INTRODUCTIONThe latter half of the 20 th century saw two developments that revolutionized the study of bacterial systematics and our conceptions about microbial diversity. The first was the use of molecular genetic information, as gained through the analysis of protein and nucleic acid sequences, for the identification, typing and classification of microorganisms. The second was the application of PCR for the recovery of DNA sequences from non-cultivable organisms.By decoupling cultivation and classification, and by providing a common yardstick by which all organisms could be compared (i.e., small subunit ribosomal DNA sequences; hereafter referred to as SSU rRNA), these procedures added a new dimension to our knowledge of the microbial world. It is no longer necessary to test organisms against a panel of known phenotypic or biochemical properties, which, in themselves, are often variable in occurrence and can lead to cases of misclassification, or carry a preconceived notion about the characters that define a particular taxonomic group.Despite the considerable advances made from applying these innovations, such an approach is not without its limitations. Characterization of a single conserved sequence, while often adequate for taxonomic purposes, does not reveal very much about the evolutionary history or biology of the organism per se. Whereas the placement of an organism (i.e., sequence) into an evolutionary framework may prompt numerous predictions about its morphology, ecology, physiology and biochemistry, virtually nothing can be said about its unique biology without cultivation or additional sequence information.Other problems can arise when attempting to classify bacteria and to determine their relationships. These derive from two sources: the specific characters used to represent an organism (the choice of data), and the algorithms used to infer relationships (the methods used to construct the phylogeny or tree). With this in mind, genotypic data hold several advantages over phenotypic characters for phylogenetic reconstruction in that: (i) variation in molecular sequences is discrete and well-defined; (ii) molecular sequences comprise hundreds, if not thousands, of characters bearing potentially useful information; (iii) molecular sequences accumulate informative changes, even when the resulting phenotypes are conserved; (iv) informational macromolecules (DNA and proteins) change according to well-defined models of sequence evolution; and (v) identical phenotypes can arise by very different genetic mechanisms or pathways, such that changes other than those attributable to common ancestry can lead to the same phenotype.Questions of character convergence are not limited to phenotypic traits: portions of molecular sequences may be identical due to chance, natural selection, or lateral gene transfer.Furthermore, because the analysis of relatively small regions of DNA can furnish so much information, phylogenetic relationships are often based on data collected from a very limited portion of the genome, w...