Measurements of soil-surface CO2 fluxes are important for characterizing the carbon budget of boreal forests because these fluxes can be the second largest component of the budget. Several methods for measuring soil-surface CO2 fluxes are available: (1) closed-dynamic-chamber systems, (2) closed-static-chamber systems, (3) open-chamber systems, and (4) eddy covariance systems. This paper presents a field comparison of six individual systems for measuring soil-surface CO2 fluxes with each of the four basic system types represented. A single system is used as a reference and compared to each of the other systems individually in black spruce (Picea mariana), jack pine (Pinus banksiana), or aspen (Populus tremuloides) forests. Fluxes vary from 1 to 10 μmol CO2 m-2 s-1. Adjustment factors to bring all of the systems into agreement vary from 0.93 to 1.45 with an uncertainty of about 10--15%. ¿ 1997 American Geophysical Union