Message 2000-10-0027: Defining species

Mon, 23 Oct 2000 15:45:39 -0700 (PDT)

[Previous by date - Re: Hybrid specifiers]
[Next by date - [unknown]]
[Previous by subject - Death of the PhyloCode?]
[Next by subject - Descendents of a species]

Date: Mon, 23 Oct 2000 15:45:39 -0700 (PDT)
From: Nathan Wilson <>
Subject: Defining species

On Fri, 20 Oct 2000, Kevin de Queiroz wrote: 

> Nathan stated:
> I'm not sure how relevant this is to the issue we've been discussing,
> but I would like to know why Nathan thinks there is no good definition
> of the term "species" (since I have argued to the contrary in several
> papers).

The crux of this issue may simply be in what we each consider to be 'good
definition'.  For me a good definition for 'species' would be one which
gives a consistent and exact line between individuals that are in a
species and ones which are not in a species. 

To directly answer your question, the source of my current opinion about
the definition of species is fairly broad including undergraduate
textbooks, popular press such as E.O. Wilson's _The Diversity of Life_, as
well as personal dialog with various fungal, bacterial and plant
taxonomists.  I was under the impression that almost all taxonomists felt
that the concept of species could not be defined perfectly.  In
particular, it is not possible in all cases to draw clear lines between
species.  I've even heard it argued by a plant taxonomist that the
existence of clearly defined species within a genus may be an indicator
that that genus is an evolutionary dead-end. 

My understanding of the most widely accepted definition of species is a
population of organisms that has no genetic exchange with any organisms
outside of that population.  (I'm perfectly happy ignoring lateral gene
transfer among bacteria for the sake of this definition.)  The problem
with this definition is that it breaks down during any process of gradual
speciation where two geographically isolated populations slowly drift
apart until some type of reproductive barrier is created between the
populations.  During this process there is necessarily a time when there
is potential genetic exchange, but none happens.  The question is at what
point do the two populations become separate species? 

When I look at this problem as an abstract computer science problem, at
the simplest level of description, the evolutionary 'tree' is a collection
of individuals (for the sake of this description clones are the same
individual) some of which can (or could) exchange genetic information
directly and others cannot.  Each individual has a set (typically two) of
parent individuals. 

The concept of speciation in this structure is a secondary phenomenon that
relates to the rules which determine which individuals can and cannot
exchange genetic information.  These rules are complex and vary
significantly from punctuated transitions where children cannot mate with
parents (changes in chromosome number might be an example of such a
change) to gradual chemical changes that prevent fertility to learned
behavioral changes to slight color or structural changes that change which
insects are attracted to a flower. 

Given all of this complexity, I was struck by the possibility that you
could base a naming system on natural groupings within the evolutionary
`tree' that depend solely on the actual relationships between the
individuals rather than depending on the formation of barriers to the
exchange of genetic information.  I was further struck by how close the
node-based and stem-based definitions used in the Phylocode are to valid
definitions of this type.  My belief (and I'm pretty sure it could be a
mathematical proof) is that the structures revealed using my slightly
modified version of these definitions will include all mono-phyletic
species (using any definition of species) as well as all larger
mono-phyletic groupings that actually exist in the evolutionary `tree'. 

My opinion is that the Phylocode would be significantly strengthened and
in no way weakened by adopting these more fundamental structural
definitions and dispensing with the reliance on any species concept.



Feedback to <> is welcome!