Noncoding chloroplast DNA sequence comparison has become a popular tool for systematics studies at low taxonomic levels. But earlier expectations of random, unconstrained, and independent nucleotide evolution in such regions have not been validated by research results. Mutation in noncoding cpDNA may be largely dependent on sequence structure and pattern, resulting in nonindependent and predictable sequence evolution. Mechanisms of molecular evolution described in these regions include slipped-strand mispairing, stem-loop secondary structures, minute and moderate-sized inversions associated with secondary structures, intramolecular and extra-regional recombination, and nucleotide substitutions. Detection of mutational mechanisms contributing to inferred mutations can enhance the alignment of insertions and deletions and the assessment of nucleotide and indel homology. The manner of noncoding cpDNA evolution described here has several important repercussions for phylogenetic analysis, effecting all levels of methodology by invalidating underlying assumptions. Unique evolutionary characteristics of noncoding cpDNA may eventually prove no more difficult to assimilate than related phenomena occuring in genic DNA; perhaps a reasonable course for now is to identify probable mutational processes evident in a data set, and carefully modify existing phylogenetic analysis procedures to more accurately reflect the underlying mutational model.

Key words: alignment, methodology, models of sequence evolution, molecular evolution, noncoding evolution, phylogeny estimation