What is the relationship among Hernandiaceae, Lauraceae, and Monimiaceae, and why is this question so difficult to answer?

Molecular and morphological phylogenetic studies in the Laurales have found that Hernandiaceae, Lauraceae, and Monimiaceae sensu stricto form a monoph yletic group. Because of the paucity of phylogenetically informative substi tutions, however, relationships among families within this clade remain unc lear. In general, molecular phylogenies may conflict because of a variety o f factors, including substitution rate variation among sites and lineages, taxon sampling, outgroup choice, and base compositional biases. We analyzed a total of 2846 aligned nucleotides from a plastid intron, three spacers, and a portion of the nuclear 26S rRNA gene in a sample of Hernandiaceae, La uraceae, and Monimiaceae; we used four outgroups with differing substitutio n rates. Despite obtaining single best topologies with maximum likelihood, minimum evolution, and parsimony approaches, family relationships remained as poorly supported as they were in the previous molecular studies. Explora tion of the data indicates that varying substitution rates across lineages or sites, insufficient taxon sampling, fast-evolving outgroups, or biased b ase composition are unlikely to explain the difficult reconstruction. Exclu sion of some of the longest branched taxa (the hemiparasite Cassytha, selec ted Hernandiaceae, and two of the four outgroups) had no effect on topologi es. To resolve relationships among the three families one could now complem ent existing five-gene data sets by adding the basal genera of Lauraceae, M onimiaceae, and Hernandiaceae, which are newly sampled here, or, our prefer red strategy, by sequencing low-copy nuclear genes for the key genera to ob tain different kinds of data.