Measurement of soil-respired CO₂ at high temporal resolution and sample density is necessary to accurately identify sources and quantify effluxes of soil-respired CO₂. A portable sampling device for the analysis of δ(13)C values in the field is described herein. CO₂ accumulated in a soil chamber was batch sampled sequentially in four gas bags and analysed by Wavelength-Scanned Cavity Ring-down Spectrometry (WS-CRDS). A Keeling plot (1/[CO₂] versus δ(13)C) was used to derive δ(13)C values of soil-respired CO₂. Calibration to the δ(13)C Vienna Peedee Belemnite scale was by analysis of cylinder CO₂ and CO₂ derived from dissolved carbonate standards. The performance of gas-bag analysis was compared to continuous analysis where the WS-CRDS analyser was connected directly to the soil chamber. Although there are inherent difficulties in obtaining absolute accuracy data for δ(13)C values in soil-respired CO₂, the similarity of δ(13)C values obtained for the same test soil with different analytical configurations indicated that an acceptable accuracy of the δ(13)C data were obtained by the WS-CRDS techniques presented here. Field testing of a variety of tropical soil/vegetation types, using the batch sampling technique yielded δ(13)C values for soil-respired CO₂ related to the dominance of either C₃ (tree, δ(13)C=-27.8 to-31.9 ‰) or C₄ (tropical grass, δ(13)C=-9.8 to-13.6 ‰) photosynthetic pathways in vegetation at the sampling sites. Standard errors of the Keeling plot intercept δ(13)C values of soil-respired CO₂ were typically<0.4 ‰ for analysis of soils with high CO₂ efflux (>7-9 μmol m(-2) s(-1)).