Abstract:
Morphometric characterization using truss network systems has been
widely used for fish identification and stock identification. In Sri Lanka,
differentiation of Oreochromis mossambicus and O. niloticus, the two major
species in reservoir fishery, using conventional morphometric characteristic
and phenotypic variations is problematic, due to the reasons that both species
share somewhat similar body shape characteristics and that the populations of
the two species have mixed characteristics as a result of introgressive
hybridization. In the present study, an attempt was made to employ truss
network analysis to differentiate O. mossambicus and O. niloticus. Truss
measurements of 100 specimens of O. mossambicus from Negombo lagoon
and 100 specimens of O. niloticus from Baira lake were analyzed using
principal component analysis (PCA). Before the analysis, two approaches of
data transformation were employed to minimize to eliminate any size effect in
the data set of truss measurements. It was concluded that the data
standardization method using a geometric mean regression techniques on the
logarithms of truss length and total length was more effective for
morphometric differentiation of the two cichlid species than using ratio of truss
length to standard length for data transformation.
