The non-Langerhans histiocytoses, a nosologic category to which juvenile xanthogranuoma (JXG) belongs, represent a heterogenous collection of disorders related to the monocyte/macrophage lineage. The dermal dendrocyte was previously proposed as the cell of origin for JXG on the basis of Factor XIIIa reactivity, a suggestion that does not fully explain the occasional xanthogranulomatous proliferations localizing exclusively to extracutaneous sites. This study applies a panel of recently developed immunohistochemical markers to JXGs and relates the phenotype of this process to new concepts of monocyte/dendritic cell ontogeny. Twenty-seven JXG, ten dermatofibromas (DF), and ten age-matched normal skin specimens were stained using standard immunohistochemistry methods, and all JXGs were fascin+ and CD68+, although 26 of 27 were reactive for HLA-DR, 25 of 27 for Factor XIIIa, 25 of 27 for LCA, 21of 27 for CD4, and 8 of 27 for polyclonal s100. Six of those eight polyclonal S100+ cases were also reactive for monoclonal S100. None of those cases was reactive for CD1a, CD3, CD21, CD34, or CD35. Eight of ten dermatofibromas were FXIIIa+; all were negative for HLA-DR, LCA, CD4, and polyclonal s100. In controls, fascin+ dendritic cells were present but did not stain for Factor XIIIa, S100, or CD4. Based on the morphologic and phenotypic overlap of the lesional cells in JXGs and plasmacytoid monocytes, it would appear that the plasmacytoid monocyte might be considered the putative normal counterpart of the major cellular population of JXGs, a proposal that helps explain the extra-cutaneous, visceral, and soft tissue location that have been reported for occasional cases of JXG. We would also conclude that neither Factor XIIIa-nor S100+ results should preclude the diagnosis of JXG, and find that reactivity for CD4 and LCA may be used to distinguish JXG from DF when the latter is heavily lipidized or the former is not.