Mean bioluminescence time-series data from young (10-day), middle-aged (30-day) and old (50-day) flies in DD. Bioluminescence counts per second (CPS) is plotted as average of raw counts at 30-minute intervals. Data are shown as mean ± SEM with error bars at each time point. 10-day cohort equals black lines and columns, 30-day cohort equals mid-grey lines and columns, and 50-day cohort equals light-grey lines and columns. (A) 8.0-luc (reporting PER in the dorsal neurons, central clock). The 30-day cohort shows higher PER amplitude and acrophase than 10-day and 50-day (p < 0.001 in all), and there is no difference between 10-day and 50-day (p = 1.00 in both). (B) XLG-luc (reporting PER in the whole animal, peripheral clocks). The 30-day and 50-day cohort show lower PER amplitude and acrophase compared to 10-day (p < 0.001 in all), 50-day is lower than 30-day (p = 0.02 in amplitude and p = 0.001 in acrophase). (C) tim-luc (reporting TIM in the whole animal, peripheral clocks). The 30-day and 50-day cohort show lower TIM amplitude and acrophase compared to 10-day (p < 0.001 in all), and 50-day is lower than 30-day in acrophase (p < 0.001) but not in amplitude (p = 0.70).