Primary afferent projections from cutaneous afferents in the forelimb and hindlimb digits to the dorsal horn (DH) were examined using 4 tracers: (1) 25% free horseradish peroxidase (HRP), (2) 2.5% wheat-germ agglutinin conjugated to horseradish peroxidase (WGA-HRP), (3) a mixture of 25% free HRP and 2.5% WGA-HRP (WGA-HRP/HRP) or (4) 0.1% HRP conjugated to cholera toxin (B-HRP). The tracer was injected intracutaneously into the digits. Three to 4 days later, the rats were perfused transcardially, transverse sections (60-microns thick) were cut and the HRP was reacted using the tetramethyl benzidine (TMB) method. The location of the label was reconstructed by camera lucida drawings. In rats which received an injection of HRP alone, no label was detected in the DH. Rats injected with WGA-HRP had projection patterns similar to those injected with WGA-HRP/HRP. Patterns of labelling with WGA-HRP differed markedly from those with B-HRP. WGA-HRP labelled cutaneous afferents projecting to Rexed's laminae I-III, with the densest label in lamina II; in contrast, B-HRP labelled cutaneous afferents projecting to laminae II-V, with the densest label in laminae III-IV. These results indicate that, for cutaneous primary afferents projecting to the DH, WGA-HRP and B-HRP labelled different subpopulations of fibers, with the B-HRP-labelled subpopulation biased toward afferents of larger diameter. Rostrocaudally, the extent of the densest fiber projections, whether labelled by WGA-HRP or by B-HRP, was essentially the same, but the extent of the less densely labelled projections was much greater with B-HRP than with WGA-HRP. Comparisons of the projection maps from each of the five digits, using either WGA-HRP or B-HRP, indicated that, as seen in transverse sections through the DH, there was extensive overlapping among the labelled cutaneous afferent fibers from adjacent, or even non-adjacent digits.