Abstract In recent years, a number of authors among whom Hilgenberg, Carey, Heezen, Egyed, Wilson, Dicke, Creer and Van Hilten must be mentioned, have suggested that the earth has expanded in the course of geological time. This hypothetical expansion can be tested by means of palaeomagnetism. Various methods are available: 1. (1) the palaeomeridian method, 2. (2) Ward's (1963) method of minimum dispersion and 3. (3) the triangulation method. These methods are discussed in some detail. Van Hilten's (1963) model of earth expansion is subjected to a thorough analysis by means of spherical trigonometry. It is shown that the ancient earth radii, calculated by Van Hilten (1963), are necessarily inexact, and that some of his results are at variance with his model. All available computed ancient earth radii are evaluated. Combined with new data, compiled by the present authors, they show that only small rates of earth expansion (including a zero rate of expansion i.e., a constant earth radius) are compatible with the data. It is pointed out that the palaeomagnetic evidence for expansion can also be explained as a result of the presence of systematic errors in the palaeomagnetic inclination of sediments, the occurrence of which is also known from other sources. Large expansion rates, such as those advocated by Hilgenberg (1962) and Carey (1958) must probably be rejected. This means that reconstructions of ancient super-continents which require these large rates of expansion are almost certainly invalid.