and how this resolution is affected by taxon sampling (Sanderson 1990). For homoplastic
characters, parsimony only provides a lower bound on the number of changes, and the
Figure 8.5 Chronogram calibrated against the geological timescale (Harland et al. 1990) focusing on
eurosid I and eurosid 11. Fossil-based estimates extracted from Magallón et al. (1999) are indicated
with thick bars. The chronogram is based on the analysis using ACCTRAN optimization for
calculating branch lengths, and has been generalized, usually down to family level, to facilitate a
graphical comparison between our molecular-based estimates and the fossil-based estimates
(Magallón et al. 1999). Non-generalized chronograms, indicating age estimates for all included taxa,
have been presented elsewhere (Wikström et al. 2001), and node numbers correspond to those
given there. With few exceptions (taxa marked with ) taxon names follow those used by Soltis et
al. (1999) and arrows indicate nodes that received less then 50 per cent jackknife support in their
phylogenetic analyses (Soltis et al. 1999, 2000). Taxa marked with have been generalized and
named to correspond with the usage of Magallón et al. (1999).
ANGIOSPERM DIVERGENCE 161