ll up with water<br />
from snow, rainfall, groundwater,<br />
a captured drainage, spring or<br />
swamp or a larger inundation.<br />
Depending on water supply,<br />
drainage and evaporation, the<br />
lake may reach the lowest point<br />
on the rim and then overspill as a<br />
waterfall if the rim is high; or as a<br />
stream, if at the outset the rim is<br />
low or becomes water eroded. At a<br />
critical point of attrition there can<br />
be catastrophic breakout fl ooding.<br />
If the crater contains an active<br />
volcanic vent (see ’diatreme‘ above)<br />
the water will have an elevated<br />
temperature and be turbid and<br />
acidic from high concentrations<br />
of dissolved volcanic gases and<br />
distinctly green, or red-brown if<br />
iron rich. Gases include carbon<br />
dioxide (CO 2 ), sulfur dioxide (SO 2 ),<br />
hydrogen chloride (HCl) and<br />
hydrogen fl uoride (HF), which may<br />
persist in solution and are lethal to<br />
invertebrate and vertebrate life.<br />
Lake Nyos, with a diatreme some<br />
80km below the lake bed, is one of<br />
only three known contemporary<br />
‘exploding’ and periodically lethal<br />
lakes, all of which are African<br />
(the others being nearby Lake<br />
Monoun, Cameroon (5°35’N,<br />
10°35’E) and Lake Kivu, Rwanda).<br />
Nyos and Monoun are located<br />
within the Oku Volcanic Field<br />
near the northern boundary of the<br />
Cameroon Volcanic Line, a zone<br />
of volcanoes, maars, calderas and<br />
other tectonic activity that extends<br />
south-west to the large, inactive<br />
Mount Cameroon composite<br />
volcano (stratovolcano) and beyond<br />
to the island of Bioko in the Gulf<br />
of Guinea, which also contains an<br />
unexplored crater lake (Flesness,<br />
pers. comm.). Nyos has periodically<br />
been supersaturated with carbon<br />
dioxide (CO 2 , forming carbonic<br />
acid) leaching from the underlying<br />
magma and with a peak lake<br />
density of approximately 90 million<br />
tonnes of CO 2 . In 1986, there<br />
was a gaseous explosion, perhaps<br />
precipitated by an earthquake or<br />
landslide, releasing approximately<br />
1.6 million tonnes of CO 2 into<br />
the atmosphere. This killed some<br />
1,800 people, 3,500 livestock, and<br />
gas in solution presumably killed<br />
fi shes and other aquatic life.<br />
Degassing pipes were installed<br />
in 2001 to prevent a repetition of<br />
the catastrophe (Kling et al. 2005).<br />
Some 2,000 times larger than Nyos,<br />
Lake Kivu has also been found to<br />
be periodically supersaturated –<br />
with evidence for outgassing every<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
89
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
90<br />
thousand years or so. The general<br />
ability of crater lakes to store<br />
carbon dioxide at depth for long<br />
periods and also release it is clearly<br />
important when calculating lake<br />
stability, contemporary carbon<br />
sequestration and ‘footprints’ –<br />
and in determining the survival,<br />
ecology and evolution of fi shes and<br />
other aquatic animal populations.<br />
In the case of large mature<br />
impact craters and inactive or<br />
dormant volcano craters, the water<br />
normally becomes thermally and<br />
eventually ecologically stratifi ed.<br />
The deep, cold, dense, aphotic and<br />
anoxic water above the lake bed<br />
is usually quite separate from the<br />
warm, less dense, sunlit surface<br />
layers which support most of<br />
the animal and plant species and<br />
biomass. Lake surface waters<br />
down to around 40m are usually<br />
life supporting and fresh but<br />
can, in some instances, be saline.<br />
The clarity or transparency (and<br />
hence transmission of sunlight,<br />
level of photosynthetic activity<br />
and primary production) can be<br />
high, but this is determined by<br />
the nature of the crater rim soil<br />
and biota above the waterline<br />
(Elenga et al. 2004; Lebamba et al.<br />
2010), nutrients, water movements<br />
(including infl ows, outfl ows and<br />
overturns of thermal strata), and by<br />
other limnological processes. Some<br />
crater lakes are of considerable<br />
maturity and scale, for example,<br />
the Lake Toba caldera, Danau Toba,<br />
Indonesia was formed around<br />
70,000 years ago, with an area<br />
of over 1,000km². By contrast,<br />
Lake Barombi Mbo is small and<br />
estimated to be biologically mature<br />
since about 25,000 to 33,000 years<br />
ago; it is considered to be the oldest<br />
radiocarbon-dated crater lake in<br />
Africa (Elenga et al. 2004; Lebamba<br />
et al. 2010). The physical origin<br />
of the lake has been estimated<br />
as around 1 million years ago<br />
(Schliewen 2005). In any event,<br />
there is contemporary evidence<br />
that substantial permanent bodies<br />
of water can form very quickly in<br />
craters, for example, the lake that<br />
The craters represent a<br />
younger, less complex (if<br />
potentially more volatile)<br />
ecosystem – a ‘microcosm’ more<br />
easily studied than the East<br />
African Great Lakes<br />
developed post 1991, following<br />
the eruption of Mount Pinatubo,<br />
Philippines.<br />
Lake colonisation and the<br />
evolution of species. Western<br />
African and other small crater<br />
lakes have attracted the attention<br />
of evolutionary biologists and<br />
conservationists mainly because of<br />
their endemic cichlid fi shes and the<br />
natural and anthropogenic threats<br />
to their survival. The craters<br />
represent a younger, less complex<br />
(if potentially more volatile)<br />
ecosystem – a ‘microcosm’ more<br />
easily studied than the East African<br />
Great Lakes. Such craters provide<br />
an opportunity to investigate<br />
stages in ecological colonisation<br />
from an initially lifeless<br />
environment, and the processes<br />
of population differentiation<br />
and speciation. While invariably<br />
occupied by invertebrates, not<br />
all western African crater lakes<br />
contain fi shes and shrimps<br />
(Schliewen 2005). For those which<br />
contain cichlids, and which are<br />
geologically isolated, the question<br />
of how they came to occupy the<br />
crater is intriguing. In some cases,<br />
there are potentially testable<br />
hypotheses of natural migration<br />
through large-scale paleo-historical<br />
indundations of water, or via<br />
crater stream outfl ows (some still<br />
extant). Notions of paleo-historical<br />
introductions of fi shes or eggs by<br />
humans or birds are less credible<br />
and diffi cult, or impossible, to<br />
test scientifi cally. Setting such<br />
possibilities aside, western<br />
African models of tilapiine cichlid<br />
speciation or adaptive radiation are<br />
being tested against the classical<br />
grand-scale eastern African model<br />
(Klett and Meyer 2002; Salzburger<br />
and Meyer 2004; Seehausen 2006).<br />
Evidently, the evolution of<br />
species fl ocks is not invariably an<br />
enclosed, lacustrine phenomenon<br />
or confi ned to cichlid taxa.<br />
However, for Salzburger and<br />
Meyer (2004): ‘Species richness<br />
seems to be roughly correlated<br />
with the surface area, but not the<br />
age, of the lakes. We observe that<br />
the oldest lineages of a species<br />
fl ock of cichlids are often less<br />
species-rich and live in the open<br />
water or deepwater habitats.’ Based<br />
initially on Lake Victoria, the<br />
general eastern African hypothesis<br />
is that haplochromine and other<br />
cichlid taxa evolved into lacustrine<br />
species fl ocks numbering in the<br />
hundreds through a process of<br />
allopatric speciation, that is, one<br />
involving periodic geographical<br />
separation of populations. It<br />
was suggested that a regular rise<br />
and fall of waters in geological<br />
time created satellite lakes to<br />
isolate cichlid populations, which<br />
then differentiated ecologically,<br />
morphologically, behaviourally and<br />
genetically into distinct species.<br />
These isolates supposedly later<br />
returned to the main lake during<br />
high paleo-historical water levels<br />
,but by that time did not interbreed<br />
with their congeners.<br />
An alternative model is that<br />
species can arise as monophyletic<br />
fl ocks within the body of a lake<br />
without such total isolation,<br />
that is, through a process of<br />
sympatric speciation. In testing<br />
these competing (but not
necessarily mutually exclusive)<br />
models, Schliewen et al. (2001)<br />
conducted a ‘gene fl ow’ study<br />
within fi ve tilapiine morphs<br />
endemic to Lake Ejagham,<br />
western Cameroon (5°44’59”N,<br />
8°59’16”E; surface areas 0.49km 2 ;<br />
maximum depth around 18m<br />
(Schliewen 2005)). Comparisons<br />
with a closely related riverine<br />
outgroup of cichlids suggest that<br />
synapotypic colouration and<br />
‘differential ecological adaptations<br />
in combination with assortative<br />
mating could easily lead to<br />
speciation in sympatry’ (Schliewen<br />
et al. 2001). More generally, it<br />
is postulated that a dynamic<br />
network of gene exchange or<br />
hybridization among populations<br />
creates a process of ‘reticulate<br />
sympatric speciation’ among<br />
Cameroonian crater lake cichlids<br />
(Schliewen et al. 1994; Schliewen<br />
1996, 2005; Schliewen and Klee<br />
2004). Comparable empirical<br />
research on post-colonisation<br />
cichlids in a young crater lake in<br />
Nicaragua also supports the idea<br />
that sympatric endemic ‘morphs’<br />
of individual cichlid species may<br />
diversify rapidly (say, within a<br />
hundred years or generations) in<br />
ecology, morphology and genetics<br />
and this can be interpreted as<br />
‘incipient speciation’ (Elmer et al.<br />
2010). Again, this is postulated to<br />
be through disruptive selection,<br />
perhaps sexual selection, mediated<br />
by female mate choice.<br />
Conservation of crater lake fi shes.<br />
The phylogenetic and associated<br />
data on crater lake cichlid species<br />
fl ocks (above) are at different<br />
levels of generality and, among<br />
other criteria, important in<br />
the evaluation of conservation<br />
priorities (Stiassny and de Pinna<br />
1994). However, a paucity of<br />
well-worked and wide-ranging<br />
studies has until recently limited<br />
such contributions (Stiassny<br />
2002; Stiassny et al. 2002). Even<br />
so, western African crater lakes<br />
are included as an important<br />
biogeographic category within<br />
standard recognised freshwater<br />
ecoregions of the world and Africa<br />
(Thieme et al. 2005; Abell et al.<br />
2008).<br />
Thieme et al. (2005) designate<br />
closed basins and small lakes<br />
as a ‘major habitat type’, whose<br />
ultimate conservation status within<br />
most of the western African block<br />
of ecoregions is under threat ‘based<br />
on projected impacts from climate<br />
change, planned developments,<br />
and human population growth’.<br />
Recent research on pollen, biomes,<br />
forest succession and climate in<br />
Lake Barombi Mbo crater during<br />
the last 33,000 years or so suggests<br />
the persistence of a humid, dense,<br />
evergreen cum semi-deciduous<br />
forest: ’These forests display a<br />
mature character until ca 2800 cal<br />
yr BP then become of secondary<br />
type during the last millennium<br />
probably linked to increased<br />
human interferences [our emphasis]’<br />
(Lebamba et al. 2010).<br />
The recent conservation status<br />
of small Cameroonian crater<br />
lakes, including Barombi Mbo,<br />
and their endemic fi shes and<br />
invertebrates, is considered in<br />
detail by Reid (1989, 1990a,b, 1995,<br />
1996) and Schliewen (1996, 2005).<br />
Such unique lake environments<br />
and endemic species are clearly<br />
of national and international<br />
importance. There is, from the<br />
outset, an inherent vulnerability<br />
of these ecosystems resulting<br />
from the geological instability in<br />
craters; their small physical size;<br />
the small size of the contained<br />
populations and their genetic<br />
isolation; and, for cichlid fi shes,<br />
their methods of reproduction<br />
and limited fecundity. Actual or<br />
potential general threats are widely<br />
familiar, including: overfi shing<br />
and other socio-economic factors,<br />
including pressure from external<br />
visiting tourists; the introduction<br />
of alien species (for example,<br />
crustaceans and fi shes (Slootweg<br />
1989)); siltation and a reduction or<br />
loss of allochthonous food supply<br />
of terrestrial plant material and<br />
invertebrates (both resulting from<br />
deforestation and slash and burn<br />
agriculture within the crater rim);<br />
adverse water level fl uctuation<br />
(from damming the lake outfl ow<br />
and from excessive abstraction);<br />
and water pollution (from natural<br />
volcanic gases, from aerial and<br />
industrial emissions travelling<br />
from a distance; and from locally<br />
applied agrochemicals, pesticides<br />
and ichthyotoxic molluscicides<br />
used to control the aquatic snail<br />
vectors of human schistosomiasis,<br />
at least endemic in Barombi Mbo).<br />
Among conservation<br />
recommendations that have<br />
been proposed by the authors<br />
(above) are: systematic Population<br />
and Habitat Viability Analyses,<br />
as formulated by the <strong>IUCN</strong><br />
Conservation Breeding Specialist<br />
Group; Red List threat assessments<br />
(as summarized in this volume);<br />
the formal designation of lakes as<br />
legally and practically protected<br />
aquatic nature reserves of national<br />
and international importance, with<br />
an accompanying conservation<br />
action plan (Lakes Barombi Mbo<br />
and Ejagham have now been<br />
designated as forest reserves<br />
(Schliewen 2005)); and ex situ<br />
programmes for the conservation<br />
breeding of species at risk,<br />
with the prospect of eventual<br />
reintroduction in appropriate<br />
circumstances (such ex situ<br />
aquarium breeding programmes<br />
have been in operation since 1999<br />
through European and North<br />
American Fish Taxon Advisory<br />
Groups). Despite the persistent<br />
threats outlined above, a survey of<br />
Lake Barombi Mbo in 2002 found<br />
all fi sh species to still be present<br />
(Schliewen 2005). However,<br />
many of the species present<br />
are threatened (even Critically<br />
Endangered), but there have been<br />
no recorded fi sh or invertebrate<br />
population declines to the point<br />
of extinction in any of the crater<br />
lakes. Nevertheless, continued<br />
vigilance, conservation monitoring,<br />
threat assessment, mitigation and<br />
protective measures remain highly<br />
appropriate.<br />
CHAPTER 3 | SPECIES IN THE SPOTLIGHT<br />
91