The changes in the eluted reactant's response curves from a pulse reactor operated in Knudsen flow [e.g. a temporal-analysis-of-products (TAP) reactor] as a function of the pulse size was interpreted as due to varying fractional coverages of the catalyst surface. These were used with the reactor modeling of the response curves to determine the active site concentration on a catalyst powder. The technique uses a transient pulse to bring about significant site coverage, and is an improvement on the present use of gas-surface concentration equilibrium because it is able to selectively determine just the active site concentration and not the entire exposed surface atom concentration. (c) 2007 Elsevier Ltd. All rights reserved.