Monday, 7 January 2008

Biological Classification and Systematics

The Linnaean classification

Scientific classification or biological classification is how species both living and extinct are grouped and categorized. Man’s desire to classify the natural world seems to be very deep rooted and the fact that many traditional societies have highly sophisticated taxonomies suggests the practice goes back to prehistoric times. However the earliest system of which we have knowledge was that of Aristotle, who divided living organisms into two groups – animals and plants. Animals were further divided into three categories - those living on land, those living in the water and those living in the air, and were in addition categorised by whether or not they had blood (those “without blood” would now be classed as invertebrates). Plants were categorised by differences in their stems.

Aristotle’s system remained in use for hundreds of years but by the 16th Century, man’s knowledge of the natural world had reached a point where it was becoming inadequate. Many attempts were made to devise a better system, but the science of biological classification remained in a confused state until the time of Linnaeus, who published the first edition of his Systema Naturae in 1735. In this work, he re-introduced Gaspard Bauhin’s binomial nomenclature and grouped species according to shared physical characteristics for ease of identification. The scheme of ranks, as used today, differs very little from that originally proposed by Linnaeus. A taxon (plural taxa), or taxonomic unit, is a grouping of organisms. A taxon will usually have a rank and can be placed at a particular level in the hierarchy.

The ranks in general use, in hierarchical order, are as follows:

Phylum (animals or plants) or Division (plants only)

The prefix super- indicates a rank above; the prefix sub- indicates a rank below. The prefix infra- indicates a rank below sub-. For instance:


Even higher resolution is sometimes required and divisions below infra- are sometimes encountered, e.g. parvorder. Domains are a relatively new grouping. The three-domain system (Archaea, Bacteria and Eukaryota) was first proposed in 1990 (Woese), but not generally accepted until later. Many biologists to this day still use the older five-kingdom system (Whittaker). One main characteristic of the three-domain system is the separation of Archaea and Bacteria, previously grouped into the single prokaryote kingdom Bacteria (sometimes Monera). As a compromise, some authorities add Archaea as a sixth kingdom.

It should be noted that taxonomic rank is relative, and restricted to the particular scheme used. The idea is to group living organisms by degrees of relatedness, but it should be bourn in mind that rankings above species level are a bookkeeping idea and not a fundamental truth. Groupings such as Reptilia are a convenience but are not proper taxonomic terms. One can become too obsessed with whether a thing belongs in one artificial category or another – e.g. is Pluto a planet or (closer to home) does habilis belong in Homo or Australopithecus; does it really matter if we lump the robust australopithecines into Australopithecus or split them out into Paranthropus?


Systematics is the study of the evolutionary relationships between organisms and grouping of organisms. There are three principle schools of systematics – evolutionary taxonomy (Linnaean or “traditional” taxonomy), phenetics and cladistics. Although there are considerable differences between the three in terms of methodologies used, all seek to determine taxonomic relationships or phylogenies between different species or between different higher order groupings and should, in principle, all come to the same conclusions for the species or groups under consideration.

Some Terminology and concepts

One of the most important concepts in systematics is that of monophyly. A monophyletic group is a group of species comprising an ancestral species and all of its descendants, and so forming one (and only one) evolutionary group. Such a group is said to be a natural group. A paraphyletic group also contains a common ancestor, but excludes some of the descendants that have undergone significant changes. For instance, the traditional class Reptilia excludes birds even though they evolved from an ancestral reptile. A polyphyletic group is one in which the defining trait evolved separately in different places on the phylogenetic tree and hence does not contain all the common ancestors, e.g. warm-blooded vertebrates (birds and mammals, whose common ancestor was cold-blooded). Such groups are usually defined as a result of incomplete knowledge. Organisms forming a natural group are said to form a clade, e.g. the amniotes. If however the defining feature has not arisen within a natural group, it is said to be a grade, e.g. flightless birds (flight has been given up by many unrelated groups of birds).

Characters are attributes or features of organisms or groups of organisms (taxa) that biologists use to indicate relatedness or lack of relatedness to other organisms or groups of organisms. A character can be just about anything that can be measured from a morphological feature to a part of its genetic makeup. Characters in organisms that are similar due to descent from a common ancestor are known as homologues and it is crucial to systematics to determine if characters under consideration are indeed homologous, e.g. wings are homologous if we are comparing two birds, but if a bird is compared with, say, a bat, they are not, having arisen through convergent evolution, a process where structures similar in appearance and function appear in unrelated groups of organisms. Such characters are known as homoplasies. Convergences are not the same as parallelisms which are similar structures that have arisen more than once in species or groups within a single extended lineage, and have followed a similar evolutionary trajectory over time.

Character states can be either primitive or derived. A primitive character state is one that has been retained from a remote ancestor; derived character states are those that originated more recently. For example the backbone is a defining feature of the vertebrates and is a primitive state when considering mammals; but the mammalian ear is a derived state, not shared with other vertebrates. However these things are relative. If one considers Phylum Chordata as a whole, the backbone is a derived state of the vertebrates, not shared with the acrania or the tunicates. If a character state is primitive at the point of reference, it is known as a pleisiomorphy; if it is derived it is known as an apomorphy (note that “primitive” trait in this context does not mean it is less well adapted than one that is not primitive).

Current schools of thought in classification methodology

Biologists devote much effort to identifying and unambiguously defining monophyletic taxa. Relationships are generally presented in tree-diagrams or dendrograms known as phenograms, cladograms or evolutionary trees depending on the methodology used. In all cases they represent evolutionary hypotheses i.e. hypotheses of ancestor-descendant relationships.

Phenetics, also known as numerical taxonomy, was developed in the late 1950s. Pheneticists avoid all considerations of the evolution of taxa and seek instead to construct relationships based on overall phenetic similarity (which can be based on morphological features, or protein chemistry, or indeed anything that can be measured), which they take to be a reflection of genetic similarity. By considering a large number of randomly-chosen phenotypic characters and giving each equal weight, then the sums of differences and similarities between taxa should serve as the best possible measure of genetic distance and hence degree of relatedness. The main problem with the approach is that it tends to group taxa by degrees of difference rather than by shared similarities. Phenetics won many converts in the 1960s and 1970s, as more and more “number crunching” computer techniques became available. Though it has since declined in popularity, some believe it may make a comeback (Dawkins, 1986).

By contrast, cladistics is based on the goal of producing testable hypotheses of genealogical relationships among monophyletic groups of organisms. Cladistics originated with Willi Hennig in 1950 and has grown in popularity since the mid-1960s. Cladists rely heavily on the concept of primitive versus derived character states, identifying homologies as pleisiomorphies and apomorphies. Apomorphies restricted to a single species are referred to as autapomorphies, where as those shared between two or more species or groups are known as synapomorphies.

A major task for cladists is identifying which is the pleisiomorphic and which is the apomorphic form of two character states. A number of techniques are used; a common approach is outgroup analysis where clues are sought to ancestral character states in groups known to be more primitive than the group under consideration.

In constructing a cladogram, only genealogical (ancestor-descendent) relationships are considered; thus cladograms may be thought of as depicting synapomorphy patterns or the pattern of shared similarities hypothesised to the evolutionary novelties among taxa. In drawing up a cladogram based on significant numbers of traits and significant numbers of taxa, the consideration of every possibility is beyond even a computer; computer programs are therefore designed to reject unnecessarily complex hypotheses using the method of maximum parsimony, which is really an application of Occam’s Razor.

The result will be a family tree – an evolutionary pattern of monophyletic lineages; one that can be tested and revised as necessary when new homologues and species are identified. Trees that consistently resist refutation in the face of such testing are said to be highly corroborated.

A cladogram will often be used to construct a classification scheme. Here cladistics differs from traditional Linnaean systematics. Phylogeny is treated as a genealogical branching pattern, with each split producing a pair of newly-derived taxa known as sister groups (or sister species). The classification is based solely on the cladogram, with no consideration to the degree of difference between taxa, or to rates of evolutionary change.

For example, consider these two classification schemes of the Phylum Chordata.

Classification Scheme A (Linnaean):

Phylum Chordata
Subphylum Vertebrata (vertebrates)
Superclass Pisces (fish)
Class Amphibia (amphibians)
Class Reptilia (turtles, crocodiles, snakes and lizards)
Class Mammalia (mammals)
Class Aves (birds)

Classification Scheme B (Cladistic):

Phylum Chordata
Subphylum Vertebrata
Superclass Tetrapoda
Subclass Lissamphibia (recent amphibians)
Superclass Amniota
Class Mammalia (mammals)
Class Reptilomorpha
Subclass Anapsida (turtles)
Subclass Diapsida
Infraclass Lepidosaura (snakes, lizards, etc)
Infraclass Archosauria
Order Crocodilia (crocodiles, etc)
Class Aves (birds)

In Scheme A, crocodiles are grouped with turtles, snakes and lizards as “reptiles” (Class Reptilia) and birds get their own separate grouping (Class Aves). This scheme considers physical similarities as well as genealogy; but the result is the scheme contains paraphyletic taxa. Scheme B strictly reflects cladistic branching patterns; the reptiles are broken up, with birds and crocodiles as a sister group Archosauria (which also included the dinosaurs). All the groupings in this scheme are monophyletic. It will be noted that attempts to append traditional Linnaean rankings to each group runs into difficulties – birds should have equal ranking with the Crocodilia and should therefore be also categorised as an order within the Archosauria; not their own class, as is traditional.

Traditional Linnaean systematics, now referred to as evolutionary taxonomy, seeks to construct relationships on basis of both genealogy and overall similarity/dissimilarity; rates of evolution are an important consideration (in the above example, birds have clearly evolved faster than crocodiles); classification reflects both branching pattern and degree of difference of taxa. The approach lacks a clearly-defined methodology; tends to be based on intuition; and for this reason does not produce results amenable to testing and falsification.

© Christopher Seddon 2008

No comments: