Trenching (Tr), root biomass regression (RR), and root excising (RE) methods were used to estimate the contribution of root (RR) and heterotrophic (HR) respiration to soil respiration (SR) in a cool-temperate deciduous forest in central Japan. The contribution ratios of RR to SR were 23 % (−16 to 46 %), 11 % (−19 to 61 %), and 115 % (20 to 393 %), as estimated by the Tr, RR, and RE methods, respectively. The contribution ratio showed clear seasonal variation with high values in summer for the Tr method, while they were undetectable for the RR and RE methods because of some methodological problems. These results suggest the Tr method is the best of the three methods used to estimate the contribution ratio of RR and HR to SR in the forest. Annual SR, RR, and HR rates, estimated by the Tr method, were 479, 369, 110 gC m−2 year−1, respectively. The seasonal variation of SR was mainly influenced by HR (77 %) throughout the year, while the influence of RR on SR was strongest in summer (46 %). This effect occurred because RR (Q 10 = 7.5) is more sensitive to temperature than HR (Q 10 = 3.2). Also, the contribution of fine RR to total RR was higher than that of coarse RR because of high respiratory activity (Q 10 and R 10) as well as the large biomass of fine roots. These results suggest that each component of SR responds differently to the same environmental factors and their relative influence on SR changes across the seasons.