In a serial learning experiment in which the first item was made unique, the controls learned the list in the usual manner. For the experimental group, each trial began at a different, randomly chosen syllable. The results show that (a) acquisition is impaired and there is a different curve for the experimental group; (b) the unique item has an anchoring effect; and (c) a plot of errors within each trial shows no serial position effect (SPE) in the experimental group. The results are interpreted to indicate that: (a) anchoring accounts for a substantial portion of the SPE, and (b) recall shows no SPE in this paradigm.