The identification and classification of young star groups in other galaxies is still a controversial topic (Battinelli 1991a). Very different estimates of OB association sizes in different galaxies were explained by Hodge (1986) by a difference in linear resolution and limiting magnitude, but the existence of two kinds of resolved star groups is also essential in this issue. The bulk of OB associations are members of larger groups, star complexes (Efremov 1978), of diameter 400–1000 pc that also include individual fainter and older stars, such as Cepheids. There exists a hierarchical, embedded sequence of young star groups: there are associations, aggregates, complexes and supercomplexes (regions). Associations and complexes are the more common ones (Efremov 1988, 1989, 1993).

Given sufficient resolution and limiting magnitude, one can see both complexes as well as brighter and smaller associations mainly inside complexes, and such is the case for M31, where Efremov, Ivanov & Nikolov (1987; hereafter EIN) found, by eye, OB associations of typically 80 pc diameter and star complexes of typically 600 pc. The latter complexes are mainly the same large groups of blue stars that were identified by van den Bergh (1964) under the name of OB associations.

The issue arises whether one has simply OB associations with a large range of sizes, the smaller ones being younger as suggested by van den Bergh 1964, or if there exist two kinds of star groups of different hierarchical level, with younger ones as constituent parts of larger and older ones, as suggested by Efremov (1978, 1989) and EIN.