In this study we examined the hypothesis that the intermediolateral cell column (IML) of the thoracic spinal cord, the nucleus from which preganglionic sympathetic neurons originate, provides an anatomical substrate through which selective regulation of sympathetic nervous system targets is accomplished. Preganglionic sympathetic neurons of rats were retrogradely labeled by the simultaneous exposure of the cervical sympathetic trunk (CST) and the adrenal medulla to Fluoro-Gold and True blue, contrasting fluorescent dyes. Retrograde labeling from these sites revealed 2 populations of sympathetic preganglionic neurons in IML whose distribution overlapped between segments T1 and T4. In regions where these 2 groups of retrogradely labeled neurons overlapped, sympathoadrenal preganglionic (SAP) neurons occupied the most lateral aspect of the nucleus. It was also determined whether individual retrogradely labeled neurons within these two groups sent axon collaterals to both the CST and adrenal medulla. Diamidino yellow, a fluorescent retrograde tracer dye that labels only nuclei, was substituted for Fluoro-Gold and used in combination with True blue to simultaneously label preganglionic sympathetic neurons projecting to either the CST or adrenal medulla. No double-labeled cell bodies were observed in spinal cords of rats treated in this manner. Thus it appeared that the efferent projections of these 2 cell populations in IML were target-specific. Immunohistochemical analysis of the relationship between nerve fibers in the IML and preganglionic sympathetic neurons was also undertaken in an attempt to classify further these 2 populations of sympathetic preganglionic neurons. Equal proportions of identified CST and SAP neurons appeared to be apposed by varicosities immunoreactive for either somatostatin or serotonin. On the other hand, when the comparison was based on whether oxytocin-immunoreactive varicosities appeared to appose these 2 populations of retrogradely labeled sympathetic neurons, a highly significant difference was revealed. That is, oxytocin-immunoreactive fibers and terminals appeared to avoid SAP neurons. Thus these data support the hypothesis that an anatomical substrate exists in spinal cord IML whereby selective regulation of sympathetic nervous system targets may be mediated. Moreover, the lack of oxytocin-immunoreactive varicosities apposing SAP neurons in IML suggests that if the paraventricular nucleus innervates SAP neurons in IML, it does so via a population of neurons that do not use oxytocin as a neurotransmitter.