Molecular systematic studies have focused largely on examination of DNA sequences of genes and spacers from the plastid genome, and of highly repeated genes or spacers from the nuclear genome. Additional sources of data are available, however, including mitochondrial sequences and single-copy nuclear sequences. Sequences for the nad1b-c mitochondrial intron were obtained via automated fluorescent sequencing from species representing numerous tribes from across the five traditional subfamilies of Orchidaceae. The sequences show a relatively low rate of base substitution -- less than that seen for the plastid gene rbcL -- but exhibit a large amount of length variation, ranging from approximately 1600 bp in Neuwiedia to 900 bp in Maxillaria. This combination of features means that sequence alignment is fairly straightforward, and provides an excellent opportunity to study the evolution of intron length. Insertions and deletions (indels) are distributed throughout the intron and often provide useful synapomorphies to unite taxa, although some indels show homoplasy. The basic insertion pattern appears in most cases to be a direct duplication of short adjacent sequences (1-10 bp). In some cases, much larger insertions of over 100 bp were observed. The data set, including both base changes and indels, is useful for cladistic analysis of the family; the patterns observed agree in large part with those from other molecular data sets and from morphology, and provide another tool for building a stable classification of the family.

Key words: cladistics, DNA, introns, Orchidaceae