Thursday, August 30, 2012

The Great Botanical Butter Battle Book

In recent decades, there have been many often drastic changes in the taxonomic system of flowering plants (and indeed all organisms).  This is due to radical changes in taxonomic philosophy that began in the middle of the 20th Century, and which were in turn the result of vastly improved techniques for determining the pattern of evolutionary relationships among organisms.  Usually referred to as phylogenetic taxonomy, the primary philosophical goal of this approach is to make our classification system reflect the evolutionary history, or phylogeny, of organisms as exactly as possible.  Our traditional classification system deviated from this ideal in many of its details, and so the classification of organisms has changed considerably in the last 50 years. 

In a recent posting, I described the fate of the Snapdragon Family, which has been taxonomically “blown to smithereens” because it was polyphyletic.  A polyphyletic taxonomic group (taxon) is one that has members descended from very different ancestors, i.e. their resemblance to one another is superficial. Similar body forms in unrelated organisms with very different anatomy and chemistry, not to mention genetics, is due to independent adaptation to similar environmental conditions, or convergent evolution.  The former classification of whales and porpoises, along with tuna and salmon, in the category of “Fishes” is an extreme example.

No one really disagrees that such strange bedfellows need to be separated and returned to their true relatives, i.e. to be arranged in monophyletic categories - by definition a common ancestor and all its descendants.   Much of the taxonomic busywork of the past century has been to ferret out illicit polyphyletic associations and rearrange organisms into a classification based on true relationship.  That’s not where the butter battle comes in, however.

 The classic Butter Battle Book by Dr. Suess satirized the human propensity for escalating silly disagreements into cataclysmic space-time-continuum-threatening warfare.  The story began with an argument between two groups of whimsical creatures, one of which believed passionately that toast should be eaten butter-side up, and the other that it should be eaten butter-side down.  Something akin to this feud began in the botanical world during the latter part of the 20th century, and it’s not over yet.  It wasn’t about polyphyletic groups, but about a much subtler distinction that results in some traditional taxa being called “paraphyletic.”

Both traditional and phylogenetic systems of classification have the form of a hierarchy of categories.  The level of the hierarchy is referred to as ranks.  Domains are currently the highest rank of taxon, and they are subdivided into Kingdoms.  Progressively smaller subcategories are given lesser rank.  The  principle ranks are Domain, Kingdom, Phylum, Class, Order, Family, Genus,  and Species, and in-between ranks (subclass, superorder, etc.) can be inserted as needed.  

Phylogenetic taxonomy has two basic rules designed to make a classification system explicitly reflect evolutionary history:

1.  Only monophyletic groups - by definition a complete branch, or clade, of the phylogenetic tree, with the common ancestor and all of its descendents, can be formally recognized as taxa at each level of branching.  
 2.  Each branch point on a cladogram results in two sister taxa, which must be given equal rank (genus, family, class, etc.).  

   Applying these two rules, it was found that many traditional taxa were neither polyphyletic nor monophyletic, but fell into a fuzzy gray area called paraphyletic.  Paraphyletic taxa are incomplete clades (violating rule 1) - some part of the clade that should be just a subclade has been improperly promoted to the same rank as the main clade (violating rule 2 since they are not sister clades).
The traditional classification of higher vertebrates was
simple and intuitive, but misleading about
evolutionary relationships.  Clades A through F
retain certain ancestral characteristics, such as
scaly skin, that we associate with reptiles, while
the birds evolved highly distinctive wings, flight
feathers and other specialized features
associated with flight.
     This is probably as clear as mud at this point.  Let me use a familiar  example from the animal world to quickly explain this concept.  Reptiles and birds are two of the traditional classes of vertebrates, but modern studies have overwhelmingly confirmed that birds descended from reptiles, in particular from a group of dinosaurs that included velociraptors and Tyrannosaurus.   While birds are a monophyletic group, the reptiles are paraphyletic (also referred to as a grade to differentiate it from a clade).  

In the greatly simplified diagram to the right, clades A through F are different groups of reptiles.  Technically, the sister group of birds is clade F, which includes their closest dinosaur relatives.  Therefore, according to the rules, birds must be have the same rank as group F.  Birds are still a distinct group, but taxonomically must be considered just a highly specialized subgroup of reptiles.  Note also that the traditional classification obscures the fact that group F is much more closely related to birds than it is to group A (turtles and their relatives).  Remember that it is the goal of phylogenetic taxonomy to better reveal these relationships in the classification system.

Taxonomically, there are actually three choices in such a situation:  1) keep the status quo, recognizing the paraphyletic Reptilia and the monophyletic Aves as two equivalent classes.  This would be the preference of some traditional taxonomists; 2) combine the reptile and bird classes together into one class (“lumping”), in which birds become one of the subclasses of reptiles along with groups A through F; or 3) keep the birds as a class, but also promote each of  the main clades of reptiles to the rank of class, creating several additional classes (“splitting”). 
An acceptable solution in phylogenetic systematics
is to lump reptiles and birds into a single class, with
birds recognized as a subclass (along with clades A
 through F)

The alternate solution recognizes many classes of
reptiles along with the birds.
The second two options are equally acceptable to phylogenetic taxonomists, but this creates a dilemma with each paraphyletic situation encountered: Do we fragment the paraphyletic taxon (reptiles) into a number of smaller classes or submerge the distinctive group at the top (birds) into one big class with the reptiles?  Either way, the taxonomy reflects the phylogenetic history of the group, but one way ends up with a lot of smaller classes (harder to remember them all) or fewer big classes (obscuring the distinctive, sometimes revolutionary, adaptations of the top group).  This ranking decision is pretty much subjective.

This is all about the formal classification used primarily by scientists.  We can still talk about reptiles and birds informally.   No one is obligated to include birds in a field guide of common reptiles, for example. 

To finally get around to plants, I'll mention first in passing that the traditional classification of flowering plants into dicots and monocots created exactly the same situation as the reptiles and birds.  Dicots turned out to consist of a grade of separate clades, with monocots occupying the bird position at the top of the tree.  It still has not been fully agreed upon as to how to rank these clades.

Walter Judd, Roger Sanders and Michael Donoghue, in a truly landmark paper in 1994, identified a number of pairs of traditional angiosperm  families, of which one was paraphyletic and theo other monophyletic.  In most cases, the monophyletic family at the top of the tree was primarily temperate in distribution and consisted primarily of herbaceous plants, while the clades branching along the grade consist mainly of woody trees and shrubs.  The pairs include: Apocynaceae (oleander family)/Asclepiadaceae (milkweed family), Araliaceae (Aralia family)/Apiaceae (carrot family), Capparaceae/Brassicaceae (mustard family), among others.  To keep this posting under control, I’ll just talk briefly about the first case.

The Apocynaceae and Asclepiadaceae share a number of characters inherited from their common ancestor.  Each clade that branched off of the Apocynaceae grade had some distinctive characters, but were not deemed worthy of family status.  The upper subclade we call Asclepiadaceae, however, had some very innovative and distinctive floral characters that allowed this group to blossom into a “family” of many related genera.  The conclusion of Judd, Sanders, and Donoghue was however that the Asclepiadacease should be submerged into the Apocynaceae.
The traditional Apocynaceae, like this
Pachypodium, have relatively
conventional flowers with petals united
into a flaring tube that often twists like a

I’ve provided this lengthy explanation as a prelude to getting into the real butter battle, which may take one or two more posts to explore fully.  It is first of all a battle between two views of how we divide up an undisputed diagram of relationship into taxonomic entities that are convenient, practical, and informative.  Phylogenetic taxonomists, in forbidding  paraphyletic taxa, are emphasizing information on phylogenetic branching patterns.  Clades are real entities that represent evolutionary history and groups of related organisms.  They are predictive of the characteristics to expect within the group.  Recognizing birds and reptiles as equivalent groups obscures the fascinating relationship of birds to dinosaurs, and the fact that Tyrannosaurus rex is much more closely related to ostriches than it is to turtles. 
The Asclepiadaceae, represented by this milkweed
(Asclepias), has the same basic features as the Apocynaceae,
but adds some unique floral features related to specialized
modes of pollination.

 Traditional taxonomists, while probably not much concerned about the reptile/bird issue anymore, believe that allowing some paraphyletic groups to be recognized as taxa is not only harmless, but also desirable in order to simplify classifications and create user-friendly, easily recognizable taxa, and  to recognize highly innovative groups, like the birds or the Asclepiadaceae, as distinctive taxa of higher rank.  Birds have achieved the level of distinctiveness that is worthy of a new class of organisms, while the other clades of reptiles are still basically reptilian.  The differences between them are technical and not readily evident to the non-expert.   They point out that if one wants to see the actual pattern of relationship, one can look directly at the phylogenetic trees.  There is no need to create cumbersome, non-intuitive classifications just to duplicate the information that is easier to see in the diagrams themselves.  This sentiment is probably stronger among botanists that among zoologists.  

  Phylogenetic taxonomists appear to be winning the day, but there are many holdouts among traditionalists.  For the latter, the distinctions between reptiles and birds, dicots and monocots, Apocynaceae and Asclepiadaceae, are all simple, intuitive,informative, and useful.

There is, moreover, another dimension to the debate brought up by traditionalists that is not so easily dismissed, and which we might sum up as the “ancestor problem.”  This arises when one views, theoretically at least, the entire sweep of evolutionary history.  The first member (common ancestor) of every genus (or family, etc.) logically had ancestors that were in an older genus.  Whenever a new genus-worthy group of species became distinct, it would have rendered its ancestral genus paraphyletic (just as the evolution of birds as a distinct new class of organisms rendered the reptiles paraphyletic).  So the evolution of new genera (or families or classes) appears to be impossible according to the rules of phylogenetic taxonomy, or else to force the retroactive fragmentation of the older taxa. I will take up this tricky issue in a future posting (See "Making the Ancestor Problem Go Away,"  October 18, 2012).