Molecular sequence data have become increasingly used for examining the evolutionary history of plants at scales ranging from relationships among the major lineages of land plants to relationships within individual genera. Phylogenetic systematics has progressed in part through the development of new molecular markers suited to particular classes of phylogenetic problems (e.g., Mathews et al., 2000; Olsen and Schaal, 1999; Palmer et al., 1988; Taberlet et al., 1991; Whitlock and Baum, 1999). Nonetheless, there are few genes that may be easily amplified from a wide range of taxa that have levels of sequence variation appropriate for studies of closely related species or studies at the population level (Schaal et al., 1998). Consequently, species and population-level phylogenies, especially involving recent radiations, most commonly utilize techniques such as inter-simple sequence repeats (Wolfe and Randle, 2001), amplified fragment length polymorphisms (Albertson et al., 1999; Bakkeren et al., 2000), and microsatellites (Billotte et al., 2001; Dayanandan et al., 1997). Unfortunately, these techniques do not provide gene genealogies, which may be useful for clarifying species limits and patterns of gene flow (Avise and Ball, 1990; Baum and Shaw, 1995), but rather provide a summation across numerous, possibly discordant, gene genealogies. The most commonly used marker for sequence analysis at low phylogenetic levels is the internal transcribed spacer (ITS) region of 18S–5.8S–28S rDNA (Baldwin et al., 1995). While ITS has been useful in elucidating some island lineages (reviewed in Baldwin et al., 1998), it is often invariant in clades that have undergone rapid or recent radiations (Ganders et al., 2000; Vargas et al., 1998). This is particularly true in island radiations, which are consequently often assessed with morphological data (Buss et al., 2001; Wagner and Funk, 1995). Additionally, ITS is present in multiple copies per genome and subject to varying degrees of concerted evolution, which can occasionally result in a distorted picture of evolutionary history (Buckler et al., 1997, Wendel et al., 1995). Palumbi and Baker (1994) described a method to find more variable regions of the genome, which they called exon-primer, intron-crossing (EPIC). The method entails using data from two or more model species to find conserved exons that flank variable introns. Primers can then be designed within these flanking exons that will amplify the intron in a broader array of taxa. Strand et al. (1997) used such a technique to design primers for amplifying introns from a diversity of plant taxa. However, they focused on markers with multiple small introns. For example, glyceraldehyde-3-phosphate dehydrogenase (G3PDH), which has been used successfully at the population level (Olsen and Schaal, 1999), has introns that are typically only 100 bp in length. Longer introns would provide more phylogenetic information per unit sequence because less exon sequence would need to be generated. Therefore, we set out to use EPIC to find longer nuclear introns that could be amplified from a diversity of plants and could provide phylogenetic information at low taxonomic levels. This paper describes one target of this effort, the nitrate reductase (NIA) gene, which is a promising candidate for phylogenetic reconstruction. The gene (sometimes abbreviated NR) has been isolated from fungi, algae, and land plants (Zhou andKleinhofs, 1996). NIA from most land plants acts as a homodimer and, with NADH as a cofactor, catalyzes the first reaction in the uptake of nitrogen from the soil, the reduction of Molecular Phylogenetics and Evolution 23 (2002) 525–528 MOLECULAR PHYLOGENETICS AND EVOLUTION