Vidic and Haaf (2004) questioned the idea that infants use head information to categorize cats as distinct from dogs (Quinn & Eimas, 1996) and argued instead that the torso region is important. However, only null results were observed in the critical test comparisons between modified and unmodified stimuli. In addition, a priori preferences for the paired test stimuli were not assessed, thus leaving open the possibility that novel category preferences for unmodified stimuli could have been blocked by spontaneous preferences for modified stimuli. Moreover, only a single cat-dog pairing and set of pig parts were used as test stimuli, thereby raising the issue of whether the infants could have been responding to idiosyncratic featural differences between particular exemplars rather than features diagnostic of whole categories. The evidence does not support the conclusion that the torso is important for infants' categorization of cats versus dogs.