Empirical hyperbolic distributions (Bradford-Zipf -Mandelbrot) for bibliometric description and prediction (Reprinted from Journal of Documentation, vol 25, pg 319-343, 1969)

Purpose - Aims to build on the work of Buckland and Hindle regarding statistical distribution as applied to the field of bibliometrics, particularly the use of empirical laws. Design/methodology/approach - Gives examples of hyperbolic distributions that have a bearing on the bibliometric application, and discusses the characteristics of hyperbolic distributions and the Bradford distribution. Findings - Hyperbolic distributions are the inevitable result of combinatorial necessity and a tendency to short-term rational behaviour. Originality/value - Supports Bradford's conclusion from his law, i.e. that to know about one's speciality, one must go outside it.