Artificial and natural classifications
Classifications that use single or at most only a few characteristics to group plants usually are artificial classifications—that is, all the plants in a single group share the same characteristics, but they are not closely related to one another genetically. Popular floras (books to identify plants of a certain area) sometimes group plants using color of their flowers, or their growth form (trees, shrubs, herbs, and so on). Although such books are useful in finding the names of taxa, they provide few clues about relationships among the taxa and hence are not predictive, which means that you can deduce nothing more about the plant other than that it exhibits the characteristics used to classify it. Natural classifications group together plants with many of the same characteristics and are highly predictive. That is, by enumerating the characteristics of a plant, one can predict the natural group to which it belongs. Taxonomic floras, for example, identify species, genera, and families by listing as many characteristics as possible concerning anatomy, morphology, cytology, ecology, biochemistry, genetics, and distribution.
Phylogenetic (phyletic) classifications
Phyletic classifications are natural classifications that try to identify the evolutionary history of natural groups. When botanists accepted Darwin's theory of evolution near the end of the last century, the reasons why some groups of plants looked alike became clear: They were related to one another by a common ancestry. The mission of taxonomy since Darwin has become a quest for evolutionary relationships, not just at the lower levels of the hierarchy, but at the upper levels as well.
The evolutionary history of a taxon is called its phylogeny. To establish phylogenies, decisions must be made concerning which characteristics are “primitive” and which “advanced”—that is, which taxon is the ancestor of the others. Early phylogenetic classifications were based primarily upon plant morphology and anatomy with great emphasis upon reproductive morphology, which is more stable and less influenced by the environment than is vegetative morphology. Today, taxonomists additionally use the techniques of biochemistry and molecular biology to add details of internal organization and mechanisms to the classifications. But phylogenies, no matter how carefully constructed, are dependent upon someone's interpretation of data, and herein lies the problem: Systematists frequently differ in their interpretations of relationships. A phylogenetic classification is a hypothesis, a scientific explanation of the data and, like any hypothesis, is subject to further testing.
Certain assumptions are necessary in phylogenetic classifications. A taxon should bemonophyletic (all of the members of the taxon should be descendants of a single common ancestor). The characters or features used to identify the taxa must behomologous, which means that they must have a common origin, but not necessarily a common function. For example, all the parts of a flower—petals, sepals, stamens, and carpels—originate in the same way as leaves from primordia in meristems. Although they now have different functions in the flower (they're not photosynthetic), some sepals and petals structurally resemble leaves. Leaves and the parts of the flower are homologous structures.
Some features that look alike do not have a common origin and are said to beanalogous. An example of analogous structures is the prickles on two groups of succulent desert plants, the cacti and the euphorbs. Cacti have spines that are modified leaves; euphorbs have thorns that are modified branches. Spines and thorns look alike and are functionally similar in that both keep animals from eating the plants. Spines and thorns are analogous. This example of analogy is also an example of convergent evolution. The cactus family and the euphorb family both developed the same morphology in response to a desert environment—the cacti in North and South America, the euphorbs in Africa and Asia. The families are not related and have no recent common ancestor.
Numerical taxonomy (phenetics). Systematists have tried many ways to make phyletic classifications more subjective. When computers became readily accessible in the 1960s, numerical taxonomy or phenetics became a popular approach. In practice, measurements were made of a large number of characters of a taxon, at least 60 per plant and often 100 or more. No special importance was attributed to any one of the characters. After the measurements were complete on hundreds of individuals, the data were analyzed statistically with computer programs and cluster analysis or other methods to show purported natural groupings of plants with overall similarities. Systematists' interpretations were thought to be minimized in this fashion.