Such expressions as that famous one of Linnæus, and which we often meet with in a more or less concealed form, that the characters do not make the genus, but that the genus gives the characters, seem to imply that something more is included in our classification, than mere resemblance. I believe that something more is included; and that propinquity of descent,—the only known cause of the similarity of organic beings,—is the bond, hidden as it is by various degrees of modification, which is partially revealed to us by our classifications (Darwin, 1859, p. 413f).

Thursday, 23 October 2008

Complexity, Pattern & Process

Species are biologically complex. The Merriam-Webster Online Dictionary define Complex as "a whole made up of complicated or interrelated parts", that is 'interrelated parts' that are "consisting of parts intricately combined". Species on the other hand are harder to define.

I will not define a species here. There are enough species concepts to go around and adding another one will not aid the growing problem of understanding species. The difficulty that many have with species is that they do not classify well. This presents us with a problem: if classification makes sense of complexity, then why can't we classify species? Why, for instance, do some claim species are never monophyletic? If species are presented by genealogical and therefore reticulated lineages that include ancestors and descendants, it will be impossible for us to classify them - to divided up related individuals into separate groups. The same argument can be made of genera and families too, but we are able to classify them. Why do species present us with this problem. The answer lies in what we mean by pattern and process.

A pattern is a repeating relationship. The character-states 0(11) for instance represent a relationship or homolog that relate the taxa A, B and C. If A = 0, B= 1 and C = 1, the relationship can be expressed as A(BC). If this taxic relationship occurs many times in different characters, it forms a pattern or homology. Process however is harder to define.

Returning to the Merriam-Webster dictionary we find the process means "a natural phenomenon marked by gradual changes that lead toward a particular result". For example, ontogeny is a process. What of the processes we are unable to see or measure? These processes can be discovered through patterns. The homology A(BC) is a discovery of evolution. That is a common and shared history. A 'hidden' process, such as the sexual behavioral traits of hadrosaurs is completely unknown to us (if they had any at all). We are able to model these traits based on assumptions, wild guesses or on the behavior of related living taxa such as birds. What ever result is generated (artificially produced) will never represent a discovery. We will never see such behavioral traits, so our models are merely speculations. The same is true for genealogical relationships. We only know who we are related to simply by observation, written documentation or by word-of-mouth. Through our DNA we are able to discover how similar we are to other people, either dead or alive, but we will never know if they are our direct descendants. Genealogy, as an unobserved process, is therefore often hypothesized. In order to make that process consistent we assign a rational hypothesis or explanatory mechanism. Genealogical relationships that move beyond our understanding (i.e., observation, recorded history etc.) rely on this hypothesized explanatory mechanism.

Explanatory mechanisms (unobserved processes) do not discover patterns. They generate artificial and ad hoc hypotheses. Observed processes however do. They provide a more robust explanation, but offer little in the way of a direct or trivial narrative (e.g., who begot whom).

Let us return to species. We assume that they are interbreeding classificatory 'units' real or otherwise. Species that are defined on an unobserved 'process' are mechanical. For some this helps conceptualize a species, for others it has little to do with classification.

Biological classification is based on patterns and the observed processes that help to uncover them. We may infer explanatory mechanism from patterns, but we will never be able to discover patterns from explanatory mechanisms. Species, as defined by explanatory mechanisms, have no place in classification. As arbitrary taxa (like genera or families), however, they make perfect sense.

Saturday, 11 October 2008

The Evolution Slogan

The term "evolution" can be used recklessly in a variety of ways: "If evolution was outlawed, only outlaws will evolve", "Paraphyly is evolution all the way" (Brummitt, 2002:40) and most recently, "Because we understand how evolution happens, we can also guess where it will go next" (Jones, 2008; see also John Wilkins's post). The two main points of contention, highlighted in the latter statement by Steve Jones, are our "understanding" of evolution and our ability to "guess".

The late Colin Patterson, ichthyologist at the then British Museum, Natural History in London, gave a presentation that questioned the term. The talk, titled "Systematics and Creationism", was given at American Museum of Natural History (Patterson, 2002) in November 1981. There Patterson noted:
"...the theory is evolutionary theory, descent with modification" (Patterson, 2002:23; see also Martin Brazeau's post)
Combined with the above slogans we may suggest that 'evolution' is: a process of descent with modification that results in paraphyly problem arises. How do we see this process?

In order to know we need to be able to observe or measure. Paraphyly, for instance, cannot be observed. It exists only when an artificially delineated taxonomic group is discovered to be monophyletic (homologous) - like 'invertebrates' or 'aliens'. Descent with modification is also difficult to see in action. Although we can see genealogy and ontogeny, they do not constitute 'descent with modification', at least not in the way Jones uses the term.

What systematists and biogeographers know is that evidence for evolution is based on retrodictions - that is past 'predictions' or patterns. These patterns are homologies or relationships - evidence for evolution. Our task as systematists is to discover whether our groups are a result of evolution, rather than poor taxonomy. Evolution should not be taken for granted - just because we know it exists doesn't mean we should stop looking. Reptiles, for example, are not an evolutionary group. They are a poorly defined taxonomic group like 'insectivores' and 'creepy, crawling things'. Discovering that taxa within the reptilia share closer relationships with taxa in mammalian than with any other taxon does not validate reptiles as an evolutionary group. The task of herpetologists is find those evolutionary groups and, not to defend existing names that have no evolutionary significance. Patterns, homologies, relationship and monophyletic groups are all the same thing: evidence for evolution.

Now we return to Jones. He, like many other evolutionary biologists, has committed a classic error - assuming that life progresses from an incomplete to complete phase: also known as 'primitive to derived'. A typical example is the 'primitiveness' or 'plesiomomorphy' of Archaeopteryx lithographica. The half bird-half reptile is always considered to be transition - fossilized in the middle of evolving. Like all living things dead or alive, Archaeopteryx is perfect in its own right. It has no hidden agenda, no purpose other than to be Archaeopteryx. If we were to assume, unwittingly and in hindsight, that it was primitive, then we are advocating some purpose or teleology, namely that Archaeopteryx was aiming to become a bird. This sort of thinking gives evolution a bad reputation and opens it up to attack from protagonists of anti-science. The logic behind it does not work. Let us assume for the moment that we could go back in time, back when Archaeopteryx was alive. We would assume, that this is a highly evolved 'reptile', a derived form. See the problem? Archaeopteryx is both derived and primitive at the same time in form and space but not in geological time. The whole 'primitive – derived' argument is based stratigraphic sequence and not evolution (homology).

To counter Jones's argument - we are complete, so is Archaeopteryx and all other life that has ever existed and will ever exist on this planet. What does this completeness say about evolution? Absolutely nothing at all. Instead it tells us of a desire for explanation.

We may think 'nothing in biology makes sense except in the light of evolution' (Dobzhansky, 1973), but without a doubt, evolution only makes sense in the light of homology. Biological classification provides us with the tools to discover relationships and a way to understand the evolution of life. Without it we are just telling never-ending stories. I am sure that in 200 millions years time, an octopod biologist, will wonder how something as incomplete and primitive as Homo sapiens lived for as long as it did.


Brummitt, R. K. 2002. How to chop up a tree. Taxon 51: 1-41.
Dobzhansky, T. 1973. Nothing in Biology Makes Sense. Except in the Light of Evolution. The American Biology. Teacher, 35:125-129.
Jones, S. 2008. Evolution is complete: so where do we go from here? Daily Telegraph Online,
Patterson, C. 2002. Evolutionism and creationism. The Linnean 18: 15-33.

Wednesday, 8 October 2008

Naef and Cephalopod Awareness Day

Cephlapodcast has announced the Second Annual Unofficial International Cephalopod Appreciation and Awareness Day. I wonder what Adolf Naef would have said?

Naef said a lot about cephalopods and their relationships. As a respected world authority on octopuses and their relatives, Naef has produced several texts that have until now only interested David and I in parts (mostly the systematic theory and methodology). Our only foray into the fascinating world of calamari occurred recently when were contacted by Jan Strugnell - a cephalopodist from the University of Cambridge and admirer of Naef's work. Jan had an interesting question - why did Naef's phylogeny and classification of Argonautoidea differ?

The Argonautoidea Lamarck, 1809 (paper nautiluses and their relatives) are divided into two subfamilies: Tremoctopodinae and Argonautinae. The former includes Alloposidae and Tremoctopodidae and, the latter Argonautidae and Ocythoidae. Naef produced the following diagram (Figure 1a) with the following accompanying text:
"The typical relationship between the 4 genera which form the family are shown in the following graph: i.e., the genera developed by secondary specialization from forms I-IV in a direct series.
The ideal or hypothetical form I corresponds to the type of the family; form II corresponds to the type of the 3 genera, form II to the type of the 2 highest genera. Form IV is the ancestral form of Argonauta. The forms derived from II and III are clearly natural groups which can be defined, despite the specialization of their recent representatives. The important morphological relationships, however, call for a division into 2 subfamiles by drawing a line between II and II, i.e. a distinction between lower (I) and a higher (II) stage of variation. This is, of course, arbitrary ... and is only intended to introduce order for practical purposes by stressing the essential and omit characters of less importance. This division stresses the distance between forms II and III and creates the two natural subfamiles: Tremoctopodinae and Argonautinae (Naef 1972: 732).
These two conflicting statements (one pictorial and one as text) pose an interesting problem. What does the diagram represent and what was Naef thinking?

Naef believed both subfamilies to be monophyletic (natural groups). What Naef did diagrammatically shows a transition from the Haeckelian tree-thinking to modern systematics.

In order to understand what Naef did we go to the introduction of his work on the Fauna and Flora of the Bay of Naples. Firstly the graph is a phylogenetic tree as we understand it today. Back in Naef's day (prior to molecular systematics) the graph was termed a 'genealogical' tree. Phylogenetic trees were considered to be true depictions of ancestor-descendant relationships, an idea rarely entertained today. In Naef's phylogenetic tree, we see that forms are placed at the nodes. Naef's forms are types, namely "an abstract but naturally possible form from which a multitude of actually existing forms may have developed ..." (Naef, 1972: 15). Forms are not ancestor-descendant relationships but rather abstract transitional series between basic and specialized forms based on the concept of metamorphosis. Incorporate theses points together and you have a tree depicting the transition of forms that relate directly to the relationships of the taxa on the branches.


Naef is depicting a back-to-front cladogram. In Naef's diagram the forms are being shown to be transitional within a dichotomous framework (Figure 1b); very much like showing a character tree within a consensus tree at the same time. A modern representation of Naef's tree would place the forms at the nodes (Figure 1c). What at first appears to be a rooted phylogentic tree, turns out to be a poorly drawn cladogram that places too much emphasis on transitional forms than it does on taxic relationships.

Naef's tree represents a transitional period between Haeckel and Hennig. Naef attempted to save us from the Haeckel's 'oak' trees that later inspired the neoDarwinian Modern Synthesis by emphasizing the importance of natural (monophyletic) groups. Naef's methodological and theoretical work should also be celebrated on Cephalopod Awareness Day.


Naef, A. (1972). Cephalopoda (systematics). Fauna and Flora of the Bay of Naples (Fauna e Flora del Golfo di Napoli), Monograph 35, Part I, [Vol. I], Fascicle II. Washington, Smithsonian Institute Libraries.