To investigate the ability ofDrosophila serrata to adapt to thermal conditions over winter at the species southern border, replicate lines from three source locations were held as discrete generations over three years at either 19‡C (40 generations) or temperatures fluctuating between 7‡C and 18δC (20 generations). Populations in the fluctuating environment were maintained either with an adult 0‡C cold shock or without a shock. These conditions were expected to result in temperature-specific directional selection for increased viability and productivity under both temperature regimes, and reduced development time under the fluctuating-temperature regime. Selection responses of all lines were tested under both temperature regimes after controlling for carry-over effects by rearing lines in these environments for two generations. When tested in the 19‡C environment, lines evolving at 19‡C showed a faster development time and a lower productivity relative to the other lines, while cold shock reduced development time and productivity of all lines. When tested in the fluctuating environment, productivity of the 7–18‡C lines selected with a cold shock was relatively lower than that of lines selected without a shock, but this pattern was not observed in the other populations. Viability and body size as measured by wing length were not altered by selection or cold shock, although there were consistent effects of source population on wing length. These results provide little evidence for temperature-specific adaptation inD. serrata —although the lines had diverged for some traits, these changes were not consistent with a priori predictions. In particular, there was no evidence for life-history changes reflecting adaptation to winter conditions at the southern border. The potential forD. serrata to adapt to winter conditions may therefore be limited.