This study determined whether the visual characteristics of a familiar (imprinted) model or the auditory characteristics of the species maternal call are more important in determining the maternal preferences of visually imprinted ducklings. Domestic mallard (Peking) ducklings were visually imprinted to a stuffed model of a mallard duck during a 30-min following trial at 24 hr after hatching. Simultaneous choice tests between the familiar mallard model and an unfamiliar red-and-white striped box at 48 hr and 72 hr confirmed the efficacy of the imprinting procedure: When both models were silent, subjects preferred to follow the familiar mallard model. However, when a recording of the mallard maternal assembly call was played from a speaker mounted inside the red box, subjects imprinted to the mallard preferred to follow the unfamiliar box rather than the familiar mallard model (Experiment 1). That preference was not due merely to the audiovisual stimulation provided by athe box, since when a recording of intermittent tones was played from the mallard model, subjects imprinted to the mallard still preferred to follow the red box emitting the mallard call (Experiment 2). Playing only the tones from the red box disrupted the stability of the subjects' imprinted preferences between the first and second tests but did not produce a preference for the box (Experiment 3). These results show that the mallard maternal call is more important than visual experience with an inanimate model in determining the maternal preferences of visually imprinted Peking ducklings.