A two-gas-fraction model previously tested with gas-phase residence time distribution tests was used to interpret experimental transient curves during K(L)a determinations made with a conventional dynamic method. Large bubbles are modelled by an axial dispersion model, whereas small bubbles are considered as perfectly mixed. The use of this two-gas-fraction model leads to a significant improvement in the fitting of the experimental curves. Furthermore, the analysis carried out suggested that what is actually observed during an experiment carried out with the dynamic method is the mass transfer of large bubbles only-even in the presence of a significant fraction of small bubbles. (c) 2007 Elsevier Ltd. All rights reserved.