Classification schemes

1. Ways to approach classification.

 Phenetic or numerical

Phylogenetic or cladistic (Cladistic school once known as phylogenetic systematics now merged with molecular phylogeny)

Evolutionary systematics (also known as the classic or traditional school)

2. Phenetic school: Classifying organisms according to their attributes.

Use any characteristic in principle. In practice use hundreds of characteristics.

Use nearest neighbor or nearest average neighbor analysis to cluster organisms.


Different statistic methods yield different results

Since there is no stated way to judge which tree should be accepted, this scheme is sometimes criticized as being subjective. ---The counter argument is that if enough characters are used, any difference due to methods should be diluted.

Does not distinguish between plesiomorphic (ancestral) and apomorphic (derived) characters.


 3. Cladistic. Groups sister species or those with the most recent common ancestor. Recognizes the branching hierarchy of ancestral relationships (think of stepping back through time).

Cladistics depends on identifying shared derived characters. Shared derived characters are the main evidence for phylogenetic Relationships. These are known as synapomorphies.

Type of homology.

A homology is a character shared between species that was also present in their common ancestor.

A derived homology is unique to a particular group of species (and their ancestor), while a shared ancestral homology is found in the ancestor of a group of species but only in some of its descendants.


In cladistics the emphasis is on producing a monophyletic grouping. It is a shared derived character that defines a clade. This method is the one currently accepted by most taxonomists.

In the diagram below the clades or groups are indicated at top for example, Vertebrates and Tetrapods. The shared derived characters for a clade are indicated below the clade, so the tetrapod "clade" all have 4 limbs, but no other clade has this characteristic and it is a "new" or derived trait (not an ancestral trait or trait found in an remote ancestor to this clade). Each branch point is known as a node.


4.  Problems with cladistic methods.

Some would argue that given the vagueness of the fossil record it is difficult to determine whether characters are homologous or analogous, yet alone derived or ancestral. Outgroups are used. Outgroups is a taxon outside the group under study that must have diverged before the group evolved. The rule of thumb is to use fossils to root your trees whenever possible, but outgroups can help the placement of branches with unknown or contested ancestry, or root trees based on molecular data that can then be compared to trees based on morphology

Some scientists feel cladistic schemes place too so much emphasis on ancestry that they ignore phenetics to the point that their trees appear counter intuitive.

Example: Aves as a class ceases to exist, because some recent "reptiles" branch off at the same time. Also, there is no character that can be associated with Aves (and its most recent common ancestor) that is not found in fossil "reptiles or dinosaurs". So Aves now is considered part of the Reptilia.

Another problem is number of levels need to evoke to remain strictly cladistic (recognize true branching points).

Example: Steps in assigning taxa to cladistic diagrams.


Example of a tree demanding too many taxa using the cladistic scheme we just went through.

5. Current textbook solution: Use Evolutionary scheme, unfortunately even if cladistics is explained as the accepted method. The evolutionary school uses bits of both previous schools. Actually this is the oldest scheme, with taxonomists doing what they have for centuries. "Be as quantitative as possible, try to pick homologous traits that result in monophyletic trees or at least trees that make sense for chronicling major adaptive events."

Groups species by homologies but does not distinguish between ancestral and derived homologies.

Allows paraphyletic grouping, although attempts to acknowledge such.

Claim that resulting trees are representative of "evolutionary" events. Aves considered a separate class because birds are characterized by wings and flight and so a very different "lifestyle" than other classes of vertebrates.


System lacks consistency. Several different classes of fish or swimming vertebrates.

One does not know by simply looking at a tree whether branching arrangements are based mainly on certain characters or the ancestor shared.

Actually this is the oldest system of the three, and the phenetic and cladistic schemes came about as ways to reform the subjectivity of this system.

Many textbooks superimpose the names of the evolutionary school (also known as the classic or traditional school) for taxa, although these do not formally exist in phylogenetics. This right now just adds to the confusion, although in time it may help define what are know as crown clades or the larger groupings in the evolutionary school such as kingdoms, and phyla.

names for nodes

6. In summary: The problems faced by any system.

Character conflict: Different suites of characters give different trees.

Not potentially a problem with phenetic classification but in reality can be.
Very true in some groups when tree constructed from molecular data is compared to tree constructed to morphological data.

Search for the perfect character. Are Interspersed elements or transposons the perfect character?

DNA sequences that occasionally insert themselves into new locations within a host genome.



Transposition events unlikely so expect a particular transposition only to be inserted in a particular position once.

Also if part of a transposition is lost, usually some of host genome goes with it, so losses can sometimes be detected.

In all systems need some method to choose among competing trees, although this should be less of a problem with cladistics.

Parsimony: Pick the tree that minimizes the amount of evolutionary change that has taken place.


Maximum likelihood. Estimates how likely is it that the particular tree would occur, given a mathematical formula that describes the probability that different types of nucleotide substitution will occur, and given a particular phylogenetic tree with know branch lengths .

Everything depends on the specified model of character change.

Genetic distance. Use the tree that minimizes the distance between taxa (basically the phenetic approach although when used by evolutionary taxonomists, selected characters may be used).

Sometimes a more independent method is used such as checking whether the tree fits what we know about biogeography. Uses data from allied fields, for example, plate tectonics.

However as we have already seen, the characters and statistical method used to cluster may yield different schemes. In cladistics for example, molecular trees are sometimes in conflict with structural trait trees.


One basic problem still remains: What is a character, one eye, two eyes or should it be no eyes, some eyes. Note that none of the schools worry that characteristics are adaptive, although the evolutionary school does tend to focus on differences in crown clades that represent a change in "lifestyle". This is why Aves (flight) was recognized as a class by this school.

Good summary, addition to my notes.




link to exercise