Illustrates that the Poisson-lognormal model provides good fits to a diverse set of distributions commonly studied in bibliometrics and scientometrics. Topics discussed include applications to the empirical data sets related to the laws of Lotka, Bradford, and Zipf; causal processes that could generate lognormal distributions; and implications for future studies. (Contains 61 references.) (LRW)