What is a species

[John Trueman]

In response to my suggestion re conceptualising species as sets of
populations having "a separate evolutionary future", Richard Jensen asked
>What about "minimally diagnosable entities"?

That is a good question!  I see three problems with the use of 'minimally
diagnosable entity' as a species concept or in species diagnoses:

First, many diagnosable entities ought not to be called species.
Daschund/greyhound might be one example.

Second, diagnosability results from the past aquisition of differences but
says nothing about whether those differences will persist.  One reason why
daschunds and greyhounds ought not to be considered two separate species is
that although their separate pasts result in present differences there is a
reasonable expectation that, over evolutionary time, they are going to
share a common future.

Third, a lack of diagnosability at the present time would be no guarantee
that two entities are not already on separate evolutionary trajectories.
The aquisition of evolutionary novelty, hence diagnosability, may long
post-date the separation of gene pools. [As a first approximation, species
splitting occurs at the branching points on the tree of life but (at least
under some processes of speciation) the aquisition of novelties occurs on
the internodes.]


I would urge that whenever the meaning of 'species' comes under discussion,
a clear distinction be drawn between
(1) entities which have had separate pasts, and
(2) entities which will have (ie, reasonably can be expected to have)
separate futures.

Three subcategories of the former class are
(1a) entities which have accumulated one or more differences and therefore
are diagnosable. (Useful if the entity is to be named.)
(1b) entities which have accumulated differences which allow the history of
relationships to be estimated. (Useful if the entity is to be classified.)
(1c) entities which are historically unique but remain diagnostically
intransigent. (Useful, eg, in modelling evolutionary processes or in
counting how many separate gene pools there are.)

And two subcategories of the latter class are
(2a) entities whose separateness extends back over evolutionary time.
(Likely to be in category 1a or 1b. Useful concept for counting present
diversity.)
(2b) entities only recently separated (possibly by human action).


However,in the end, different species concepts and different species
recognition procedures always will be necessary for different biologic
purposes.

======================================================================
John Trueman
Bioinformatics Unit
Research School of Biological Sciences
Australian National University        -------------------------------------
Canberra, ACT 0200,  AUSTRALIA       | Parsimony is nowhere inconsistent  |
                                     | it's your aligment which is wrong. |
ph: +61 6 249 4840                    -------------------------------------
fax: +61 6 249 5573
=======================================================================