What do terms like monophyletic, paraphyletic and polyphyletic mean?

Question

Answer

                       Aves
                        /
                       /
          Crocodilia  /
Mammalia    \    Dinosauria
     \       \      /
      \       \    /
       \       \  /
   Synapsida  Reptilia
         \      /
          \    /
           \  /
         Amniota

Here are examples of all three types of group:

• Consider the group consisting of all the animals in this diagram - that is, Amniota. This group is monophyletic because it consists of a single animal together with all of its descendants. The Dinosauria, including the modern birds, is another monophyletic group, sometimes defined as the most recent common ancestor of Igunanodon and Megalosaurus together with all its descendants.

  Monophyletic groups are also called clades, and are generally considered as the only ``natural'' kind of group. They are very important in phylogenetic classification.

• Now consider the group consisting of the non-avian dinosaurs (which is what people usually mean by the informal term ``dinosaurs''). This is a paraphyletic group, because it can't be defined simply as ``this animal plus all its descendants'', but must be described as one clade minus another: in this case, Dinosauria minus Aves.

  The ``non-avian dinosaurs'' make up a singly paraphyletic group because only one clade need be omitted from its base definition. Groups may also be doubly paraphyletic, thrice paraphyletic, etc., depending on how many sub-clades they omit.

• Finally, consider the group of ``warm-blooded animals'', which consists of Mammalia and Aves. This is a polyphyletic group - a totally unnatural assemblage - because it can't even be expressed as a paraphyletic group, that is, a clade minus one or more of its subclades. Such groups are not used at all in phylogenetic work since they are a purely artificial construct. In terms of common descent, a ``warm-blooded animals'' grouping makes no more sense than a Synapsida-plus-Crocodilia group - though this is not to say the notion of a warm-blooded group may not be useful in some informal discussions.

So far, so straightforward. The only wrinkle in this scheme is that some workers use the word ``monophyletic'' in a sense that includes what we have described here as paraphyletic groups. These people then use ``holophyletic'' to describe what are usually called monophyletic groups. It's tempting in the face of this ambiguity just to abandon the word ``monophyletic'' and use a holophyletic/paraphyletic dichotomy, but this terminological abuse is probably not widespread enough to merit such extreme measures. It's just something to be on the watch for.

Because clades are so important, there is common notation for specifying them (taken from the Phylocode: see note 9.4.1 in http://www.ohiou.edu/phylocode/art9.html).

• Clade(A+B) is a node-based definition meaning ``the most recent common ancestor of A and B, together with all its descendents'', or equivalently, ``the least inclusive clade containing A and B''. This may be extended to forms such as Clade(A+B+C), Clade(A+B+C+D), etc., in the obvious way.

  For example, Neosauropoda is defined as Clade(Saltasaurus + Diplodocus), that is, the most recent common ancestor of Saltasaurus and Diplodocus together with all its descendants. And Eutitanosauria, the ``true titanosaurs'' can be defined as Clade(Saltasaurus + Argyrosaurus + Lirainosaurus).

• Clade(A<--B) is a stem-based definition meaning ``A together with everything that shares a more recent common ancestor with A than with B''. This too may be extended into forms such as Clade(A<--B, C)

  For example, Coelurosauria is often defined as Clade(Neornithes <-- Allosaurus), that is, modern birds and everything sharing a more recent common ancestor with them than with Allosaurus. And Eusauropoda, the group of ``true sauropods'', is defined essentially by listing a lot of taxa that are not included in it, as Clade(Saltasaurus <-- Barapasaurus, Ohmdenosaurus, Vulcanodon, Zizhongosaurus).

  Stem-based clades are useful for neatly partitioning a node-based clade. For example, within the Avetheropoda, which is defined as Clade(Neornithes + Allosaurus), the two subgroups are the Carnosauria, defined as Clade(Allosaurus <-- Neornithes), and the Coelurosauria, defined as its complement: Clade(Neornithes <-- Allosaurus).

• Clade(X in A) is an apomorphy-based definition meaning ``the first species to possess character X synapomorphic with that in A, together with all its decendants.''

  Whatever the hell that means.

These notations are not standard in formal technical literature, but appear frequently on the Dinosaur Mailing List.

It's unfortunate that this notation is so clumsy. The following, more concise, alternative notation is sometimes used (notably in Mike Keesey's admirable Dinosauricon):

• {A+B} is equivalent to Clade(A+B)
• {A>B} is equivalent to Clade(A<--B). (Yes, the arrow points in the opposite direction. That's a tragedy, but we seem to be stuck with it.)
• There's no way to indicate apomorphy-based definitions, but that's OK because no-one ever uses these anyway.
• {A-B} can be used, in the absence of the prejuducial word ``Clade'' in the notation, to indicate the paraphyletic group formed by removing the clade B from the enclosing clade A. Mildly naughty, but pragmatic.

As examples of this last, we might describe the informal grouping ``non-avian dinosaurs'' as {Dinosauria-Aves}, and the ``traditional reptiles'' as {Reptilia-Aves}.