The partitioning of soil respiration rates into the component processes of rhizospheric respiration (because of live roots and those microorganisms that subsist on root exudations) and heterotrophic respiration (because of decomposer microorganisms that subsist on the oxidation of soil organic matter) is difficult to accomplish through experimental observation. In order to minimize disturbance to the soil and maximize preservation of the natural relationships among roots, rhizospheric microorganisms, and decomposers, we conducted a girdling experiment in a subalpine forest dominated by lodgepole pine trees. In two separate years, we girdled trees in small forest plots (5–7 m in diameter) and trenched around the plots to sever invading roots in order to experimentally stop the transport of photosynthate from needles to roots, and eliminate rhizospheric respiration. Soil respiration rates in plots with trees girdled over 1 year prior to measurement were higher than those in plots with trees girdled 2–3 months prior to measurement. These results suggest that any stimulation of respiration because of the experimental artifact of fine root death and addition of labile carbon to the pool of decomposer substrates is slow, and occurs beyond the first growing season after girdling. Compared with control plots with nongirdled trees, soil respiration rates in plots with girdled trees were reduced by 31–44% at the mid-summer respiratory maximum. An extreme drought during one of the 2 years used for observations caused greater reductions in the heterotrophic component of soil respiration compared with the rhizospheric component. In control plots, we observed a pulse in K2SO4-extractable carbon during the spring snowmelt period, which was absent in plots with girdled trees. In control plots, soil microbial biomass increased from spring to summer, coincident with a seasonal increase in the rhizospheric component of soil respiration. In plots with girdled trees, the seasonal increase in microbial biomass was lower than in control plots. These results suggest that the observed seasonal increase in rhizospheric respiration rate in control plots was because of an increase in rhizospheric microbial biomass following ‘soil priming’ by a spring-time pulse in dissolved organic carbon. Winter-time, beneath-snow microbial biomass was relatively high in control plots. Soil sucrose concentrations were approximately eight times higher during winter than during spring or summer, possibly being derived from the mechanical damage of shallow roots that use sucrose as protection against low-temperature extremes. The winter-time sucrose pulse was not observed in plots with girdled trees. The results of this study demonstrate that (1) the rhizospheric component of soil respiration rate at this site is significant in magnitude, (2) the heterotrophic component of soil respiration rate is more susceptible to seasonal drought than the rhizospheric component, and (3) the trees in this ecosystem exert a major control over soil carbon dynamics by ‘priming’ the soil with sugar exudates during the late-spring snowmelt period and releasing high concentrations of sucrose to the soil during winter.