Growth of African catfish (Clarias garipinus) is characterized by large size heterogeneity resulting in poor growth performance and cannibalism. Strategies that limit size heterogeneity in C. gariepinus culture are therefore advocated. This study determined the effects of grading frequency on the growth performance and cannibalism of C. gariepinus in tank culture system. Three treatments designated as G0 (no grading), G2 (grading every 2 weeks) and G4 (grading every four weeks) in a Completely Randomized Design (CRD) was applied in triplicate. The tanks were stocked with 200 fingerlings each. Sampling was conducted weekly to measure length and weight of fish. Mortalities were recorded and dead fish observed under dissecting microscope to ascertain that it is due to cannibalism. Growth performance in terms of final mean weight, weight gain, specific growth rate and food conversion ratio were significantly (P < 0.05) affected by frequency of fish grading. The C. gariepinus graded every two weeks grew better than those graded every four weeks and were all above the no grading treatment. Significant changes in fish heterogeneity were observed after day 42. The size variation was significantly affected by grading frequency. The mean TL of fish graded every 2 weeks was the highest, followed by grading every four weeks and finally the no grading treatment was the lowest. Fish size heterogeneity (CV% and skewness) was consistently the highest in C. gariepinus in no grading treatment followed by fish where grading was done every four weeks while it was lowest in treatment where feeding was done every two weeks. Mortality owing to cannibalism was affected by grading frequency where highest cannibalism mortality occurred in C. gariepinus where there was no grading followed by grading every four weeks and intermediate at grading frequency every four weeks. Meanwhile other mortality by other causes such as such as wounds and suffocation in fish did not differ with grading frequency. Keywords: Cannibalism, Coefficient of variation, Fish grading, Fish growth, Growth heterogeneity