Toothless predators: feeding performance and techniques among the piscivores within Lake Tana's endemic barbus species flock (Pisces:Cyprinidae)

With more than 2000 fish species the Cyprinidae is the largest family of vertebrates. Lake Tana, a large lake (3050 km²) situated in the NW‐ highlands of Ethiopia, harbours, as far as we know the only remaining intact species flock of large (max. 100 cm FL) cyprinid fishes (15 Barbus spp.). One of the most intriguing aspects of this endemic Barbus species flock is the large number of piscivores (8). Cyprinid fishes seem not well designed for piscivory, they lack teeth in the oral jaw, have a small slit‐shaped pharyngeal cavity and all lack a stomach with low pH for digesting large prey. Many barbs are benthivorous species, like the ancestral barb in Lake Tana's isolated system. Why then is piscivory, which is rare among cyprinids, so common in Lake Tana Barbus? The aim of present study was to compare the performance and techniques of these piscivorous Barbus with known piscivores from other fish families. We studied prey handling times over prey size, prey capture using high‐speed movies, and assessed the effect of prey size on performance and prey selection in the field. Performances were explained by functional morphology of their feeding system. Overall, Lake Tana's piscivorous Barbus perform relatively 'poor', compared to piscivores from other fish families. For example, Lake Tana's piscivores are only able to handle prey fish smaller than 16% of their own body length. However, Lake Tana lacks potential piscivorous competitors, rendering the piscivorous Barbus by far the 'best' and apparently highly successful. They have adapted to all available macro‐habitats (littoral, offshore pelagic and offshore benthic), using different techniques (ambush, pursuit and cruising), a unique scenario for barbs.     

Major histocompatibility genes in the Lake Tana African large barb species flock: evidence for complete partitioning of class II B, but not class I, genes among different species

The 16 African large barb fish species of Lake Tana inhabit different ecological niches, exploit different food webs and have different temporal and spatial spawning patterns within the lake. This unique fish species flock is thought to be the result of adaptive radiation within the past 5 million years. Previous analyses of major histocompatibility class II B exon 2 sequences in four Lake Tana African large barb species revealed that these sequences are indeed under selection. No sharing of class II B alleles was observed among the four Lake Tana African large barb species. In this study we analysed the class II B exon 2 sequences of seven additional Lake Tana African large barb species and African large barbs from the Blue Nile and its tributaries. In addition, the presence and variability of major histocompatibility complex class I UA exon 3 sequences in six Lake Tana and Blue Nile African large barb species was analysed. Phylogenetic lineages are maintained by purifying or neutral selection on non-peptide binding regions. Class II B intron 1 and exon 2 sequences were not shared among the different Lake Tana African large barb species or with the riverine barb species. In contrast, identical class I UA exon 3 sequences were found both in the lacustrine and riverine barb species. Our analyses demonstrate complete partitioning of class II B alleles among Lake Tana African large barb species. In contrast, class I alleles remain for the large part shared among species. These different modes of evolution probably reflect the unlinked nature of major histocompatibility genes in teleost fishes.Electronic Supplementary Material Supplementary material is available for this article at .An erratum to this article can be found at     

Role of thyroid hormone in the ontogeny and morphological diversification of <Emphasis Type="Italic">Barbus intermedius sensu</Emphasis> Banister, 1973 of Lake Tana in Ethiopia

The role of thyroid hormone in the ontogeny of the Large African barb Barbus intermedius sensu Banister, 1973 (Teleostei; Cyprinidae)—the presumed ancestral species for the species flock of the large African Barbs of Lake Tana (Ethiopia)—has been studied. It has been shown that the thyroid hormone influences the timing and rate of many morphogenetic processes and a change in its level causes heterochronies, affecting the definitive morphology of fish. Thus, it can be assumed that even a slight change in activity of the thyroid axis during ontogeny can be the cause of wide morphological variability in B. intermedius and may underlie explosive diversification of the species flock of Large African Barbs of Lake Tana.     

Predicting and testing resource partitioning in a tropical fish assemblage of zooplanktivorous 'barbs': an ecomorphological approach

Morphometrics on 25 critical feeding structures predicted conspicuous specializations in Barbus tanapelagius (pursuit hunting for zooplankton), Labeobarbus brevicephalus (surface dwelling pump-filter-feeder on zooplankton) and Barbus pleurogramma (particulate feeding on tough, benthic food), whereas for Barbus humilis intermediate values predicted few constraints and specializations in feeding. These potential niches, set by fish size and structural constraints, were tested by comparing gut contents collected during a 24 months sampling programme on Lake Tana, Ethiopia. Zooplankton dominated the diet of B. tanapelagius (75% of gut volume) and L. brevicephalus (39%). The guts of B. pleurogramma and juveniles of the large labeobarbs showed an array of benthic food types, whereas B. humilis had the widest food niche, both zooplankton (40% of gut volume) and benthic invertebrates. Although the pelagic species showed the largest spatial overlap, their size, feeding modes and utilization of zooplankters differed: L. brevicephalus preyed predominantly on the larger zooplankton ( Daphnia sp.) and B. tanapelagius also on smaller species ( e.g. Bosmina sp. and cyclopoid copepods). The spatial segregation between B. tanapelagius (pelagic) and the juvenile labeobarbs (littoral) indicated the possibility for a small pelagic barb fishery without negative effects on the labeobarb stocks. The ecomorphological approach using the 'Food-Fish Model' appeared to predict competitive positions and resource partitioning appropriately, and is of major importance to evaluate food web interactions.     

Phenetic relationships and probable ways of morphological diversification of the African large barbs (<Emphasis Type="Italic">Barbus intermedius</Emphasis> complex) in Lake Tana (Ethiopia)

29 quantitative and 18 qualitative skull characters were studied in the African large barbs (Barbus intermedius complex) from Lake Tana (Ethiopia). Based on results of principle components analysis phenetic relationships of 12 morphotypes were assessed. External features diagnostic of the morphotypes were found in individuals substantially different in the skull characters. The obtained data along with the data on morphological diversity of large barbs from Ethiopia beyond the Lake Tana basin suggest origination of the Lake Tana complex of forms from a complex of forms that existed in the upper reaches of the Blue Nile before the lake came into being, and the extant morphotypes of the Lake Tana barbs originated from three or four forms that composed the ancestral riverine local complex.     

Genetic divergence among morphotypes of Lake Tana (Ethiopia) barbs

The genus Barbus is a paraphyletic assemblage that includes three different ploidy levels: diploid, tetraploid and hexaploid. 'B.' intermedins , which inhabits Lake Tana (Ethiopia), is a hexaploid taxon that forms a 'species flock' consisting of at least 14 morphotypes. Thirty-one presumptive allozymic loci were analysed in a large sample of 'B.' intermedius. No diagnostic loci were observed between the morphotypes, but substantial differences in allele frequencies were found. One morphotype, Acute, differs significantly from the others. The presence of this genetic differentiation within the lake strongly suggests that the morphotypes are diverging lineages. Because of some parallel characteristics with cichlid species flocks, we put forward the hypothesis that the morphotypes of barbs in Lake Tana represent a species flock that originated via sympatric speciation with more rapid morphological than molecular differentiation. If the differentiation and the reproductive isolation are real, it is probable that the morphotypes already are or will soon reach the status of species.     

Assessment of the role of chemoreception in the mate choice in barbs of the Barbus intermedius complex from Lake Tana, Ethiopia

The role of chemoreception in mate choice was studied in barbs of the complex Barbus intermedius from Lake Tana. It was found that chemical communication may be used in reproductive interactions of the barbs, and that males are responsible for the choice. Mature males tend to prefer ripe females and avoid both immature females and mature males, or are indifferent to them. It was shown that selectivity in mate choice by certain barb morphotypes could reduce the probability of interbreeding between different morphotypes, but not provide for reproductive isolation between them. It was hypothesized that the mate groups are finally formed as a result of direct interactions between potential partners, mediated by tactile reception or specific behavioral patterns.     

The role of the thyroid gland hormone in ontogenesis and morphological diversification of Barbus intermedius sensu of Lake Tana in Ethiopia

The role of thyroid gland hormone in the ontogenesis of the Large African Barb Barbus intermedius sensu Banister, 1973 (Teleostei; Cyprinidae)--the supposed ancestral species which produced the species flock of the large African Barbs of the Lake Tana (Ethiopia)--has been studied. It has been shown that the thyroid hormone influences the period of many morphological processes and a change in its level causes heterochrony, reflecting on the definitive morphology of fish. Thus, it can be assumed that even a slight change in activity of the thyroid axis in the process of ontogenesis can be the cause of wide variability of the morphological features of B. intermedius and is at the basis for the fast diversification of the species flock of Big African Barbs of Lake Tana.     

Lake Tana large barbs diversity: Developmental and hormonal bases

Helerochronies caused by changes in the hormonal status are proposed to play a major role in the diversification of the Lake Tana barbs of Labeobarbus (=Barbus) genus (Dgebuadze and Mina, 2008). To test this hypothesis, we (i) compared temporal patterns of cranial development in L. intermedius, likely ancestral for Tana species flock, and L. brevicephalus and L. megastoma proposed to be paedomoiphic and peramoiphic, respectively; (ii) studied the role of thyroid hormone (TH) in the regulation of L. intermedius development; (iii) evaluated the effect of TH changes on the definitive morphology, and (iv) evaluated the possible role of TH in diversification of Tana barbs. Skull development was found to be heterochronicaily changed in L. brevicephalus and L. megastoma as compared to L. intermedius; revealed heteroclironies may account for differences in skull proportions distinguishing these barbs, TH was shown to play a major role in the regulation of ontogeny. Changes in TH level cause heterochronies in the development of a skull, axial skeleton, fins and fin girdles, squamation, and dentition. Heteroclironies were shown to influence adult morphology of these structures. These findings support the hypothesis of the involvement of developmental and hormonal factors in the diversification of the Lake Tana large barbs.     

Polymorphism of skull proportions in large African barbs <Emphasis Type="Italic">Barbus intermedius</Emphasis> sensu Banister (Cyprinidae) from Lakes Awasa and Langano (Rift Valley,Ethiopia)

After performing principal component analysis (PCA) of 13 characters (indices) of skulls of large African barbs (Barbus intermedius complex sensu Banister, 1973) from Lakes Awasa and Langano (Ethiopia) distributions of individuals in coordinates BL and PC1 were studied where BL is the basal skull length and PC1 is the first principal component. It is shown that in both the lakes there are groups of barbs whose individual ontogenetic trajectories are situated within separated ontogenetic channels. Separation of the channels is by far less distinct than that of the channels of generalized and specialized forms of barbs from Lake Tana and not correlated with differences in external traits. It is suggested that the revealed polymorphism is connected with food resource partitioning and might serve as a prerequisite of further divergence of sympatric forms of large African barbs.     

Resource partitioning by Lake Tana barbs predicted from fish morphometrics and prey characteristics

We develop a food-fish model (FFM), whichquantitatively relates properties of aquaticfood types (size, shape, escape velocity,habitat, mechanical properties and chemicalquality) to feeding structures of cyprinidfish. The model is based on functionalmorphology and experiments on search, capture,selection, and internal processing of food byfish. The FFM shows which food properties aremost critical in feeding and how fish canoptimise coping with them. Relative food sizeimposes the highest demands, followed by preyvelocity, food habitat and mechanicalproperties. These overrule taxonomic affinitiesof food types. Highly demanding food types(large, fast prey, suspensions of plankton,benthic prey and mechanically tough items)impose incompatible morphological requirementson fish. We apply the FFM to the endemic Barbus species flock of Lake Tana (Ethiopia),since the structural diversity of its 14species reflects recent adaptations to trophicniches. We predict their potentials inutilising different food types by quantitativecomparisons of 35 parameters, measured for eachspecies, with the values for each foodspecialist derived from the FFM. These dietpredictions are tested against gut contentsfrom 4,711 fish, sampled over seasons andhabitats. Gut contents and predictions show agood overall fit. The value of the model isshown by its resolution in predicting resourcepartitioning among the barbs. For the 14 barbsa trophic hierarchy with six major trophicgroups is reconstructed which closely matchesthe predictions. Trophic specialists (> 65%by volume of a single food type) are alsostructurally specialised, whereas less extremeanatomical structures characterise trophicgeneralists, allowing them to switch betweenfeeding modes. Trophic generalists are bestdefined by behavioral flexibility, sincefeeding modes integrate both fish and foodcharacters. The FFM is of practical use inevaluating the role of morphological diversityin an ecosystem and enables the analysis oftrophic interactions in fish communities and ofthe cascading effects by environmental change.Such an approach can be instrumental in thedevelopment of management strategies forfisheries and in conservation of biodiversity.     

Lake Tana large barbs: phenetics, growth and diversification

Among the Lake Tana large barbs of about 10 cm SL only representatives of the' acute'morphotype can be distinguished, in the size range 10–20 cm SL' bigmouth big eye' can be identified also. As for the rest, very few individuals can be confidently affiliated with a particular morphotype most of them looking like' intermedius'. Even within the range of 20–30 cm SL some individuals are still difficult to identify. Principal component analysis of cranial characters revealed discrete groups of morphotypes. Differences in both external and cranial characters of the morphotypes result from divergence which is most pronounced when fish are 4–5 years old and 20–25 cm SL. This divergence cannot be related exclusively with differences in the growth rates of individuals representing different morphotypes. Differences in food composition between the morphotypes probably increase in parallel with their morphological divergence. Differences between the morphotypes in the lateral line (11) and the gill rakers (Sb) counts were revealed using ANOVA. Comparison of the Lake Tana Barbus complex of forms with that previously known from Lake Lanao (Philippines) suggests that in both lakes the different forms arose sympatrically but that sympatric speciation in Lake Lanao has advanced further than in Lake Tana.     

Spatial and Temporal Variation of Cestode Infection and its Effects on Two Small Barbs (Barbus humilis and B. tanapelagius) in Lake Tana, Ethiopia

Pseudophyllidean cestodes as Ligula have a complex life cycle with cyclopoid copepods as first intermediate host, zooplanktivorous fish as second, and piscivorous birds as final host. We studied the effects of diet, season and habitat occupation on the prevalence of plerocercoid larvae of the tapeworm Ligula intestinalis in two closely related small barbs and the effects of the parasites on the barbs life histories in Lake Tana (Ethiopia) during 1 year. In all affected barbs L. intestinalis caused retardation in gonad development, maturation at reduced size and lower absolute fecundity. Infection rate, averaged over all habitats was significantly higher in B. tanapelagius (10%) than in B. humilis (6%). Below a threshold of 48 mm the infection rate was zero for both barbs, this coincided with a very low proportion of copepods in their diets, increasing up to 90 and 55%, respectively, for their largest size class (81–90 mm). The relatively high infection rate in B. tanapelagius is explained by its obligatory zooplanktivorous feeding behaviour, ingesting a relatively high proportion of infected cyclopoid copepods. This is in contrast with B. humilis, which is a polyphagous species, feeding both on zooplankton and benthic invertebrates. Significant seasonal effects in infection rates were observed. In both barb species infection rates were lower during the breeding season. Only for B. tanapelagius a significant negative correlation was observed between rain fall and infection rate, probably caused by an increased turbidity that decreases feeding efficiency on zooplankton. Habitat type had also a significant effect on infection rate. Barbus humilis showed a much higher infection rate in shallow clear water (10%) than in shallow turbid water (3%), whereas B. tanapelagius showed much higher infection rates in the shallow sublittoral (13%) than in the deeper pelagic (7%). Most likely, birds predate more efficiently on barbs in shallow clear waters than in shallow turbid and deep waters.     

Growth, biomass, and production of two small barbs (Barbus humilis and B. tanapelagius, Cyprinidae) and their role in the food web of Lake Tana (Ethiopia)

Growth, biomass and production of two small barbs (Barbus humilis and Barbus tanapelagius) and their role in the food web of Lake Tana were investigated. From length–frequency distribution of trawl monitoring surveys growth coefficient, Φ′ values were estimated at 3.71–4.17 for B. humilis and 3.70–4.14 for B. tanapelagius, respectively. Values for B. humilis were confirmed in pond experiments. Mean biomass of the small barbs was 13.3 kg fresh wt ha⁻¹, with B. humilis being most abundant in the littoral and sub-littoral zones, whereas B. tanapelagius was most abundant in the sub-littoral and pelagic zones. The two small barbs had a production of 53 kg fresh wt ha⁻¹ year⁻¹. Although their P/B ratios of about 4.0 were relatively high for small cyprinids, both their biomass and production were low in comparison with other small fish taxa in other tropical lakes. Of the zooplankton production only about 29% was consumed by the small barbs. However, they did not utilize calanoid copepods, which were responsible for approximately 57% of the zooplankton production and it is likely that small barb production was food limited during certain periods of the year. Piscivorous labeobarbs consumed about 56% of the small barbs production annually, but additionally, Clarias gariepinus, and many bird species were also preying on them. Therefore, limitation of Barbus production by predation during certain periods in the year cannot be excluded.     

Reproductive segregation among the Barbus intermedius complex of Lake Tana,Ethiopia. An example of intralacustrine speciation?

Reproductive segregation among the large barbs Barbus intermedius complex, Cyprinidae , of Lake Tana, Ethiopia, was investigated in the mouths and upstream reaches of rivers tributary to the lake, during the spawning seasons of 1993 and 1994. The percentage of running fish of four Barbus morphotypes [Acute (Ac), Bigmouth big-eye (Be), Intermedius (In), and Shorthead (Sh)] peaks with the highest water level of the lake. Large differences are apparent in the minimum fork length of running fish among morphotypes [c. 12 cm for male Sh and 28 cm for male Bigmouth small-eye (Se)]. Significant differences occur in temporal, but especially in spatial patterns of the relative abundance of running morphotypes, when the lake is compared with the four major river mouths and their upstream reaches. These different relative abundances are clearest among Ac, Be, In, Sh and Troutlike (Tr). Also, the absence from the rivers of some morphotypes that are usually abundant in the lake itself [Barbel (Ba), Dark (Da), Lip (Li), and Zurki (Zu)], and the abundant presence in the rivers of Tr, rare in the lake, strengthen the hypothesis that spatial and/or temporal spawning segregation is present among the morphotypes. There are indications that Bighead (Bh) and Zu spawn in the lake itself, while other morphotypes were only found running in or near rivers. Combined with the morphological, ecological, and genetic differences, these data suggest the existence of several Barbus species, forming a unique species flock, which probably evolved within the lake. The presence of several, instead of one Barbus species complicates management of sustainable fisheries and protection of the lake's biodiversity. Present results show the importance of the rivers for Barbus reproduction and stress the urgent need for careful management of fishing activities in spawning times and areas.     

Observations on reproduction of the Lake Tana barbs

During the rainy season (September–November 1994), in tributaries of Lake Tana (Ethiopia), the Gumara, Gelda and Negashu, barbs of diVerent morphotypes were ready to spawn, running and spent. The most numerous were barbs similar to Barbus intermedius common in many rivers of the Ethiopian Highlands. They were also found spawning in the Gumara and Gelda Rivers in the dry season (February 1996, 1997). There were no evident diVerences in time and place of spawning of barbs representing diVerent morphotypes. Only zurki, bigmouth mini-eye and bighead were not found in the rivers. Spatial and temporal segregation may contribute to reproductive isolation of morphotypes spawning in rivers, but if reproductive isolation exists the main role must be played by diVerences in mating behaviour. Morphotypes of the Lake Tana barbs have diVerent migration patterns. Barbs similar to the riverine B. intermedius may spend their lives in rivers or in the lake, moving to and fro. Barbs of acute, bigmouth big-eye and, perhaps, other morphotypes migrate downstream as alevins or fry, stay in the lake for several years and migrate back into rivers for spawning after reaching maturity.     

Changes of Skull Proportions in Ontogeny of Barbus canis from Lake Kinneret and of the Lake Tana large-mouthed barbs, Barbus intermedius Complex

Skull proportions of Barbus canis individuals drastically change during their life. Yearlings have skull proportions characteristic for barbs of the NEC phenetic group (sensu Mina et al. 2001 Environ. Biol. Fish. 61: 242–252), but with age become similar to those of the Lake Tana large-mouthed morphotypes which acquire their specific features at 3–7 years old while at earlier ages they have skull proportions of a NEC form identified as 'intermedius' sensu Nagelkerke et al. (1994 Environ. Biol. Fish. 39: 1–21). The large-mouthed Lake Tana barbs originated from an ancestral NEC form through peramorphosis, but in relation to B. canis they are paedomorphic in skull proportions. B. canis supposedly originated from a NEC form as a consequence of an adaptive radiation that took place in western Asia.     

Diet segregation between two sympatric 'small'Barbus spp: an experimental study of mechanisms

Gut contents of two co‐occurring species of 'small' diploid barbs (<10  L _F cm) in Lake Tana revealed that zooplankton is the major diet component for B. tanapelagius (75% based on volume), but less prominent in B. humilis (40%). Functional response experiments in the laboratory were conducted to elucidate the mechanisms causing this difference. The type of functional response by the two 'small' barbs under different microcrustacean zooplankton densities (10, 20, 40, 60 and 80 ind.l⁻¹) was examined. The functional response of B. tanapelagius to increasing prey densities corroborates with Holling Type II model, whereas B. humilis exhibits a Type III functional response. Predation rate is higher for B. tanapelagius at low zooplankton density (<40 ind.l⁻¹) and equals the level of B. humilis at higher densities (>40 ind.l⁻¹). This suggests that at lower zooplankton densities B. humilis is a less efficient forager on zooplankton prey items than B. tanapelagius . In Lake Tana average zooplankton density is relatively low (<35 ind.l⁻¹). Under these food conditions, B. humilis is forced to feed on other food items (e.g. benthic invertebrates), whereas B. tanapelagius primarily feeds on zooplankton. The feeding potentials of the two 'small' barbs, as deduced from their morphology explain their different performances and their segregation in space and food resources.     

Individual ontogenetic trajectories and ontogenetic channels of three forms of large African barbs (<Emphasis Type="Italic">Barbus intermedius</Emphasis> complex). Analysis of experimental data

Changes of external morphological characters in three forms (morphotypes) of African barbs from Lake Tana (Ethiopia) belonging to Barbus intermedius complex were traced using electronic tags. It was shown that within age interval from two to four years individual ontogenetic allometric curves often cannot be described with a single power function equation. Individual ontogenetic trajectories in the principal components space within this interval strongly differ in shape meandering within ontogenetic channels outlined on the basis of cross-sectional data. Morphological differences between siblings are found that do not allow to classify some of them with the morphotype of their parents.