Current theories of the primary physical processes in the photosynthetic unit are based on a model in which excitations diffuse throughout a lattice and become trapped at a specialized centre. Despite the apparently well-defined nature of the problem, master equation and random walk calculations have given different answers for the average trapping time, and some doubts have been expressed about the equivalence of the methods. The equivalence is made explicit here, and the two-dimensional random walk solution by ten Bosch & Ruijgrok (1963) is found to be incorrect. The asymptotic dependence of trapping time on N, the number of sites per trap, found to be proportional to N ln N by Pearlstein (1966) and by Robinson (1967) in square networks, has been verified and extended to the triangular case. Some general kinetic considerations are presented and applications to the theory of photosynthesis are briefly discussed. © 1968.