Geometric morphometrics approach to the study of ecomorphological divergence among cichlid fishes from East African Great Lakes
Cichlid fishes from the East African Great Lakes are well known for resource partitioning along the trophic, spatial, temporal and behavioural axes, which subsequently enhances their coexistence. However, attempts to investigate the role of morphology in such resource partitioning, has been partly hampered by the lack of robust techniques to quantify morphological features that can be linked to function. Hence in this thesis, I used the novel technique, Geometric Morphometrics (GM), to quantify morphological variations, within and between species and/or trophic guilds, in order to elucidate the role of morphology in resource partitioning. Using same technique, the hypothesis of having ecomorphologically equivalent species between lakes was also tested, and lastly, a possibility of using GM data in phylogenetic systematics was investigated. Taking into account the geometry of the organism, statistical robustness and ability to visualize shape changes through deformation grids, are some of the factors that make GM the best technique to use.
Firstly, GM was used to quantify body shape and/or size variation within and among three coexisting Mbuna Petrotilapia species from Lake Malawi which are believed to differ in colouration only. MANOVA revealed significant differences in body shape among species but not between sexes. Pairwise comparisons based on Mahalanobis distances revealed that P. nigra was significantly different from P. genalutea and P. “mumbo blue”, but the latter two were not significantly different. Deformation grids revealed that P. nigra had longer and deeper head, shorter gape, shallower body depth than its congeners. ANOVA performed on centroid size revealed that males were bigger than females in all species. The observed sexual size dimorphism and interspecific morphological variations were discussed in terms of sexual and ecological selection models, respectively. Secondly, the hypothesis that Petrotilapia spp. is [sic] not morphologically different from Lake Tanganyika’s Petrochromis spp. was tested. Such hypothesis was rejected based on the significant differences in body shape among species revealed through MANOVA, and ordination along the first two CV axes showed clear separation of the two genera in the morphospace with Petrochromis fasciolatus being intermediate. The splines generated revealed that Petrochromis spp. had deeper anterior body, larger gape, shallower caudal peduncle than Petrochromis spp. [sic] However, posthoc multiple comparisons revealed nonsignificant differences between P. fasciolatus and Petrotilapia genalutea and P. “mumbo blue” suggesting that these three species display best example of convergent evolution.
Thirdly, body shape variation between two non-mbuna sympatric benthophagous species
from Lake Malawi was investigated. Significant differences were revealed and the
splines generated indicated that Ctenopharynx pictus had longer and deeper head,
larger gape and shorter abdominal cavity (which signified shorter intestine) than
Otopharynx sp. “heterodon nankhumba”. Such variations in head morphology and
intestinal length were correlated with ecological separation, possibly facilitating
coexistence of the two species. Fourthly, body shape variations were examined within
Lake Malawi’s trophic guilds, algal and zooplankton feeders, in order to understand
the role of cichlid body design in partitioning of resources along the trophic and
spatial axes. MANOVA revealed significant differences within zooplankton feeders in
which Copadichromis borleyi had shorter gape, smaller eyes, shorter caudal
peduncle relative to Ctenopharynx pictus, and for within algal feeders,
Labeotropheus fuelleborni had shorter subterminal gape, shorter head than
Petrotilapia genalutea. The fact that most pronounced variation between
species within trophic guild was found in the head region called for further
examination of the head region bony elements, viz; neurocranium, lower jaw and
premaxilla. Significant differences were revealed whereby Ctenopharynx pictus
had longer neurocranium, longer ventral-directed vomer, larger orbit, shorter
ascending arm, more compressed articular bone relative to Copadichromis borleyi.
And between algal feeders, L. fuelleborni had shorter neurocranium, smaller
orbit, ventral-directed vomer, longer ascending arm, increased height of articular
process than P. genalutea. These differences, both in body shape and trophic
morphology, were discussed in relation to trophic and feeding microhabitat
differentiation that subsequently enhances coexistence among the species. Finally,
the applicability of GM data in phylogenetic systematics was investigated by comparing
dendrograms generated from AFLP molecular marker and landmark-based data. Both the N-J
and UPGMA dendrograms, generated from AFLP data, supported separate clustering of mbuna
and non-mbuna species. GM-generated UPGMA dendrogram was not congruent to either
AFLP-generated N-J or UPGMA dendrograms, thus revealing lack of GM data to pick
phylogenetic signal.