The fine root systems of three tropical montane forests differing in age and history were investigated in the Cordillera Talamanca, Costa Rica. We analyzed abundance, vertical distribution, and morphology of fine roots in an early successional forest (10–15 years old, ESF), a mid-successional forest (40 years old, MSF), and a nearby undisturbed old-growth forest (OGF), and related the root data to soil morphological and chemical parameters. The OGF stand contained a 19 cm deep organic layer on the forest floor (i.e., 530 mol C/m2), which was two and five times thicker than that of the MSF (10 cm) and ESF stands (4 cm), respectively. There was a corresponding decrease in fine root biomass in this horizon from 1128 g dry matter/m2 in the old-growth forest to 337 (MSF) and 31 g/m2 (ESF) in the secondary forests, although the stands had similar leaf areas. The organic layer was a preferred substrate for fine root growth in the old-growth forest as indicated by more than four times higher fine root densities (root mass per soil volume) than in the mineral topsoil (0–10 cm); in the two secondary forests, root densities in the organic layer were equal to or lower than in the mineral soil. Specific fine root surface areas and specific root tip abundance (tips per unit root dry mass) were significantly greater in the roots of the ESF than the MSF and OGF stands. Most roots of the ESF trees (8 abundant species) were infected by VA mycorrhizal fungi; ectomycorrhizal species (Quercus copeyensis and Q. costaricensis) were dominant in the MSF and OGF stands. Replacement of tropical montane oak forest by secondary forest in Costa Rica has resulted in (1) a large reduction of tree fine root biomass; (2) a substantial decrease in depth of the organic layer (and thus in preferred rooting space); and (3) a great loss of soil carbon and nutrients. Whether old-growth Quercus forests maintain a very high fine root biomass because their ectomycorrhizal rootlets are less effective in nutrient absorption than those of VA mycorrhizal secondary forests, or if their nutrient demand is much higher than that of secondary forests (despite a similar leaf area and leaf mass production), remains unclear.