Thirty-two honey bee colonies, Apis mellifera bandasii, 16 in traditional basket hives and 16 in Langstroth hives were used to assess propolis production for 19 months in the highlands of Ethiopia. The colonies of each hive type were randomly and evenly assigned to treatment groups; one induced to collect more propolis by creating openings in the hives and as control groups. Significant differences (F_1,302 = 60.14, P < 0.0001) in propolis yields of 24.2 ± 22.5 g/hive and 9.5 ± 5.8 g/hive were obtained for induced and control group Langstroth hives, respective­ly. Traditional basket hives were also significantly different (F_1,302 = 119.8, P < 0.0001) for propolis yields of 12.7 ± 8.6 g per hive and 4.8 ± 2.4 g per hive for induced and control groups, respectively. The results show that simple induction of colonies for more propolis pro­duction in both traditional basket and Langstroth hives is possible. There were no sig­nificant effects (F_1,30 = 0.20, P = 0.6545) on honey yield between induced and control groups for either hive type. Significant propolis yield variations within groups and between months were also observed.