The correlation of certain plant distributions with the presence or absence of calcium carbonate in the soil is well known. Hope-Simpson (1938) has defined as calcicoles 'species affecting the more important types of calcareous soil and rare on, or absent from, acid soils, and calcifuges as the reverse'. Other pairs of terms are also in use to describe the same observations. There is some danger that the semantic similarity of the terms may be taken to imply a reciprocally related explanation for each of the calcicole and calcifuge distribution types. With this reservation, the terms are useful as field descriptions, much as fever in medicine describes a class of symptoms with many underlying causes. The experiments to be described have been concerned only with 'the calcicole problem'. Much work has been done on this subject and numerous summaries have been produced, e.g. by Black (1957), Grime (1960), Hartwell & Pember (1918), Hewitt (1952), Jensen (1952), Lundergardh (1931), Rorison (1956), Russell (1961), Salisbury (1920), Schmehl et al. (1950), Steele (1955), Webb & Hart (1945), Wilde (1954) and Zlatnik (1928). To add another would serve no useful purpose. There are, however, four complicating factors, the importance of which has not been generally recognized. These affect the design of experiments, and must therefore be mentioned first of all. First, as was pointed out by Salisbury (1920), most calcareous soils are well drained. He suggested that the calcicoles Fagus sylvatica, Juniperus communis, Clematis vitalba, Buxus sempervirens, Helleborus foetidus and the moss Pleurochaete squarrosa are really plants of well-drained soils. The calcicole status of beech and juniper is questionable (for example in the New Forest and Rothiemurchus Forest respectively) and the restriction of box to calcareous slopes may be related to the instability of these areas (Pigott & Walters 1953), but the possibility of such an explanation must be borne in mind, since most of the plants used to investigate the calcicole problem have been chalk downland plants. Secondly, a number of plant taxa are calcicole only at the limits of their range. They may be called 'marginal calcicoles'. Praeger (1950) considered Stellaria palustris, Arenaria verna, Sedum acre, Origanum vulgare and Primula veris as calcicole in Ireland but not in Britain. A survey of the soil preferences of the plants used in earlier investigations of the calcicole problem is given in Table 1. The information has been condensed from accounts in various Floras. It seems therefore that about two-thirds of these plants are calcicole in Britain in special circumstances; at the limit of their range, in well-drained soils, or as 'ruderals and weeds' in open disturbed ground. The experiments of Webb & Hart, and Steele