The effect of changing feeds on individual feed intake and feeding and dominance ranks in groups of African catfish Clarias gariepinus was investigated. Following feeding on a commercial feed groups (n = 3) of six African catfish were either fed fish meal (FM42) or maize gluten (MG35) based feeds for 5 days before being switched to the other feed for 5 days. Energy intake was significantly lower on FM42 than on MG35, dry matter intake and protein intake were significantly lower on FM42 than on commercial feed and this occurred whether FM42 was fed first or second. There were no significant differences between intake of MG35 and commercial feed. Thus, the action of changing the feed on its own did not affect feed intake since the decrease was shown to be feed‐specific to FM42. Six types of agonistic behaviours were identified and used to assign dominance rank. There were no correlations between dominance and feeding ranks. This was due to non‐linear hierarchies with one dominant fish in each group. Feeding ranks were more stable when feeding MG35 than FM42. Feeding rank stability (Kendall's coefficient of concordance) was significant in five out of six groups fed MG35 (compared with three out of six fed FM42). Feeding rank stability was higher in five out of the six groups when they were fed MG35 than when the same group was fed FM42. The experiment provided evidence that the introduction of a novel feed can, but does not necessarily, alter feed intake and that feed can influence the stability of feeding ranks.