(1) This paper examines removal and other experimental studies on the role of territorial behaviour in the limitation of bird densities. Experimental design is discussed, as are the types of conclusions that can be drawn.

(2) Experiments have been conducted on more than 40 species from a wide range of taxonomic groups. Most provided evidence for density limitation, probably mediated by territorial behaviour, and for the existence of a non-territorial sector able to take territories when territory owners were removed or when additional habitat was made available. Experiments in the breeding season indicated that some surplus birds, although younger on average then territorial birds, were sexually mature and able to breed when given the chance. In some species, replacement birds bred less well than other territorial birds, but in other species no difference was apparent between the two groups.

(3) Other experiments indicated that, while density was limited by territorial behaviour in good habitats, this was not the case in poor habitats, and that some individuals would move from poor to good habitat when territory owners were removed from good habitat.

(4) In some species, non-territorial birds of breeding age were present in the population despite the existence of vacant territorial sites, while in others replacements were observed on good territories but not on poor ones. The implications were that site quality influenced whether settlement, defence and breeding occurred, and that some individuals had more stringent site requirements than others. In some seabirds' sites in the centre of a colony were more attractive than sites on the edges.

(5) Many land-bird species that have been studied show two peaks of territorial activity each year, in autumn and in spring, and a limitation on breeding density can occur at either or both seasons, depending on conditions.

(6) The fact that densities of many bird species fluctuate greatly from year to year is not inconsistent with the limitation of density by territorial behaviour. Several mechanisms are apparent through which density might be limited by territorialism at different levels in different years, so that surplus non-territorial birds are present in years of low territorial density, as well as in years of high territorial density.

(7) Experiments have shown that density is limited by the presence of territorial birds, in some species at different levels in different areas or years. The next step is to find whether density is regulated by the presence of territorial birds over a period of years in a density-dependent manner. For this, observational data are required to find whether the proportion of birds that is excluded by territorial behaviour each year varies with the total number available for settlement. Such studies can be made only on species in which non-territorial birds can be counted accurately, as well as territorial ones.