Autoradiography with 125I-neurotensin in normal and enucleated hamsters was used to define the distribution of receptors for this peptide in the superficial layers of the superior colliculus (SC). Neurotensin binding sites were densely distributed in the stratum griseum superficiale (SGS), and results from the enucleated animals indicated that they were not located on retinal axons. The effects of neurotensin on individual superficial layer cells were tested in single-unit recording experiments. Neurotensin was delivered via micropressure ejection during visual stimulation (n = 75 cells), or during electrical stimulation of either the optic chiasm (OX; n = 47 cells) or visual cortex (CTX; n = 29 cells). In comparison with control values, application of neurotensin decreased visual responses of all SC cells tested to 54.1 ± 34.9% (mean ± standard deviation; range of decrement 7.5 to 100%; nine cells showed no effect or an increase in visual activity, which for four of these was ≥30%). Neurotensin application also reduced responses to electrical stimulation of either OX or CTX, respectively, to 65.8 ± 36.5% of control values (range of decrement 2.6 to 97.4%; 12 neurons showed a weak increment ≤ 30%) and 68.0 ± 38.5% (range of decrement 3.3 to 100%; five cells showed no effect or an increment, in one case ≥ 30%). Of the 25 neurons tested with both OX and CTX stimulation, the correlation of evoked response suppression by neurotensin was highly significant (r = 0.70; P < 0.001). This suggests that the suppressive effects of neurotensin were common to both pathways. To test whether the inhibitory effects of neurotensin were presynaptic or postsynaptic, Mg2+ ions were ejected iontophoretically to abolish synaptic responses, and the neurons (n = 16) were activated by iontophoresis of glutamate and then tested with neurotensin. Neurotensin reduced the glutamate-evoked responses to an average 59.3 ± 37.9% of control values (range 2.3 to 92.5%; one cell showed an increment >30%). This result suggests that the site of action of neurotensin is most likely postsynaptic.