Plants are generated by a series of cell divisions in meristematic tissues. The apical meristem is formed during the early stages of embryogenesis and consists of two outer cell layers (L, and L, or tunica) and the inner body (L3 or corpus). Axillary meristems originate from the apical meristem and have the same histogenic arrangement (Reeve 1948). New apical meristems can also be formed adventitiously from somatic cells in many types of tissues. However, plants generated by the outgrowth of such adventitious meristems are often genetically different from the original plant. The term ‘somaclonal variation’ was introduced to describe the genetic variation in plants regenerated from any form of cell culture (Larkin & Scowcroft 1981). The term has been adopted widely, but is used in various senses, especially in practical discussions (Soh 1987). In the present article ‘somaclonal variation’ indicates (i) genetic variation in plants originating from adventitious meristems (usually formed in vitro by various cell types, such as somatic cells in organs, cells in callus or single cell cultures, and germ cells), and (ii) genetic variation in cell and callus cultures themselves. Somaclonal variation constitutes a major problem in present-day micropropagation and is one of the great stumbling blocks for micropropagation via somatic embryos. It also impedes the application of biotechnological breeding techniques. The effects of biological (genotype, explant type), medium (plant growth regulators) and physical (duration of culture) factors on somaclonal variation have been noted, but so far basic knowledge is fragmentary and only a few general principles have emerged. In the study of somaclonal variation, the absence of a straightforward, rapid assay to measure its extent is a major obstacle (Orton 1983b). Reviews on somaclonal variation focus on its potential for breeding (Larkin & Scowcroft 1981; Evans 1989), on cytological aspects (Bayliss 1980; Lee & Phillips 1988; Pijnacker & Sree Ramulu 1990) or on its origins and causes (Orton 1983a; Karp & Bright 1985; Gould 1986; Sree Ramulu 1987). In the present paper, I will discuss methods to assess the extent of somaclonal variation (page 134). First, brief overviews are given on backgrounds (pages 130 and 131) and on practical aspects (page 133).