The most objective rank of the taxonomic hierarchy is the species. There are a number of species concepts and, thanks to differences among different groups of organisms and arguments among systematists as to the best definition, there is no definition fits all. We can, somewhat simplistically, lump definitions into two major classes that we can call the biological species concept and the morphological species concept.
For our purposes, a biological species consists of populations or groups of populations that can potentially interbreed freely within and among themselves, but which are reproductively isolated from other such populations or groups of populations. Reproductive isolation refers to genetic factors that usually prevent successful passage of genetic material (gene flow) between populations. Geographic isolation is not considered a reproductive isolating mechanism except in cases where the populations are within easy cruising range of each other but so limited by genetically set ecological factors that they do not meet. Note that some gene flow is allowed, but only when it is so minor that it doesn't materially affect the evolution of the populations involved.
The Morphological Species Concept encompasses groups of populations descended from a common ancestor that are as similar in appearance to one another as are members of a biological species and are separated from other such groups by a morphological gap as great as that between biological species. Although often used to define species of organisms that reproduce asexually, fossil species usually are morphological species since data required to assess reproductive isolation seldom are present. The portion of the definition that compares differences to those between biological species is somewhat problematical in that cryptic biological species (sibling species) show few morphological differences and thus must be ignored for purposes of definition. Cryptic species normally require genetic information for recognition.
In practice, we often use the biological species concept but assess the likelihood of reproductive isolation based on morphological criteria. For example, if two populations have the opportunity to interbreed but show no (or rare) morphologically intermediate individuals, it generally is assumed that little or no gene flow is occurring between them.
The species (especially biological species) is considered the most objective of rank of the taxonomic hierarchy. Even if we don't know whether two populations would interbreed if given the chance, there is some actual state of reproductive status between them. Other ranks, such as genera and families are more subjective, depending more on the weight that taxonomists place on the similarities and differences. One taxonomist might consider a group of species to represent one genus, while another having the same data might consider them to represent two or more genera. Generally, an unstable concensus is reached that may eventually break down because of greater biological knowledge or because of changes in philosophy between generations. In the latter case, the tendency has been for the formation of larger units (such as genera) followed by later taxonomists subdividing these into smaller units followed by later lumping these once more into larger units. Those that advocate larger units often are called lumpers and those favoring smaller units splitters.
Lumping and splitting because of philosophical differences are different from those required by advances in biological knowledge. A species, for example, may be found to be more closely related to members of a genus different than the one it has been placed in; movement into the other genus is required by the new biological knowledge. Many species have been shuffled about through the years, sometimes shuffled back and forth between genera as more and more data accrue. Relationships essentially are hypotheses and, as data accumulate, hypotheses may be strengthened or weakened.
In the early days of mammalian research, places where mammals were collected tended to be scattered about, with lots of gaps between collection stations. It often was found that two stations had kinds of mammals that were similar, but yet recognizably different. A sample of one kind of mice, for example, might be overall similar to one from another station, but with some differences in color, in length of tail, size of ears, etc. Since these were recognizably different they often were named as separate species, usually within the same genus. As geographic gaps were filled in, however, not uncommonly the mammals from the intermediate sites were intermediate in character states. Eventually, then, these animals were recognized as actually being members of the same species despite some differences. Various terms were used to designate them taxonomically, with eventually the term "subspecies" winning out.
The subspecies is the only formal rank recognized below the level of species (we do not recognize varieties or races as separate entities from subspecies under the International Code of Zoological Nomenclature). It shares with supraspecific categories the lack of objectivity, but this seems to be a somewhat greater mental problem for biologists than for, e. g., the genus category. Possible reasons are that there are more potential subspecies than genera and that the subspecies concept is (or should be) a population concept, whereas the supraspecific categories are not, at least in the Mendelian sense.
(Biological) species and subspecies are populational concepts in the sense that gene flow is at least potentially possible among members of the taxon. With the subspecies concept, there is a defined geographic component to the definition. A widely accepted definition is that of Mayr and Ashlock (1991:43): "A subspecies is an aggregate of phenotypically similar populations of a species inhabiting a geographic subdivision of the range of that species and differing taxonomically from other populations of that species." Mayr and Ashlock also recognize that there may be temporal separation rather than geographic separation, and so fossil subspecies differing from modern populations may be recognized.
Several features need to be emphasized. For one, it is the populations that are phenotypically similar, not necessarily individuals. Thus individuals within a subspecies may differ from other individuals in morphological character states, but do not form separate subspecies, instead merely indicating that the population is polymorphic. Another important point is the geographic component. Since members of different subspecies are members of the same species, then by definition of the biological species, they can interbreed given the opportunity; sympatric populations interbreeding would quickly break down any differences between them and the subspecies would disappear as separate entities. Indeed, the major practical question asked to determine whether two taxonomically different populations are members of the same species (and thus represent subspecies of that species) or not (that is, are different biological species) is, "Do they occur sympatrically anywhere during the breeding season with no evidence of free interchange of genetic material?" If they do, then they likely are separate species. The geographic component is so prominent, that subspecies sometimes are called geographic races.
Let me head off what seems to be a common misconception among beginning biologists. Namely, that a subspecies somehow is an entity entirely separate from a species. Subspecies have no existence separate from the species of which they are a part. Thus either a species has no subspecies (are monotypic: there are no subdivisions of it that differ taxonomically from other subdivisions) or it is made up of two or more subspecies (are polytypic), and there is nothing left over that is not part of one of the constituent subspecies.
Other important points to consider is that virtually every population differs to some degree from every other population of a species due to stochastic processes and, perhaps, microevolutionary events within the population. Thus variation among populations making up a single subspecies is expectable. In earlier days,it was considered (by some) to be legitimate to name as a subspecies any population that could be shown to be statistically different from other populations of the species. Eventually it became evident to most that, given a large enough sample size, any two reasonably sized populations could be shown to differ. Since any utility at all would be lost by naming each local population, the idea of "taxonomic" difference comes into play—but, as noted, how much difference is required for it to be a taxonomic difference is subjective, and there has been limited success in getting systematists to agree on the magnitude of that difference.
The populational aspect also causes some problems with individuals who have difficulty thinking populationally. For example, by the conception of most systematists, there is no requirement that every individual of a subspecies be phenotypically identifiable as belonging to that subspecies as opposed to a different subspecies. The reason, of course, is that it is the populations that must differ taxonomically. To make the point, we might go to the absurd and postulate one population of a species consisting of 80% albinos and a different population consisting of 5% albinos. The two populations definitely are different (and I won't argue whether this is a taxonomic difference or not), but having in hand an albino individual does not allow that individual to be identified to subspecies unless the geographic location from which it was taken is known.
Another problem is that there may be independent occurrences of phenotypically similar populations in different geographic areas (when considered the same subspecies, these are called polytopic subspecies). This may be because of parallel evolution due to environmental selective features—the common example of dark populations of various animals on basaltic lava beds is typical. I have argued elsewhere that they should not be considered members of the same subspecies because the defining taxonomic characters have not been inherited from a common ancestor.
Yet another problem is based on selection of characters. Often phenotypic characters vary independently geographically. For example, size might vary from north to south, while tail length might vary from east to west. Some workers have suggested that subspecies should be named only when characters are correlated geographically.
Many characters vary clinally (a cline is a character gradient through space [or time]). Some systematists would recognize subspecies in a geographically continuous, clinal distribution only if there is a narrow geographic region of rapid change in character states (a step cline). Otherwise, any boundary drawn between subspecies along the cline would be entirely arbitrary.
Since, by definition, individuals of different subspecies will interbreed when they have the opportunity, we normally expect to find a zone of intergradation between subspecies that do come into contact with one another geographically. Any attempt to name individuals within the zone as belonging to a particular subspecies is of little value (sometimes to emphasize the intermediate characteristics, the "x" symbol may be used; e.g., Peromyscus maniculatus blandus x rufinus. However, the "x" symbol more commonly is used to indicate a hybrid between two species).
In practice, the subspecies may be a useful tool, but must be used with caution and pretty much requires familiarity with the particular species and what philosophy has been used to name the subspecies within that species.
Last Update: 18 Jan 2008
Centennial Museum and Department of Biological Sciences, The University of Texas at El Paso