Treatment of paraphyletic taxa (1 Viewer)

Paraphyly is certainly expected to exist for a time after every speciation event or population divergence. Paraphyly encompasses a wide range of phenomenon though, and whether you use it for or against splitting depends on the type. Really, arguments for or against splitting ultimately rely on other information (how divergent and reproductively isolated populations are) rather than the paraphyly itself… or at least I think they should.

In some cases you have species A nested among several monophyletic lineages of species B, each of which is distinct and diagnosable. In cases like this, I think paraphyly supports splitting up the overall group, because this pattern would not exist if there was gene flow between the distinct populations of species B. Of course, this says nothing about the evolution of complete reproductive isolation, and it is entirely plausible that A could be incompatible while different lineages of B still are all compatible. But since we’re usually talking about allopatric populations, we move into hypotheticals.

For an example of this, here is the tree from the Caprimulgus paper used as evidence for splitting C. arizonae from C. vociferous, because of the placement of C. saturatus between them. Another example would be the Aphelocoma jays under discussion.

http://i199.photobucket.com/albums/aa220/slybirdsly/Papers/Han-2010.png

Another situation involves recently diverged species that are reproductively isolated. In this case there hasn’t been enough time for lineages to become genetically distinct. This pattern of paraphyly is usually widespread sharing of the same haplotypes with no monophyletic groups. An example includes Tympanuchus grouse. The species are divergent in breeding displays and are thus likely reproductively isolated, but without that information it would be impossible to tell if there were divergent populations or not from the genetic data alone. In these cases it is perfectly acceptable to have paraphyletic species.

http://i199.photobucket.com/albums/aa220/slybirdsly/Papers/Oyler-McCance-2010.png

A third source of paraphyly would be gene flow. I still wouldn’t use this type of paraphyly for or against splitting without knowledge of the hybrid zone and the extent of gene flow and selection against hybridization.

In answer to the numbered questions:

1 – No, it is more of a gradient. If a taxon ends of scattered all over a larger tree and the pieces are broken up among many other groups, that is polyphyly. Paraphyly is more generally talking about the decendents of one node in the tree or one taxon. I guess you could say polyphyly is extreme paraphyly.
3- You don’t really say a taxon is monophyletic with respect to itself. Monophyly of a taxon can only be determined by knowing its relationships with all closely related taxa.
4- In the One True Evolutionary Tree, every node (branching point) on the tree is monophyletic. In human-constructed taxonomy, we attempt to assign a taxon to each node in the tree but often get it wrong. Paraphyly is a human construct.
6- Not if we know about it. Any phylogenetic tree constructed with extant taxa invariably will be wrong because we are missing data on extinct species. If we know about extinct species, we can include them. If not, oh well
2&5 – Reciprocal monophyly simply means both species are monophyletic groups, as opposed to such situations where species B may form a monophyletic group within another species A, such that species A is paraphyletic but species B is monophyletic.

A good review of paraphyly would be Funk and Omland 2003:
http://www.umbc.edu/biosci/Faculty/OmlandLabWebpage/NewPages/papers/FunkOmlandARev.pdf

Mysticete said:

using paraphyly as evidence for support of separate species .

What are people's feelings for this?

Click to expand...

I think it is nonsense, and paraphyletic species are OK.

Opposite is result of the wrong over-simplification of evolution. That is:

1) evolution always goes in neat branching splitting. In fact, two lineages can hybridize and lump into a paraphyletic population until the level of a different genus.

2) that rate of evolution is constant (independently of selection, habitat etc.!). So if one lineage got status of the species, all other lineages also got the status of species. In fact it is not true.

In practice, not allowing paraphyletic species, genera, families etc. will make classification impractical. Because we will find many paraphyletic species. Not allowing paraphyletic species gives only two choices: 1) lump dissimilar birds into one species. 2) split, and have a swarm of species which are identical in most respects, often un-identifable in many plumages and have large hybrid populations.

Some people are tempted into over-splitting birds, because it may give more push for conservation of populations. In fact, mammal people made U-turn. After splitting African elephants into savanna and forest species, they re-lumnped them again. Splitting would leave many elephants as unknown species or hybrids with uncertain conservation status.

Sorry I abandoned this post...Was in Hokkaido all last week without internet

FYI Mammal Species of the World still recognizes 2 species of Elephant in Africa, and they are pretty much the closest thing in the world to a standardized checklist for mammals, so I think it is a oversimplication to say that mammal workers have abandoned recognizing two species.

Paraphyly at species level and paraphyly at higher classification levels are also two different scenarios. Paraphyly at higher taxonomic levels is almost universally frowned upon nowadays, so I approve of not recognizing paraphyletic genera, families, etc (all of which are arbitrary anyway).

Species-level paraphyly

Ross (in press). The incidence of species-level paraphyly in animals: a re-assessment. Mol Phylogenet Evol. [abstract]

Zachos 2014

Zachos 2014. Paraphyly—again!? A plea against the dissociation of taxonomy and phylogenetics. Zootaxa 3764(5): 594–596. [preview]

Ca 20% of animal species are paraphyletic!

Lets play out the scenario: what would happen, if somebody tried not to allow paraphyletic species?

One option is that 20% of described, distinctive animal species are sunk into subspecies.

The other option requires estimating, how many splits would be needed to make all species monophyletic? The best case is that every paraphyletic species can be made monophyletic by splitting into only two species. This would raise the number of species by 20%, and 33% of animal species would belong to very similar species pairs.

This is, however, unrealistic little. If average number of splits would be 5, the number of species will grow by 100% and over 50% of animal species would belong to groups of very similar species. In the example of polar bears, making monophyly among bears would require making much more than 5 brown bear "species".

One result would be that distinctive species and evolutionary isolated lineages "drown" in a mass of look-alikes. Resources for conserving an average threatened species will also drop significantly. For new interest in "new species" is unlikely to raise the total conservation capital twice, and split populations are much more likely to be threatened.

Well paraphyly based on mtDNA is not the same as paraphyly of the species tree. After correcting for introgression and incomplete lineage sorting, how much will be left?

It is interesting that in the paper by Zachos (the second of Richard's posts from today) uses examples such as whether birds really are reptiles. I do not see him argue at all at the species level - even though it is very conceivable that he intends the paper to also be applied at the species level.

At the species level, I think there are some good arguments as to why speciation can be faster in one subspecies than in others.

Niels

Ben Wielstra said:

Well paraphyly based on mtDNA is not the same as paraphyly of the species tree. After correcting for introgression and incomplete lineage sorting, how much will be left?

Click to expand...

This is exactly right.

If you define species based on their ability to interbreed, then you have to allow for paraphyletic (i.e. species tree paraphyly) species because the ability to interbreed is a symplesiomorphy. If you define species by their history, then you need not have any paraphyletic taxa.