Re: Sergey Gavrilets and the adaptive landscape

in article dkebci$2f8a$1@xxxxxxxxxxxxxxxxxxx, phillip smith at
philge@xxxxxxx wrote on 11/3/05 4:52 PM:

> I have started reading a copy of this book and though I have only part
> way through chapter 2. So maybe I am being a bit premature with my
> criticism.
>
> There are some area's I would have to take issue with. Perhaps these
> are addressed later on in the book. But I would have thought that a
> book on fitness landscapes would have explained early on what fitness
> is. With out a clear description of what this nebulous concept is for
> the purpose of his models, I find it hard to understand the value of
> the model.

I agree, although I wouldn't fault Gavrilets much for this. Fitness is an
inherently nebulous concept IMHO, and the theory of natural selection has
been developing with this limitation from the beginning.

> There are some disturbing statements in the book. for example
> page 32 "it is straight forward to find mathematically the average
> fitness of a population if we know the individual fitness"
> Unfortunately it is impossible to know the individual fitness of any
> individual never mind every individaul the whole population.

Empiricists routinely deal with this problem by measuring concrete
components of the nebulous factor we call fitness. Lifetime reproductive
success is one of the best measures used, but there are few systems where we
can obtain this information. Other measures are often used operationally,
like survival until fledging in birds. These operational measures are
generally agreed to be important components of fitness, and thus reasonable
proxies, even though we never know how these components interact to form
fitness itself. If this was clear, the concept of fitness would not be
nebulous. IMHO, this is not a failure of appreciating how the components
interact from the bottom up to yield fitness; rather it is that the concept
of fitness itself is inherently nebulous and cannot be 'fixed'.

I readily admit that my view of fitness is not held universally. John Edser
strongly disagrees, for example. However, I have yet to hear any arguments
that erode my confidence in this viewpoint.

> The other
> problem with average fitness with regards to fitness landscapes is
> that we need know the fitness of every possible individual, inorder to
> describe the landscape which means a infinite population. So how we can
> calculate the fitness of unknown individuals I don't know.
> Perhaps these issues are addressed later in the book. I hope so.

The fitnesses of all individuals can be known when exploring a model through
computer simulation, or by assumption in analytical models. These are the
approaches taken by Gavrilets. I don't see his model being tested in
biological systems by trying to measure everyone's fitness; however, I think
there are other sorts of predictions that would emerge from this paradigm
that could be tested. I have ordered the book myself, and I will have to
read it before I could suggest what kinds of testable predictions could be
made.

Regards,

Guy Hoelzer


.

Relevant Pages

  • Re: Sergey Gavrilets and the adaptive landscape
    ... book on fitness landscapes would have explained early on what fitness ... individual never mind every individaul the whole population. ... It has a very strong implication for the adaptationist versus drift ...
    (sci.bio.evolution)
  • Re: Sergey Gavrilets and the adaptive landscape
    ... > book on fitness landscapes would have explained early on what fitness ... > individual never mind every individaul the whole population. ... to be how they are (and how inEPT scientists/philosophizers were in the ...
    (sci.bio.evolution)
Quantcast