Environmental Biology of Fishes

, Volume 99, Issue 5, pp 473–485

Morphological variations of wild populations of Nile tilapia (Oreochromis niloticus) living in extreme environmental conditions in the Kenyan Rift-Valley

  • Titus Chemandwa Ndiwa
  • Dorothy Wanja Nyingi
  • Julien Claude
  • Jean-François Agnèse
Open Access
Article

Abstract

In this study, we use geometric morphometric method to establish morphological differences between natural populations of Nile tilapia from two extreme environmental conditions (high temperature and salinity) in Kenya, and compare them to two populations from regions experiencing less extreme conditions. To determine genetic influence on morphology, we correlated genetic data with morphological data. The study observed significant morphological differences between all studied populations, including three closely related hot spring populations (Bogoria, Chelaba and Turtle Springs), and two populations with similar genetic background inhabiting saline environments in Lake Turkana basin (Turkana and Crocodile Lake populations). In addition, allometric growth patterns of the seven populations differed significantly, demonstrating that developmental changes (plastic or adaptive) operated between these populations. Positive correlation between morphometric and genetic data confirmed the influence of genetic factors on morphology. All observed differences were attributed either to genetic and/or environmental factors, which seemed to play a major role in influencing morphology of wild Nile tilapia populations. We recommend further studies to be carried out under controlled conditions to confirm the role of temperature, pH and salinity in morphological diversification of Nile tilapia.

Keywords

Morphometrics Salinity Temperature Nile tilapia Rift-Valley Kenya 

Introduction

Cichlids are tropical freshwater fishes well known for their high rates of speciation often resulting in rapid radiation. They exhibit remarkably high levels of genetic and morphological diversity, which affect their morphology, ecology, behaviour and genomes (Nelson 1994; Barlow 2000; Chakrabarty 2005). A large extent of cichlid diversity is observed in three main East African Lakes (Victoria, Tanganyika and Malawi) where 1800 to 2000 species (60 % of all cichlid species worldwide) are present in species flocks (Greenwood 1984; Turner et al. 2001). These cichlid species flocks are primarily monophyletic groups of closely related species inhabiting the same ecosystem, and evolved from a single ancestral species (Greenwood 1974). Within these three East African Lakes, some cichlid species went through rapid morphological diversification and speciation, leading to formation of many new species within a few thousand generations (Sage et al. 1984; Meyer et al. 1990; Owen et al. 1990). In Lake Victoria for instance, the radiation of monophyletic haplochromine species occurred within a geologically short time of about 750,000 years and resulted to the emergence of approximately 500 cichlid species (Greenwood 1984; Abila et al. 2004; Kerschbaumer and Sturmbauer 2011). Additionally, evolutionary older cichlid assemblages of Lake Tanganyika have shown the highest degree of morphological differentiations (Salzburger et al. 2002).

Under the ecological theory of diversification (Schluter 2000), a strong relationship exists between speciation events and colonization of new environments, or niche diversification, which may in turn induce changes in organism’s behaviour, morphology and physiology (Price et al. 2003). Generally, selection pressures that exist in new environments favour divergence of populations from their ancestral forms. A classic example is provided by the case of the three spined sticklebacks (Gasterosteus aculeatus) in Canada, where two sympatric forms (limnetic and benthic) emerged as a consequence of ecological divergence (Schluter and McPhail 1993). This example among others (Meyer 1987; Day et al. 1994; Imre et al. 2002) has provided evidence of morphological diversifications as a result of environmental variations. However, Gagnaire et al. (2013) proposed that ecological divergence may actually involve multiple facets of species divergence on which divergent selections may operate, hence the need for an integrative approach to study the process of speciation. Some of the factors proposed to influence divergence include variations in quantitative phenotypic traits, regulation in gene expression and differential transmission of particular allelic combinations (Gagnaire et al. 2013).

The Nile tilapia (Oreochromis niloticus), the subject of our study, is a teleost fish of the family Cichlidae, has its native range in Africa and is widely known for its importance in aquaculture. This species has been recognized for its ability to tolerate a wide range of salinity, dissolved oxygen and temperature (Farmer and Beamish 1969; Avella et al. 1993), high fecundity and rapid growth rate (Welcomme 1967) and an omnivorous mode of feeding (El-Sayed 1999). Due to these attributes, the species has the ability to adapt rapidly to new environments, showing a wide range of biological responses to different environmental conditions both in culture and in nature (Schofield et al. 2011; Grammer et al. 2012). As a result of this plasticity (or of the high degree of physiological tolerance found in the Nile Tilapia), the species has been able to successfully establish itself in extreme environmental conditions such as temperate winter conditions (Peterson et al. 2005; Grammer et al. 2012), hot springs (Trewavas 1983; Nyingi et al. 2009; Ndiwa et al. 2014) and saline waters (Schofield et al. 2011). Studies by Collyer et al. (2005) in pupfish (Cyprinodon tularosa) noted that the geographical spread of species across a broad range of ecological conditions is accompanied by equally diverse morphological variations, which are strongly correlated to the environmental conditions.

Within the Kenyan Rift Valley, Nile tilapia has colonized numerous newly formed habitats within a system characterised by a number of drainages comprising a series of rivers, springs and lakes. This complex system is a host to four out of the seven described sub-species of Oreochromis niloticus by Trewavas (1983): O. n. baringoensis from Lake Baringo; O. n. sugutae from River Suguta; O. n. vulcani from Lake Turkana. Recently Nyingi et al. (2009) discovered a natural population from Lake Bogoria Hotel genetically differentiated from but nevertheless close to the Lake Baringo population. The purpose of the current study is to investigate whether populations of Nile Tilapia have undergone significant morphological diversifications, and whether these differences can be related to environmental factors (either as the result of phenotypic plasticity or adaptive evolution).

According to Collyer et al. (2005, 2007), variations in salinity have the potential to affect fish body shape. These authors reported a general decrease in body circumference to overall body size at high salinity on pupfish (Cyprinodon Tularosa). Similarly, high water temperatures can influence both the growth rate and development rate of an organism (Atkinson 1995). In Zebra fish (Danio rerio) for instances, raising of juveniles under four different developmental temperatures resulted to highly variable phenotypes (Sfakianakis et al. 2011). A similar study by Georgakopoulou et al. (2007) on European seabass (Dicentrarchus labrax, L) showed that samples reared under lower temperatures (15 °C) were more slender than those reared under higher temperature (20 °C). Even though no studies have been carried out to establish the effect of water pH on fish shape, extreme changes in water pH have been reported to cause fish deaths by altering other aspects of water chemistry, for instance increase in pH increases toxicity of ammonia (Schofield 1976; Wurts and Durborow 1992).

Taking into consideration the local genetic variations observed in three recently discovered populations of O. niloticus within the hot springs (Chelaba, Lake Bogoria Hotel and Turtle) of Loboi Swamp (Nyingi et al. 2009; Ndiwa et al. 2014), we hypothesized that these fish populations may also differ in shape due to the influence of high salinity and high temperature. In addition to the Loboi Swamp fish, another population of O. niloticus found in the northern part of the Kenyan Rift-Valley also experiences extreme environmental conditions. The Crocodile Lake, one of the three Lakes located on the Central Island of Lake Turkana is saline, with electrical conductivity as high as 10,590 μS/cm and alkaline with a pH of around 10.0 (Avery 2010). Depending on the level of Lake Turkana, Crocodile Lake can either be submerged or emerged. During high water level the lake has been known to disappear (Hay 1976; Cerling 1979) while in low water levels, it remains isolated. The last time the two lakes were reported to have connected was in 1902 (Hopson 1982). So far only three species of fish have been reported to survive in Crocodile Lake namely: Clarias gariepinus, Haplochromis sp., and O. niloticus.

Both the Loboi and Crater Lake populations have adapted to extreme conditions, high temperature and high salinity levels respectively, while other closely related populations (Lakes Baringo, Turkana and Victoria) do not experience such conditions. These populations thus present an opportunity to observe how extreme conditions affect the morphology of the fish. To achieve this goal, we compared individuals from seven populations, four with extreme conditions (hot springs; Lake Bogoria Hotel Spring, Chelaba Spring, Turtle Spring and saline; Crocodile Lake populations), and three that experience more favourable environmental conditions (Lakes Baringo, Turkana and Victoria). Six of the studied Nile tilapia populations occur in their natural habitats except the Lake Victoria population, which was introduced into Lake Victoria from Lake Albert in Uganda (Fryer and Isles 1972; Trewavas 1983).

The study provides an insight into morphological responses of these fish populations to extreme temperature and salinity. Occurrence of phenotypic differences between the populations studied will confirm that the environment or/and the genetic background play an important role in evolution of cichlids, while lack of phenotypic differences will imply that the high degree of tolerance in Nile Tilapia is related to its propensity to invade new environment, and in keeping a very canalized phenotype.

Materials and methods

Specimens from three hot spring populations of Loboi swamp drainage system (Chelaba, Lake Bogoria Hotel and Turtle) and four other lakes (Baringo, Turkana, Crocodile and Victoria) were studied (Table 1; Fig. 1). The three hot springs are characterised by elevated water temperatures (approximately 36 °C), low dissolved oxygen concentration (3.5–3.9 mgl−1) and slight acidity pH (6.4–6.9) (Ashley et al. 2004). The springs drain into Loboi swamp of about 1.5 km2, which is covered by dense vegetation of Typha domingensis and Cyperus papyrus. As the water exits the swamp, it is poorly aerated (Dissolved Oxygen (DO) = 0.2–1.1 mgl−1) and has a higher pH (7.4–8.3) (Ashley et al. 2004). Comparatively, Crocodile Lake is a crater formed island lake within the Lake Turkana. It is characterised by relatively higher water pH (10) and almost three times higher salinity (conductivity, 10,590 μS/cm) than Lake Turkana, which has pH and a conductivity of 9.3 and 3420 μS/cm respectively (Avery 2010). Lake Baringo, a fresh water Lake located about 52 km from the Loboi swamp has DO concentration ranging from 5.89–6.7 mg/L, temperature range of 22.1–31.8 °C, pH ~ 8.5 and conductivity of about 580 μS/cm (Omondi et al. 2014). Lake Victoria, also fresh water body is characterized by water temperatures ranging from 23.8 to 31.5 °C, pH 6.2–9.2, and conductivity of approximately 170 μS/cm (Ochumba 1990; Crul 1995).
Table 1

Names of the seven populations of O. niloticus studied, their respective sample sizes (N), coordinates of study sites, habitat characteristics, surface water temperature range, conductivity (salinity) and size of the water body

Population

Number

Co-ordinates

Habitat

Temperature

Conductivity

Surface area

Bogoria Hotel Spring

41

0°21’44”N, 36°03’04”E

Hotspring

32–36 °C

<100 μS/cm

1.5 Km2

Chelaba Spring

30

0°21’30”N, 36°02’58”E

Hotspring

32–36 °C

<100 μS/cm

1.5 Km2

Turtle Spring

24

0°21’44”N, 36°02’41”E

Hotspring

32–36 °C

<100 μS/cm

1.5 Km2

Crocodile Lake

28

3°29’34”N, 36°01’54”E

Saline

25–30 °C

10,590 μS/cm

<1 Km2

Lake Turkana

38

3°30’46”N, 35°54’53”E

Saline

26–27 °C

3420 μS/cm

6405 Km2

Lake Baringo

43

0°36’43”N, 36°01’31”E

Normal

22–32 °C

580 μS/cm

130 Km2

Lake Victoria

33

0°38’17”N, 33°58’35”E

Normal

23–32 °C

170 μS/cm

68,800 Km2

Fig. 1

Map showing the location of all the six natural populations and one introduced population of O. niloticus: Lake Turkana at Lake Turkana drainage system, Crocodile Lake in the Central Island of Lake Turkana, Lake Baringo at Lake Baringo drainage system, Turtle Spring, Chelaba Spring and Bogoria Hotel Spring in the Loboi Swamp, and Lake Victoria (introduced) at L. Victoria drainage system

Fish specimens were collected from the hot springs (Chelaba, Lake Bogoria Hotel and Turtle) and Lakes Baringo, Victoria, Turkana and Crocodile using seine nets and scoop nets. Fish were anaesthetised and killed using an overdose of MS-222 before fixing in 4 % formalin solution for seven days. The formalin was washed and voucher specimens preserved in 70 % ethanol, and are curated as reference material at the National Museums of Kenya. The standard lengths (distance from the tip of the snout to the base of caudal fin) of the fish were measured using a digital Vernier calliper in order to estimate their sizes.

Body shape of the samples was quantified using landmark based geometric morphometric method (Rohlf and Marcus 1993). Images of preserved samples of fish were obtained by using a flatbed scanner (Herler et al. 2007) and imported into the program tpsUtil version 1.53 (Rohlf 2012). A total of 14 landmarks (Fig. 2) were chosen based on their capacity to capture overall body shape, and their coordinates (x,y) digitized on the left hand side of each sample using the program tpsDig version 2.16 (Rohlf 2010). The coordinates were saved in a tps format data file and used for further morphometric analysis. A total of 237 adult fish specimens from seven wild population of Nile tilapia were analysed. Maturity of the specimens was ascertained by observing their gonads. Their standard length sizes ranged from 40 to 207 mm.
Fig. 2

Position of 14 different landmarks used in the study: (1) anterior tip of the snout with mouth closed, (2) centre of the eye, (3 & 4) anterior and posterior insertions of the dorsal fin, (5) dorsal origin of caudal fin, (6) point halfway between dorsal and ventral origin of caudal fin, (7) ventral origin of caudal fin, (8) posterior insertion of anal fin, (9) anterior origin of anal fin, (10) anterior origin of pelvic fin, (11) juncture of the ventral edge of operculum with the ventral outline, (12) most posterior end of the mouth, (13) anterior origin of pectoral fin and (14) most posterior end of operculum

Data analysis

The data file was imported into R software program and analysed using Momocs (Bonhomme et al. 2014), Geormoph (Adams and Otarola-Castillo 2013; Dryden 2013) and Ape packages (Padis et al. 2004; Padis 2012). Coordinates were scaled to unit centroid size, translated, rotated and optimally superimposed through the generalized partial Procrustes Analysis. Procrustes coordinates were projected into Eucledian tangent shape space for further analysis (Dryden and Mardia 1998; Claude 2008). In order to establish morphological variations between the different populations of fish, we used principal component analysis (PCA) of the superimposed coordinates of the landmarks. Variations were visualised by plotting multivariate ordination of specimens on the first two principal component (PC) axes (PC1 and PC2). Contribution of original variables on corresponding eigenvectors or PC loadings were analysed to establish shape features that had the greatest influence on each PC.

Nested MANCOVA models were conducted on the principal components (PC) of shape data in order to assess the influence of locality, and environmental factors. Habitat and locality were considered as explanatory variables and centroid size as a covariate in the model. Size was added as a covariate to understand the effect of habitat and locality on allometric growth patterns.
$$ Response\sim Size+ Habitat+ Habitat: Size+ Habitat: locality+ Habitat: Locality: Size $$

To represent and appraise morphological differences between the localities, a linear discriminant analysis was performed on the dataset of the Procrustes coordinates.

The Fst values for 5 populations of Nile tilapia (Chelaba, Bogoria, Turtle, Turkana, Baringo and Victoria) used in the current study were obtained from our previous studies (Ndiwa et al. 2014). Alternatively, the Fst values for Crocodile Lake and Lake Victoria populations were obtained from unpublished data. All our Fst values were calculated from microsatellite molecular data using Genepop software (Rousset 2008). A mantel test analysis (Mantel 1967) was performed through permutation to correlate the Fst Values and Mahalanobis distance. This analysis aimed at confirming the influence of genetic factors on the morphology. Lastly, minimum spanning network based on Mahalanobis distance was constructed on R based on Ape package (Padis et al. 2004; Padis 2012) in order to visualize phenotypic relationship between the seven wild populations of Nile tilapia.

Results

Comparison of body sizes of the fish samples from the seven populations of Nile tilapia (Fig. 3) indicated minimal variations between specimens from Lake Baringo, Chelaba Spring, Crocodile Lake and Turtle Spring. However, greater body size differences were observed between the latter populations and populations from Bogoria Spring, Lake Turkana and Lake Victoria. Body size comparison among individuals of same population showed least variations within Crocodile Lake (Fig. 3).
Fig. 3

Size distribution of the fish samples collected from the seven localities

The first shape PC accounted for 23 % of the observed variations. This PC was mostly related with the variation of the head, caudal peduncle and anal fin (Fig. 4). The PC2 on the other hand accounted for 14.5 % of variation mostly within the depth of the body and caudal peduncle (Fig. 4a).
Fig. 4

a Differentiation of the seven populations of Nile tilapia based on Principal Component Analysis (PCA). On fish shape images (PC 1, 2), Grey = mature (large samples); Black = small (young) samples. In the scatterplot, cro = crocodile; Tur = Turkana; Che = Chelaba; Bog = Bogoria; Trt = Turtle; Bar = Baringo; Vic = Victoria b Differentiation of the seven populations of Nile tilapia using Linear Discriminant Analysis (LDA). On fish shape images (LDA 1, 2), Grey = mature (large samples); Black = small (young) samples. In the scatterplot, cro = crocodile; Tur = Turkana; Che = Chelaba; Bog = Bogoria; Trt = Turtle; Bar = Baringo; Vic = Victoria

Even though there were overlaps between most of the populations under study, significant morphological differences were observed between all populations. A strong separation was observed between the hot spring populations and both Crocodile Lake, and Lake Turkana populations based on PC1. However, PC2 showed greater differentiation between Lakes Turkana and Crocodile populations than the latter (Fig. 4a).

Surprisingly, we observed that populations of Lakes Turkana and Victoria were similar morphologically. These results were unexpected since the two systems have different ecological characteristics.

Differences between populations and habitats

Both differences in habitats and localities significantly explained the shape variations between the seven populations of Nile tilapia studied (Table 2). Linear discriminant analysis based on the sampling locations correctly assigned 84.8 % of the individuals to their respective localities. All samples from Lake Victoria were assigned to their real group (Table 3). However, most of individuals from the hot springs populations that were not correctly assigned to their respective localities were either assigned to another hot spring population(s), or to L. Baringo (Table 3). Assignment of individuals from other localities is shown on Table 3.
Table 2

Results of multi-variate analysis of covariance (MANCOVA) performed on body shape considering all the samples

Factor

df

Pillai

approx. F

num. df

den. df

P-value

Size

1

0.8282

40.167

24

200

<0.0001

Habitat

2

1.2667

14.464

48

402

<0.0001

Habitat:Size

2

0.3991

2.088

48

402

<0.0001

Habitat:Locality

4

2.0469

8.864

96

812

<0.0001

Habitat:Locality:Size

4

0.5402

1.321

96

812

0.0268

Table 3

Linear discriminant analysis (LDA) assignment of individuals to different localities and groupings according to habitat types, percentages of assignation are in brackets

Origin of the fish

False assignation

Habitat grouping

Lake Bogoria Hotel Spring (82.5 %)

Chelaba Spring (10 %)

Hotspring - Hotspring

Lake Baringo (5 %)

Hotspring - Normal

Turtle Spring (2.5 %)

Hotspring - Hotspring

Chelaba Spring (78.13 %)

Bogoria Spring (12.5 %)

Hotspring - Hotspring

Turtle Spring (9.37 %)

Hotspring - Hotspring

Turtle Spring (76.19 %)

Chelaba Spring (4.76 %)

Hotspring - Hotspring

Crocodile Lake (4.76 %)

Hotspring - Saline

Lake Baringo (14.29 %)

Hotspring - Normal

Lake Baringo (91.43 %)

Bogoria Spring (8.57 %)

Normal - Hotspring

Lake Turkana (94.59 %)

Crocodile Lake (5.41 %)

Saline - Saline

Crocodile Lake (83.87 %)

Turtle Spring (3.23 %)

Saline - Hotspring

Lake Turkana (3.23 %)

Saline -Saline

Lake Baringo (9.67 %)

Saline - Normal

Lake Victoria (97.06 %)

Lake Baringo (2.94 %)

Normal - Normal

Linear discriminant analysis (LDA) clustered the seven populations of O. niloticus into two distinct groups. The first cluster consisted entirely of individuals from Lake Victoria, while the second cluster comprised the other six populations (Fig. 4b).

However, partial separations were observed among the hot spring populations and also among Lakes Turkana and Crocodile populations (Fig. 4b). Shape variations related with both LDA1 and LDA2 were attributed to differences within the head and the caudal region.

Because we observed a significant interaction of locality with size, we compared small sized individuals to large sized individuals from each locality in order to establish their differences and similarity in growth. Generally, we observed a decrease in head size and length of the mouth with increase in size in all the six populations except Crocodile Lake, where no marked differences were observed between the heads of the small and large sized samples. We also observed differences in body depth between the small and large sized samples (Fig. 5).
Fig. 5

Growth patterns in the seven populations of Oreochromis niloticus. Grey colour represents the shape of the large samples while the black colour represents the shape of the small sized individuals

The closely related hot spring populations displayed different patterns of growth irrespective of their close spatial relationship. In Bogoria spring, small sized individuals were characterized by more elongated mouths compared to the other two spring populations. On the other hand, the heads of small sized individuals from Chelaba Spring showed slight elongation with greater dorsal displacement of their mouths. We also observed posterior displacement of the anal fin as individuals increase in size for Bogoria and Chelaba populations. Finally, large sized samples from Chelaba and Turtle Springs had elongated caudal peduncle length compared to small sized individuals, while no differences were observed for Lake Bogoria Hotel Spring population (Fig. 5).

Growth patterns of the Lake Turkana system (Turkana and Crocodile) populations differed significantly among themselves, and among other O. niloticus populations. For instance, great change in shape was observed between the small sized samples and the large sized samples involving drastic decrease of the depth of the body and caudal peduncle. Dorsal displacement of the mouth and anterior displacement of the eye, dorsal fin and operculum were also noted. Smaller differences were observed between small and large samples in Lake Turkana; anterior elongation of the mouth for small sized individual, anterior displacement of the lower jaw and pelvic fin of mature individuals, and slight elongation of the caudal peduncle (Fig. 5).

Within Lakes Victoria and Baringo populations, marked differences were observed between small and large sized samples. Comparison between their growth patterns showed that Lake Victoria samples had deeper caudal peduncles that increased in length with maturity of the fish, while Lake Baringo samples had slender caudal peduncles, which also increased in length with fish growth (Fig. 5).

Additionally, correlation of Fst values and Mahalanobis distance based on Mantel test analysis showed existence of significant relationship (P < 0.05) (Fig. 6).
Fig. 6

Correlation of Fst values against Mahalanobis distance showing positive correlation

Minimum spanning network computed from Mahalanobis distances between localities showed close proximity between the three hot spring populations. Additionally, Lake Turkana population showed close relationship to Lake Victoria population, while Crocodile Lake and Lake Turkana populations were separated (Fig. 7).
Fig. 7

Minimum spanning network showing overall similarity in body shape among the O. niloticus populations

Discussion

Morphological analysis revealed that specimens studied were first differentiated by their sizes. Fishes from Lakes Turkana and Victoria were the largest with a mean standard length sizes of 122.8 ± 4.38 mm (mean ± standard error) and 163.3 ± 4.40 mm, respectively. The smallest specimens were from Turtle Spring and Crocodile Lake, and exhibited mean standard length sizes of 76.6 ± 6.11 mm and 70.1 ± 1.72 mm, respectively. Differences were also observed in terms of variability of sizes; some samples had higher size range (Lake Victoria) than others (Crocodile Lake). These observed differences may partly represent artefacts due to sampling, and may not completely reflect the actual differences of fish sizes in different localities despite the knowledge that Lakes Turkana and Victoria host large specimens of O. niloticus (Trewavas 1983) due to their larger sizes and greater depths, compared to biotopes investigated in the present study.

The PC analysis revealed a large amount of shape differences whereby variations of the head, caudal peduncle and anal fin accounted for 23 % (Fig. 4a). These variations allowed discrimination of the Lake Victoria specimens from those of other Lakes (Baringo, Turkana and the hot springs). Within these latter populations, fishes from Lake Turkana were placed between the two groups of fishes.

Differentiation of the Lake Victoria population from all other population based on linear discriminant analysis was even clearer (Fig. 4b). There was also a clear separation between the hot springs specimens on one hand, and Lake Turkana and Crocodile Lake specimens on the other hand. Specimens from Lake Baringo appeared to fall between these two groups of specimens. All observed variations on both LDA1 and LDA2 were related to differences within the head and the caudal region. These results are congruent with the genetic variation observed by Agnèse et al. (1997), who demonstrated that fishes from Lake Baringo and Lake Turkana are genetically close. Later on, Nyingi et al. (2009) and Ndiwa et al. (2014) observed that fishes from the hot springs are even more closely related to fishes from Lake Baringo. If one assumes that fishes from Crocodile Lake should have a genetic background identical to the fishes of Lake Turkana, and that fishes from Lake Victoria introduced from Lake Albert (Fryer and Isles 1972; Trewavas 1983) should be less genetically related to the other fish studied, then morphological differentiation seems to be congruent with genetic differentiation. Indeed, our studies based on mantel test confirmed the role of genetic factors in influencing morphology. In addition, links between morphological and genetical differentiation have been reported by Stewart and Albertson (2010), who integrated both genetic and geometric morphometric approaches in studying scale eating cichlids of Lake Tanganyika, which had been reported to show frequency balanced polymorphism in the left-handedness and right-handedness of the mouth, an adaptation previously reported by Takahashi et al. (2007). Their morphometric results strongly supported the genetic basis of jaw laterality.

In our study, we also observed different allometric patterns of growth in all the seven populations of Nile tilapia. Considering the fact that most of the localities differ greatly in their environmental conditions, the observed allometries may be partly attributed to environmental differences, which might have acted on developmental stages, and shaped the differences observed between O. niloticus populations. Whether differences are adaptive or the by-product of phenotypic plasticity is still speculative at that stage. However, common garden experiments are necessary in order to establish the role of various environmental factors on fish morphology.

Shape changes found in our study could be functional and environmentally related. Indeed, our observations are congruent to findings by Ramler et al. (2014) who investigated the effect of different temperatures on body size and shape on stickleback (Gastrerosteus aculeatus). The authors reported significant differences between individuals across different temperature groups. Similarly, study on zebrafish (Danio rerio) juveniles reported occurrence of highly variable phenotypes induced by diverse thermal conditions during the early ontogenic stages (Sfakianakis et al. 2011). However, results obtained by Fujimura and Okada (2008) comparing (during their developmental stages) the lower jaw bones of a cichlid fish (Haplochromis chilotes) from Lake Victoria and another Riverine cichlid (O. niloticus) contradicted our findings. Irrespective of the habitat differences, the authors noted that most shape changes had a similar pattern during the growth of the two species showing that the development of these fish were relatively robust.

When looking at closely genetically related populations, one can observe a certain amount of morphological differentiation. The first case concerned specimens from the three hot springs. Specimens from Bogoria and Chelaba or Bogoria and Turtle springs were unevenly distributed along the first and second axis of the LDA analysis respectively (Fig. 4b). Differences between Chelaba and Bogoria Springs were mainly due to variations in body length and body depth, while differences between Bogoria/Chelaba and Turtle were as a result of variations in head length and body depth.

If we consider that the genetic background of these three populations is the same (Ndiwa et al. 2014), then these observed morphological differentiations could be attributed to phenotypic plasticity of the species. Very likely, environmental characteristics of these hot springs are slightly different. Bogoria Spring is the biggest of the three hot springs and runs a few hundred meters before entering the swamp. The two other springs have a smaller flow rate and enter in the swamp after only few meters. Nevertheless, as only a few genes can strongly act on morphological differentiation, one cannot exclude that morphological differences observed between the three hot spring populations was due to genetic differences. Ndiwa et al. (2014) observed that gene flow between these populations is limited by the swamp, which is covered by a dense vegetation of Typha domingensis and Cyperus papyrus. Decomposition of organic matter, and shading provided by these vegetation is responsible for low oxygen concentration and low temperatures, which act as barriers to free movement of the fish from one place to another. Similarly, Ndiwa et al. (2014) also observed that these fish are slightly introgressed by O. leucostictus genes from escapees of this species originating from aquacultural ponds within the drainage. The account of possible introgression of alien genes acting on morphological differentiation cannot be excluded.

The second case of genetically related populations in our study concerns fish from Lakes Turkana and Crocodile. There was a clear morphological differentiation between these two groups of fish as shown on Fig. 4b. Specimens from Crocodile Lake were generally smaller than those from Lake Turkana, and differed in body depth, head size and caudal length.

These two populations are genetically very close (unpublished results) since Crocodile Lake population originated from Lake Turkana about one century ago. Most likely, the observed differences can be as a result of phenotypic plasticity. These lakes offer two quite different environments to the fish. Lake Turkana is a large alkaline (pH 9.3) and slightly saline lake (conductivity 3420 μS/cm) hosting numerous fish species (48 following Hopson 1982). Comparatively, Crocodile Lake is a small size lake (about 0.32 km2), with high alkalinity (pH 10) and salinity (conductivity 10,590 μS/cm). Only three species of fish are present, two Cichlids (Aplochromis sp., O. niloticus) and one Clariid (Clarias gariepinus).

In conclusion, our study observed that populations of Nile tilapia inhabiting different localities displayed differentiated body shapes. The study also noted correlation between genetic distance and morphological distance. Despite these findings, further laboratory studies are necessary in order to confirm the importance of genetic and environmental factors in influencing morphological development. This can be achieved by breeding genetically identical siblings in different environmental conditions: first under low versus high salinity, and secondly under low versus high temperature.

Notes

Acknowledgments

We thank Mr. Joseph Gathua (National Museums of Kenya, NMK), Mr. Gilbert Kosgei (Kenya Wetlands Biodiversity Research Group, KENWEB) and Dr. Judith Nyunja (Kenya Wildlife Service, KWS) for their support during collection of samples. We are also grateful to KENWEB, NMK, Institut Recherche pour le Développement (IRD) and French Embassy, Nairobi, Kenya for their financial support during this study.

References

  1. Abila R, Barluenga M, Engelken J, Meyer A, Salzburger W (2004) Population-structure and genetic diversity in a haplochromine cichlid fish (corrected) of a satellite lake of Lake Victoria. Mol Ecol 13(9):2589–2602CrossRefPubMedGoogle Scholar
  2. Adams DC, Otarola-Castillo E (2013) Geomorph: an R package for the collection and analysis of geometric morphometric shape data. Methods Ecol Evol 4:393–399CrossRefGoogle Scholar
  3. Agnèse JF, Adépo-Gouréne B, Abban EK, Fermon Y (1997) Genetic differentiation among natural populations of the Nile tilapia Oreochromis niloticus (Teleostei, Cichlidae). Heredity 79:88–96CrossRefPubMedGoogle Scholar
  4. Ashley GM, Maitima MJ, Muasya AM, Owen RB, Driese SG, Hover VC, Renaut RW, Goman MF, Mathai S, Blatt SH (2004) Sedimentation and recent history of a freshwater wetland in a semi-arid environment: Loboi Swamp, Kenya, East Africa. Sedimentology 51:1301–1321CrossRefGoogle Scholar
  5. Atkinson D (1995) Effects of temperature on the size of aquatic ectotherms: exceptions to the general rule. J Therm Biol 20:61–74CrossRefGoogle Scholar
  6. Avella M, Berhault J, Bornancin M (1993) Salinity tolerance of two tropical fishes, Oreochromis aureus and O niloticus I Biochemical and morphological changes in the gill epithelium. J Fish Biol 42:243–254CrossRefGoogle Scholar
  7. Avery S (2010) Hydrological impacts of Ethiopias Omo basin on Kenya’s Lake Turkana water levels and fisheries. Final Report prepared for the African Development Bank, TunisGoogle Scholar
  8. Barlow G (2000) The cichlid fishes: the natures grand experiment in evolution. Persuus publishing Cambridge, MassachusettsGoogle Scholar
  9. Bonhomme V, Picq S, Gaucherel C, Claude J (2014) Momocs: outline analysis using R. J Stat Softw 56(13)Google Scholar
  10. Cerling TE (1979) Palaeochemistry of Plio-Pleistocene Lake Turkana, Kenya. Palaeogeogr Palaeoclimatol Palaeoecol 27:247–285CrossRefGoogle Scholar
  11. Chakrabarty P (2005) Testing conjectures about morphological diversity in cihlids of Lakes Malawi and Tanganyika. Copeia 2:359–373CrossRefGoogle Scholar
  12. Claude J (2008) Morphometrics with R. Springer, New YorkGoogle Scholar
  13. Collyer ML, Novak JM, Stockwell CA (2005) Morphological divergence of native and recently established populations of white sand pupfish (Cyprinodon Tularosa). Copeia 2005(1):1–11CrossRefGoogle Scholar
  14. Collyer ML, Stockwell CA, Adams DC, Reiser MH (2007) Phenotypic plasticity and contemporary evolution in introduced populations: evidence from translocated populations of sand pupfish (Cyprinodon Tularosa). Ecol Res 22:902–910CrossRefGoogle Scholar
  15. Crul RCM (1995) Limnology and hydrology of Lake Victoria, Comprehensive and Comparative Study of Great Lakes. UNESCO/IHP-IV Project M-5.1. UNESCO publishingGoogle Scholar
  16. Day T, Pritchard J, Schluter D (1994) A comparison of two sticklebacks. Evolution 48:1723–1734CrossRefGoogle Scholar
  17. Dryden IE (2013) Shapes: Statistical shape analysis. R package version 1.1–6. Retrieved on march 2006 from http://www.maths.nottingham.ac.uk/~ild/shapes
  18. Dryden IE, Mardia KV (1998) Statistical Shape Analysis. Wiley, ChichesterGoogle Scholar
  19. El-Sayed AFM (1999) Alternative dietary protein sources for farmed tilapia, Oreochromis spp. Aquaculture 179:149–168CrossRefGoogle Scholar
  20. Farmer GJ, Beamish FWH (1969) Oxygen consumption of Tilapia nilotica in relation to swimming speed and salinity. Z Fish Rex Bd Canada 26(11):2807–2821CrossRefGoogle Scholar
  21. Fryer G, Isles TD (1972) The cichlid fishes of the Great Lakes of Africa. Oliver and Boyd, EdinburghGoogle Scholar
  22. Fujimura K, Okada N (2008) Shaping of the lower jaw bone during growth of Nile tilapia Oreochromis niloticus and Lake Victoria cichlid Haplochromis chilotes: a geometric morphometric approach. Develop Growth Differ 50:653–663CrossRefGoogle Scholar
  23. Gagnaire PA, Normandeau E, Pavey SA, Bernatchez L (2013) Mapping phenotypic, expression and transmission ratio distortion QTL using RAD markers in the Lake Whitefish (Coregonus clupeaformis). Mol Ecol 22:3036–3048CrossRefPubMedGoogle Scholar
  24. Georgakopoulou E, Sfakianakis DG, Kouttouki S, Divanach P, Kentouri M, Koumoundouros G (2007) The influence of temperature during early life on phenotypic expression at later ontogenetic stages in sea bass. J Fish Biol 70(1):278–291CrossRefGoogle Scholar
  25. Grammer GL, Slack WT, Peterson MS, Dugo MA (2012) Nile tilapia Oreochromis niloticus (Linnaeus, 1758) establishment in temperate Mississippi, USA: multi-year survival confirmed by otolith ages. Aquat Invasions 7(3):367–376CrossRefGoogle Scholar
  26. Greenwood PH (1974) The cichlid fishes of Lake Victoria, East Africa: the biology and evolution of a species flock. Bull Br Mus Nat Hist 6:1–134Google Scholar
  27. Greenwood PH (1984) African cichlids and evolutionary theories. In: Echelle AA, Kornfield I (eds) evolution of fish species flocks. University of Maine at Orono press, MaineGoogle Scholar
  28. Hay RL (1976) Geology of Olduvai gorge. University of Carlifornia Press, BakelyGoogle Scholar
  29. Herler J, Lipej L, Makovec T (2007) A simple technique for simple imaging of live and preserved small fish specimens. Cybium 31:39–44Google Scholar
  30. Hopson AJ (1982) Lake Turkana; A report on the findings of the Lake Turkana Project 1972 1975, vol 16. Overseas Development Administration, LondonGoogle Scholar
  31. Imre I, McLaughlin RL, Noakes DLG (2002) Phenotypic plasticity in brook charr: changes in caudal fin induced by water flow. J Fish Biol 61(5):1171–1181CrossRefGoogle Scholar
  32. Kerschbaumer M, Sturmbauer C (2011) The utility of geometric morphometrics to elucidate pathways of cichlid fish evolution. Int J Evol Biol 2011:8. doi:10.4061/2011/290245
  33. Mantel N (1967) The detection of disease clustering and a generalized regression approach. Cancer Res 27(2):209–220PubMedGoogle Scholar
  34. Meyer A (1987) Phenotypic plasticity and heterochrony in Cichlasoma managuense (Pisces, Cichlidae) and their implications for speciation in cichlid fishes. Evolution 41:1357–1369CrossRefGoogle Scholar
  35. Meyer A, Kocher TD, Basasibwaki P, Wilson AC (1990) Monophyletic origin of Lake Victoria cichlid fishes suggested by mitochondrial DNA sequences. Nature 347:550–553CrossRefPubMedGoogle Scholar
  36. Ndiwa CT, Nyingi D, Agnèse JF (2014) An important natural genetic resource of Oreochromis niloticus (Linnaeus, 1758) threatened by aquaculture activities in Loboi drainage, Kenya. PLoS One 9(9):e106972CrossRefPubMedPubMedCentralGoogle Scholar
  37. Nelson JS (1994) Fishes of the world, 3rd edn. Wiley, New YorkGoogle Scholar
  38. Nyingi D, De-Vos L, Aman R, Agnèse JF (2009) Genetic characterization of an unknown and endangered native population of the Nile tilapia Oreochromis niloticus (Linnaeus, 1758) (Cichlidae; Teleostei) in the Loboi Swamp (Kenya). Aquaculture 297:57–63CrossRefGoogle Scholar
  39. Ochumba PBO (1990) Massive fish kills within the Nyanza Gulf of Lake Victoria. Kenya Hydrobiol 208:93–99CrossRefGoogle Scholar
  40. Omondi R, Kembenya E, Nyamweya C, Ouma H, Machua SK, Ogari Z (2014) Recent limnological changes and their implication on fisheries in Lake Baringo, Kenya. J Ecol Nat Environ 6(5):154–163CrossRefGoogle Scholar
  41. Owen RB, Crossley R, Johnson TC, Tweddle D, Kornfield I, Davison S, Eccles DH, Engtrom DE (1990) Major low levels of Lake Malawi and their implecation for speciation rates in cichlid fishes. Proc R Lond B 240:519–553CrossRefGoogle Scholar
  42. Padis E (2012) Analysis of phylogenetics and evolution with R, Second edn. Springer, New YorkCrossRefGoogle Scholar
  43. Padis E, Claude J, Strimmer K (2004) APE: analysis of phylogenetics and evolution in R language. Bioinformatics 20:289–290CrossRefGoogle Scholar
  44. Peterson MS, Slack WT, Woodley CM (2005) The occurrence of nonindigenous Nile Tilapia, Oreochromis niloticus (Linnaeus) in coastal Mississippi: ties to aquaculture and thermal effluent. Wetlands 25:112–121CrossRefGoogle Scholar
  45. Price TD, Qvarnstrom A, Irwin DE (2003) The role of phenotypic plasticity in driving genetic evolution. Proc R Soc Lond B 270:1433–1440CrossRefGoogle Scholar
  46. Ramler D, Mitteroecker P, Shama LNS, Wegner KM, Ahnelt H (2014) Nonlinear effects of temperature on body form and developmental canalization in the threespine stickleback. J Evol Biol 27:497–507CrossRefPubMedGoogle Scholar
  47. Rohlf FJ (2010) tpsDig, digitize landmarks and outlines, version 2.16. Department of Ecology and Evolution, State University of New York at Stony BrookGoogle Scholar
  48. Rohlf FJ (2012) tpsUtil, file utility program, version 1.53. Department of Ecology and Evolution, State University of New York at Stony BrookGoogle Scholar
  49. Rohlf FJ, Marcus LF (1993) A revolution in morphometrics. Trends Ecol Evol 8:129–132CrossRefGoogle Scholar
  50. Rousset F (2008) Genepop 007: a complete re-implementation of the genepop software for Windows and Linux. Mol Ecol Resour 8:103–106CrossRefPubMedGoogle Scholar
  51. Sage RD, Loiselle PV, Basasibwaki P, Wilson AC (1984) Molecular versus morphological change among cichlid fishes of Lake Victoria. In: Echelle AA, Kornfield I (eds) Evolution of fish species flocks. University of Maine at Orono press, OronoGoogle Scholar
  52. Salzburger W, Meyer A, Baric S, Verheyen E, Sturmbauer C (2002) Phylogeny of the Lake Tanganyika Cichlid species flock and its relationship to the central and East African Haplochromine cichlid fish faunas. Syst Biol 51(1):113–135CrossRefPubMedGoogle Scholar
  53. Schluter D (2000) The ecology of adaptive radiation. Oxford University Press, OxfordGoogle Scholar
  54. Schluter D, McPhail JD (1993) Character displacement and replicate adaptive radiation. Trends Ecol Evol 8:197–200CrossRefPubMedGoogle Scholar
  55. Schofield CL (1976) Acid precipitation: effects on fish. Ambio 5(5–6):228–230Google Scholar
  56. Schofield PJ, Peterson MS, Lowe MR, Brown-Peterson NJ, Slack WT (2011) Survival, growth and reproduction of nonindigenous Nile tilapia (Oreochromis niloticus (Linnaeus, 1758). I Physiological capabilities to various temperatures and salinities. Mar Freshw Res 62:439–449CrossRefGoogle Scholar
  57. Sfakianakis DG, Leris I, Laggis A, Kentouri M (2011) The effect of rearing temperature on body shape and meristic characters in zebra fish (Danio rerio) juveniles. Environ Biol Fish. doi:10.1007/s10641-011-9833-z Google Scholar
  58. Stewart TA, Albertson RC (2010) Evolution of unique predatory feeding apparatus: functional anatomy, development and a genetic locus for jaw laterality in Lake Tanganyika scale-eating cichlids. BMC Biol. doi:10.1186/1741-7007-8-8 PubMedPubMedCentralGoogle Scholar
  59. Takahashi R, Watanabe K, Nishida M, Hori M (2007) Evolution of feeding specialization in Tanganyikan scale-eating cichlids: a molecular phylogenetic approach. BMC Evol Biol 7:195. doi:10.1186/1471-2148-7-195 CrossRefPubMedPubMedCentralGoogle Scholar
  60. Trewavas E (1983) Tilapiine fishes of the genera Sarotherodon, Oreochromis and Danakilia. British Museum (Natural History), LondonGoogle Scholar
  61. Turner GF, Seehausen O, Knight ME, Allender C, Robinson RL (2001) How many species of cichlid fishes are there in African lakes? Mol Ecol 10(3):793–806CrossRefPubMedGoogle Scholar
  62. Welcomme RL (1967) Observations on the biology of Introduced species of Tilapia in Lake Victoria. Revue Zool Bot Afr 76(3–4):249–279Google Scholar
  63. Wurts WA, Durborow RM (1992) Interaction of pH, carbon dioxide, alkalinity and hardness in fish ponds. SRAC Publication No 464Google Scholar

Copyright information

© The Author(s) 2016

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors and Affiliations

  • Titus Chemandwa Ndiwa
    • 1
    • 2
  • Dorothy Wanja Nyingi
    • 1
    • 2
  • Julien Claude
    • 3
  • Jean-François Agnèse
    • 4
  1. 1.Department of Clinical Studies, College of Agriculture and Veterinary SciencesUniversity of NairobiNairobiFrance
  2. 2.Kenya Wetlands Biodiversity Research TeamNational Museums of KenyaNairobiKenya
  3. 3.Département Forme, UMR IRD 226 CNRS 5554, Institut des Science de l’EvolutionUniversité de Montpellier 2MontpellierFrance
  4. 4.Département Conservation et Domestication, UMR IRD 226 CNRS 5554, Institut des Science de l’EvolutionUniversité de Montpellier 2MontpellierFrance

Personalised recommendations