The Responsibilities of Taxonomists, Part 2:
Taxonomic Categories and the Quest for a Natural Classification Scheme

Dorothy E. Tuthill and Gregory K. Brown
Bromeliad Research Lab, Department of Botany,
University of Wyoming, Laramie, WY 82071


Classification of entities seems to come naturally to human beings, as we seek generalizations that make communication easier between us. For professional and non-professional botanists alike, it is much easier to say “bromeliad,” by which we mean members of the family Bromeliaceae, than it is to say something like “I enjoy working on/looking at/growing those herbaceous plants that form perennial rosettes with short axes, that may be terrestrial, saxicolous or epiphytic, having roots that may be modified into holdfasts, and with terminal inflorescences that may be scapose or simple, compound or simple,” and so on until we have included all of the characters that comprise “bromeliad.” Scientific classification, called taxonomy, endeavors to provide names and descriptions at all levels of classification, from the very broad, such as the kingdom Plantae, to the smallest unit that we can recognize, such as Tillandsia fasciculata var. densispica forma alba, a name that conveys a wealth of information to someone familiar with it.

            The fundamental taxonomic unit is the species, whose name is the Latin binomial that we are familiar with. For example, the species name Tillandsia fasiculata Swartz includes the genus name Tillandsia, indicating a bromeliad with leaves having smooth margins and scales (usually), plumose seeds, separate petals that lack appendages, etc., and fasiculata, the specific epithet, which means clustered (perhaps the flower spikes or maybe the plants themselves), but implies also a number of other characters that separate T. fasiculata from other Tillandsias. (See TABLE 1).

TABLE 1. Taxonomic placement of Tillandsia fasciculata in the grand scheme of living organisms.

Common name
Domain Eukaryota organisms with nuclei
Kingdom Plantae plants
Division Magnoliophyta flowering plants
Class Liliopsida monocots
Order Poales grass order
Family Bromeliaceae pineapple family
Subfamily Tillandsioideae Tillandsia subfamily
Genus Tillandsia tillandsia
Species Tillandsia fasciculata Swartz cardinal airplant
Variety T. fasciculata var. densispica
Form T. fasciculata var. densispica

The third part of a species name is the name of the person who originally described the species. This is necessary because sometimes two people, unaware of each other’s work, use the same name for different plants. Although only one of the uses can be retained according to the rules of botanical nomenclature it is important that subsequent botanists know which concept of the name is in use. If a later taxonomist decides that a species has been assigned to the wrong genus, his (or her) name gets added also. For example, when Jason Grant moved Vriesea penduliscapa Rauh to Tillandsia (Grant 1993), the species’ new name became Tillandsia penduliscapa (Rauh) J.R. Grant.

Taxonomists don’t all agree as to what constitutes a species. Zoologists often use the biological species concept (BSC) of Ernst Mayr (1963), which defines a species as a group of interbreeding or potentially interbreeding populations, but this concept is often not useful to botanists. Plants, and bromeliads in particular, tend to breed across what appear to be reasonable species boundaries, and even across genera. In a practical sense, plant taxonomists look for groups of individuals (or populations) that are alike, and that can be distinguished from other, similar groups. Traditionally, this has meant “look alike” since plant taxonomy has relied heavily on morphology. If we see variation within species we can call particular groups subspecies, varieties or forma. Technically, the category subspecies is greater than variety, but usage tends to be a matter of preference, with authors using one or the other. Smith and Downs (1974, 1977, 1979) recognized varieties but did not consider any subspecies in the Bromeliaceae. Forma refer to particular characters that are known to be variable, such as flower color.

Genera are made up of species that share one or more characters. To continue with the Tillandsia example, species in Tillandsia are very much like species in the genus Vriesea, with the exception that all Tillandsia have petals that lack appendages, while all Vriesea have appendages on their petals (FIGURE 1). In defining genera, taxonomists try to pick characters that they think show evolutionary signal; that is, they want genera that are natural products of evolution, and not artificial collections of unrelated species.

FIGURE 1. Scanning electron micrograph of petal appendages on Vriesea carinata. In this case the appendages (A) are also called scales because of their shape. S = stamen filament. P = petal.

In the bromeliads, the morphological concepts of species and genera have worked pretty well in some groups, but poorly in other groups. The characters that are visible to humans are also the characters that provide the interface between the plant and the entire ecosystem of which it is a part. Therefore, these characters tend to be under strong selective pressure to ensure survival of the species. Consider leaves, for example. Bromeliads that live in shaded environments such as forest floors tend to have broader leaves because a large leaf surface helps the plant intercept more light. If the environment in which the plant lives becomes drier, then there may be less overhead leaf cover and more sunshine, and over generations leaves may become more narrow as light is no longer limited. Narrow leaves are more resistant to water loss, an advantage in a drier habitat. Likewise, features of the inflorescence are determined in large part by the need to attract pollinators, so that unrelated species relying on similar pollinators share characters that are attractive to those pollinators.

Evolutionary change has been rapid in the Bromeliaceae. We know that the family as a whole is quite young, younger than the South American continent, yet the number of species in the family exceeds that of all other New World families except the Cactaceae. Within this short time period bromeliads have diverged greatly, both in number of species and morphological characters. However, it is often the case that only some characters evolve while others remain the same. Flowers may change radically to attract a new pollinator, while the rest of the plant remains unchanged, so that plants look alike except when in flower. In another lineage the flowers may remain similar though the vegetative parts become drastically different, so that the plants look very different until they bloom. And, as mentioned before, unrelated plants can evolve to look alike as a response to similar selective pressures. All of these changes in myriad directions can make the sorting out of species and genera very difficult.

Fortunately, plant taxonomists have acquired new tools to help with the sorting, and to help determine how species and genera are related. The study of evolutionary relationships is called systematics, and a favored tool of systematists is phylogenetics. Phylogenetic analyses involve searching for patterns of character changes within lineages, which in turn involves determining which characters are primitive and which are derived (advanced). The result is a “tree” diagram, where related species are found on the same branch and most closely related species are on adjacent twigs (FIGURE 2). These analyses can be performed on morphological data, but are more commonly done with molecular data, i.e. DNA sequences. DNA sequences provide many more characters than morphology and are independent of morphology, thus allowing systematists to test the traditional, morphology-based taxonomic ideas with new molecular data.

FIGURE 2. A phylogenetic tree depicting the relationships among some members of the Tillandsioideae. Brocchinia, Bromelia and Navia are “outgroup” taxa, included to determine which character states are primitive (possessed by all bromeliads) and which are derived within the subfamily. Catopsis and Glomeropitcairnia appear to be more closely related to each other than to other tillandsioids. Vriesea vittata, V. viridiflora and V. gladioliflora are together on a branch near Mezobromelia, but V. espinosae and V. malzinei are on a different branch with representative species of Tillandsia. This tree indicates that Tillandsia and Vriesea, as currently recognized, are not natural groups. Tree based on ndhF sequences; redrawn and simplified from Terry et al. (1997).

The hope, of course, is to find all members of a genus on the same branch, all members of the same subgenus on the same subbranch and so on. When that is not the case we need to reevaluate our traditional concepts of the group. For example, Terry et al. (1997), using sequences from the chloroplast gene ndhF, found some Vriesea species on the Tillandsia branch of the tree (Figure 2). This supported earlier suggestions (Utley 1983, Gardner 1986, Brown & Terry 1992, Grant 1993) that presence or absence of petal appendages, traditionally used to separate those two genera, is not always a “good” taxonomic character, because it places species into groups that do not reflect their true relationships.

Because bromeliads have diversified rapidly and recently, most gene sequences that are used for phylogenic analyses in other plants don’t show enough differences to fully resolve the relationships within groups of closely related bromeliads. Therefore, bromeliad systematists are still searching for genes and gene regions that show enough variation to help answer these questions. And to further complicate matters, we know from previous experience that not all genes tell the same evolutionary story. Because of these problems, and the conservative nature of most systematists, Terry et al. (1997) did not propose to change the names of any Vriesea species to Tillandsia based on their work with the ndhF gene. However, if more analyses based on more genes support their conclusions, name changes may be in order. Although changes may be frustrating to non-systematists, when based on sound and thorough information they improve the value of the taxonomic system. Ideally, taxonomic groups should be natural groups; natural classification systems provide maximum information because evolutionary history is included. And maximum information is what taxonomy is all about, so that we don’t need to describe our favorite plants in detail every time we want to express our admiration of a single beautiful bromeliad!


This project has received partial funding from a National Science Foundation grant (DEB-0129446) to G.K.B.

Literature Cited

Brown, G.K. and R.G. Terry. 1992. Petal appendages in Bromeliaceae. American Journal of Botany 79: 1051—1071.

Gardner, C.S. Preliminary classification of Tillandsia based on floral characters. Selbyana 9: 130—146.

Grant, J.S. 1993. True Tillandsias misplaced in Vriesea (Bromeliaceae: Tillandsioideae). Phytologia 75: 170—175.

Mayr, E. 1963. Animal Species and Evolution. Cambridge, MA: Belknap Press.

Smith, L.B. and R.J. Downs. 1974, 1977, 1979. Bromeliaceae. Flora Neotropica Monograph 14, parts 1—3.

Terry, R.G., G.K. Brown and R.G. Olmstead. 1997. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) using ndhF sequences. Systematic Botany 22: 333—345.

Utley, J. 1983. A revision of the middle American thecophylloid Vrieseas (Bromeliaceae). Tulane Studies in Zoology and Botany 24:1—81. [an error occurred while processing this directive]