The respective roles in the imprinting process of parts (IMHV) of the left and right hyperstriatum ventrale of the chick brain were examined by destroying first one and then the other IMHV in a two-stage operation. One hundred and eight chicks were dark-reared to ≃ 19 h post-hatch and exposed to a training stimulus for 2 h. Chicks were anaesthetised ≃ 3 h after the end of training. Lesions were placed in either (i) right IMHV (N = 18 birds), (ii) left IMHV (N = 18) or (iii) left or right hyperstriatum accessorium (HA; N = 18). Fifty-four chicks served as sham-operated controls. Chicks were returned to the dark incubator, and, 15–20 h after the operation, the chicks’ approach towards the training stimulus and to a second novel stimulus was measured (Test 1). After this test the chicks were again anaesthetised and a second lesion was made, this lesion being placed in the corresponding structure (IMHV or HA) of the hemisphere contralateral to that with the first lesion. The chicks’ preferences were measured 15–20 h later (Test 2). In Test 1, all birds strongly preferred the training stimulus. In Test 2, sham-operated controls and HA chicks continued to prefer the training stimulus as did chicks with the initial lesion in the left IMHV. However, chicks with the initial lesion in the right IMHV failed to show a preference for the training stimulus. Thus, if the right IMHV is destroyed first the presence of the left IMHV is crucial for retention. In contrast, if the left IMHV is destroyed first the presence of the right IMHV is not crucial for retention: chicks continue to prefer the training stimulus after the right IMHV has been lesioned. In these circumstances, therefore, some region outside IMHV takes on a memory function. The results imply that at least two memory systems are formed during imprinting. One of these involves the left IMHV, the other does not. The putative second system is fully able to sustain recall in the normal chicks by ≃ 26 h after training: if bilateral lesions to IMHV (N = 28 chicks) are made at this time, retention, measured 15–20 h later, is not significantly different from that of sham-operated control chicks (N = 25).