Rats with hippocampal aspiration lesions and controls were trained on delayed nonmatching to sample with small complex goal boxes, presented trial uniquely. A series of experiments then used pairs of large or small boxes, presented repeatedly. The lesions impaired choice accuracy when the rats were tested with large empty boxes but not when small boxes containing 3-dimensional objects were used. There was a comparable impairment when the rats were tested with pairs of large complex boxes, which contained arrays of objects, identical to those used in the smaller boxes but necessarily spaced further apart. Subsequent experiments revealed that the lesion deficit with large boxes was reduced by insertion of a continuous line of distinctive objects and eliminated by trial-unique presentation of large boxes. The results are discussed in terms of (non) spatial accounts of hippocampal function and the compensatory effects of novel object cues. We conclude that, for hippocampal rats, spatial cues, although useless, can nonetheless be profoundly disruptive.