Flux chambers are useful and convenient tools for measuring gas emissions at soil and water surfaces in agricultural, ecological, environmental, and engineering studies. In this experiment, a closed chamber and a dynamic chamber were tested to study their general behavior and to identify factors affecting flux measurement. The experiment was designed and conducted on the basis of a previous study where the behavior of these flux chambers was simulated using mathematical models. Emission of a volatile solvent (CH2Cl2) from a constant source was measured at the surface of a soil layer by both closed and dynamic chambers. Measurements from the closed chamber tests show that the average flux calculated over a placement time (ti-t0) by a linear model is smaller than the initial flux at t0=0 but greater than the temporal flux at ti-ti-1. The results from the dynamic chamber tests indicate that the steady-state flux may underestimate the actual flux when the chamber is operating at low airflow rates but overestimate the actual flux at high airflow rates. The underestimate at a low airflow rate is probably due to a depression on the diffusive flux at the enclosed soil surface, while the overestimate is due to a pressure deficit present within the chamber headspace that induces an advective flux from the covered soil matrix. The vacuum system operating the dynamic chamber in this experiment was found to be a predominant source of the pressure deficit. The air permeability of soil matrix and its surface condition are demonstrated to be important factors that determine how significant the effect of the pressure deficit is. In general, the experimental results agree with the simulation results reported previously. When using closed chambers, it is recommended that appropriate nonlinear models be used to calculate flux. When using dynamic chambers, which are more desirable, relatively high airflow rates should be employed and the pressure deficit within the chamber headspace should be measured and minimized. ¿ 1998 American Geophysical Union