Phase resetting responses to daily temperature gradients are relatively slow and cumulative. After entrainment to 12-h dark/12-h light cycles (lights on from noon to midnight at constant 25°C for ≥4 days) groups of Canton-S male and female flies were exposed to 0, 1, 3, or 5 'anti-phase' daily temperature gradients (16°C at midnight to 28°C at noon to 16°C at midnight in constant darkness; see profile Z7T4C4 in Additional file 2) before being released into free running conditions at 25°C in constant darkness. The phase of peak and offset of activity during the free running circadian period length (τ) was determined from median activity records and plotted relative to the number of temperature cycles included in the experiment (A). The corresponding circadian activity profiles (mean ± S.E.M.) for female and male Canton-S flies are represented in panels (B) and (C), respectively, with individual experiments identified by the number of included temperature cycles (immediately above the peak) and increasingly lighter shades of gray. The shaded and black-and-white bars underneath the activity profiles represent the daily temperature gradient and light (L)/dark (D) treatments, respectively, used during previous entrainment.