Introduction
Quantitative analyses of guild structures of living and fossil mammals have a relatively long history (e.g. Valverde, 1964; Van Valkenburgh, 1988; Legendre, 1989; Gunnell et al., 1995), although carnivores have often been excluded from older studies. However, some studies have been published dealing with general carnivore ecomorphology (e.g. Van Valkenburgh, 1992, 1999; Werdelin, 1996; Van Valkenburgh et al., 2004; Wesley-Hunt, 2005), or structures of single guilds (e.g. Dayan et al., 1989; Viranta and Andrews, 1995; Dayan and Simberloff, 1996; Jones, 2003; Hertler and Volmer, 2008). Few of these studies, however, have combined more than two parameters (e.g. body mass and diet or body mass and locomotion). In addition to body mass, diet and locomotor patterns can satisfactorily be estimated for fossil taxa (see Morlo, 1999, for an example using these three parameters in an analysis of creodont guilds). A similar methodological approach has been applied to compare several carnivore guilds (Morlo, 1999; Nagel and Morlo, 2000, 2003; Morlo and Gunnell, 2003, 2005a,b, 2006; Nagel et al., 2005; Stefen et al., 2005; Morlo and Nagel, 2007). In this chapter, we augment these studies with the addition of a set of guild analyses from the Paleocene to the Recent. Having guild structure established on the three parameters, two guilds can be tested against each other by principal component analysis (PCA) to clarify which parameters are mainly responsible for the differences.