Fluorescent compounds which are currently used as retrograde tracers were tested in the cat peripheral motor system and compared with horseradish peroxidase (HRP). The tracers were either injected into forelimb muscles or applied to the proximal end of transected forelimb nerves. The remaining muscles of the limb has been carefully denervated. Following intramuscular injection all fluorescent compounds labeled spinal cord motoneurons, the DAPI compounds labeled endothelial cells in addition. In the nerve application mode tracer positive motoneurons were only observed when propidium iodide (PI) and the DAPI compounds were used, whereas bisbenzimide (BB), nuclear yellow (NY) and primuline did not label any cells. The fluorescence of BB labeled motoneurons were predominantly located in the cytoplasma. NY positive motoneurons showed a different localization of the fluorescent label between the different neurons of the same motornucleus: in some neurons it was exclusively located in the nucleus, in others predominantly in the cytoplasma, in the majority in both compartments. The intracellular distribution of the BB and the NY label was independent of the pH of the fixation fluid. The fluorescent tracers labeled the motoneuronal cell columns in their complete rostrocaudal extent and in a position identical to the one obtained with HRP. However, some substances (PI, fast blue) labeled less motoneurons of a motornucleus than did HRP, none of the fluorescent tracers labeled more. The results are discussed under several aspects: use of the investigated fluorescent compounds as single tracers; use of several tracers in the same animal to map collateral projections of one and the same neuron; use of several tracers in the same animal to establish the topographical relation between several independent neuronal populations.