Pitfall trap capture for Sitophilus oryzae (L.), S. zeamais (Motschulsky), S. granarius (L.), Tribolium confusum (Duval), T. castaneum (Herbst); Oryzaephilus surinamensis (L.), Cryptolestes pusillus (Schonherr), Rhyzopertha dominica (Fabr.), and Prostephanus truncatus (Horn) in millet was assessed by visual and time‐lapse video recordings. The behaviour of different beetle species in arenas containing millet was monitored over 24 h and the frequency of encounters with the trap rim resulting in capture recorded. The capture efficiency of four types of pitfall traps (i.e. polystyrene, polythene, glass and tin‐plated steel can) with rims exposed or submerged below the millet surface level were compared. Capture was related to beetle size, locomotory rate, and beetle behaviour at the trap rim as well as trap design and placement. The lighter and smaller species were least captured. Glass jars were more effective than plastic and metal containers. Traps placed with their rims submerged below the grain surface level were more efficient than those with rims exposed. Capture rate was unrelated to trap size. The frequency of encounters with trap rims was not correlated with capture rate. Three types of avoidance behaviours at the trap rims i.e. probing, skirting and spontaneous retreat, were related to capture rate, spontaneous retreat being the most effective escape mechanism and probing least.