The agglomerate size distribution for an assembly of primary particles coagulating with constant kernel K is usually derived from the solution of the relevant deterministic equations. We present an alternative stochastic approach to the problem.