Historically, chemical taxonomy of plants developed out of the chemistry of natural products, and was limited essentially to observations of distributional correlations between groups of plants and certain of their chemical products. The possibilities of the application of chemical knowledge to the problems of systematics have intrigued both chemists and biologists for many years. The classic work of Baker and Smith (1) in 1920, on the oils of Eucalyptus, is generally cited as an example of such an application. The work of these investigators surpassed that of much later periods in that, throughout their studies, the specific oil components of Eucalyptus were always identified. In later work, the chemistry was often treated only superficially, but inferences of great phylogenetic significance were extracted from the meager data (2).Most chemical studies devoted primarily to systematic problems in plants are concerned with groups of secondary constituents such as alkaloids, terpenes, and various water-soluble pigments. Within a particular taxonomic group the presence of certain unusual compounds may help to delimit the group (for example, the alkaloids of Senecio or Veratrum; the isothiocyanates, or mustard oils, of Cruciferae). Sometimes, distinctive chemical features provide sound bases for speculation about SC~E=IENCE corollary, the biosynthetic and enzymatic bases of the two systems, should be of great interest. There are numerous other examples of compounds whose distributions make it difficult to determine whether the compounds are indicative of a phylogenetic connection or of independently evolved pathways. Many alkaloids, the isothiocyanates, certain nonprotein amino acids, some steroids, some sesquiterpenes, the carbon glycosides of flavones, and even some fatty acids (such as petroselenic acid of the Umbelliferae, Araliaceae, and Simarubaceae) are substances whose taxonomic distributions may reflect one to several independent evolutionary appearances.As a general guide to the systematic evaluation of comparative chemical data, one may use the "percentage of frequency rule" (3). In principle, this rule states that the taxonomic distribution of a member of a group of biosynthetically related substances is inversely related to the chemical complexity of the member. Therefore, the more complex members of a chemical series occur in taxa which are evolutionarily more advanced. As long as the principle is not applied as a rigorous criterion for phylogenetic positioning it is useful. However, numerous examples, involving loss mutations and other evolutionary mechanisms, may be called to mind which limit the reliability of the percentage of frequency rule. Figure 2 illustrates the broad application of the rule to a few selected groups of phenolic compounds, all of which are considered to have a close biosynthetic relationship. An apparent contradiction to the percentage of frequency rule is encountered among quinones. Excluding coenzymes (for example, ubiquinone), the monocyclic benzoquinones have a more rather b...