Soil respiration in tropical forests is a major source of atmospheric CO2. The ability to partition soil respiration into its individual components is becoming increasingly important to predict the effects of disturbance on CO2 efflux from the soil as the responses of heterotrophic and autotrophic respiration to change are likely to differ. However, current field methods to partition respiration suffer from various methodological artefacts; root–rhizosphere respiration is particularly difficult to estimate. We used trenched subplots to estimate root–rhizosphere respiration in large-scale litter addition (L+), litter removal (L−) and control (CT) plots in a lowland tropical semi-evergreen forest in Panama. We took a new approach to trenching by making measurements immediately before-and-after trenching and comparing them to biweekly measurements made over one year. Root–rhizosphere respiration was estimated to be 38%, 17% and 27% in the CT, L+, and L− plots, respectively, from the measurements taken immediately before and one day after trenching in May–June 2007. Biweekly measurements over the following year provided no estimates of root–rhizosphere respiration for the first seven months due to decomposition of decaying roots. We were also unable to estimate root–rhizosphere respiration during the dry season due to differences in soil water content between trenched and untrenched soil. However, biweekly measurements taken during the early rainy season one year after trenching (May–June 2008) provided estimates of root–rhizosphere respiration of 39%, 24% and 36% in the CT, L+, and L− plots, respectively, which are very similar to those obtained during the first day after trenching. We suggest that measurements taken immediately before and one day after root excision are a viable method for a rapid estimation of root–rhizosphere respiration without the methodological artefacts usually associated with trenching experiments.