Cladistic analysis has been used for more than 20 years to reconstruct the phylogenetic relationships of fossil catarrhine species and genera (e.g. Delson & Andrews, 1975; Eldredge & Tattersall, 1975; Delson, 1977; Delson et al., 1977; Tattersall & Eldredge, 1977; Andrews, 1978, 1992; Corruccini & McHenry, 1980; Harrison, 1982; Skelton & McHenry, 1986; Wood & Chamberlain, 1986, 1987; Andrews & Martin, 1987; Chamberlain & Wood, 1987; Strasser & Delson, 1987; Stringer, 1987; Wood, 1988, 1991, 1992; Skelton & McHenry, 1992; Lieberman et al., 1996; Begun et al., 1997; Cameron, 1997; Rae, 1997; Strait et al., 1997). However, it is now apparent that, in contrast to the situation with higher-level primate taxa (Harrison, 1993), few of the relationships supported by these analyses can be considered to be reliable. This is demonstrated by the small increases in length required to alter the topologies of the most parsimonious cladograms. For example, the addition of only one step converts the Homo monophyly seen in Wood's (1991) most parsimonious cladogram into Homo paraphyly, as well as altering the relationships of A. africanus (Wood, 1992). Likewise, the addition of two steps to the cladogram preferred by Strait et al. (1997) results in Homo paraphyly (Wood & Collard, 1999). These examples are taken from the hominin palaeontological literature, but they could easily have been taken from studies of Miocene hominoids, Eurasian pliopithecids, or fossil Old World monkeys (e.g. Harrison, 1993; Rae, 1997). The unreliability of the most parsimonious cladograms is also illustrated by the results of Corruccini's (1994) bootstrap re-analysis of hominin data from Wood & Chamberlain (1986), Skelton et al. (1986), Chamberlain & Wood (1987) and Skelton & McHenry (1992).