I take your points, Jane (though I may have some reservations about your
curve-fitting example), but my point, backed up by my examples, is that
branching systems represent MANY patterns.  I think this is especially
important when we consider anastomoses (cross-connections).  When there are
enough of them, we have a network (like capillary beds, or strangler-fig
roots), not a topological tree, but I don't know of any good way of
deciding where to draw the line.

Martin

2012/12/18 Jane Shevtsov <[email protected]>

> Hi Martin,
>
> On Mon, Dec 17, 2012 at 3:17 PM, Martin Meiss <[email protected]> wrote:
>
>> While there similarities between river branching systems and various
>> biological branching systems, there is no meaningful level of analysis at
>> which the can be said to be the same.  Branching systems can be
>> characterized by number of orders of branching (and there are various ways
>> of counting orders), diameter ratios between parent branches and daughter
>> branches, diameter of daughter branches as a function of branching angle,
>> distances between branching points, number of anastomoses, etc.
>>  Biological
>> branching systems represent engineering solutions to problems of
>> optimizing
>> mechanical strength, gas exchange, fluid dynamics (energy efficiency),
>> redundancy (e.g., finding alternative routes in case of blockages),
>> mechanical flexibility, repair mechanisms, and more.  Even trees (in the
>> botanical sense) show vast diversity in the details of their branching
>> systems: pines vs. maples vs. strangler figs vs. rose bushes.  Then there
>> is the diversity of branching systems in animals: airways, nerves, axons
>> and dendrites, arteries, veins, capillary beds, and more.
>>
>> Given this diversity, is it reasonable to refer to branchings systems as
>> "a
>> pattern"?
>>
>
> Yes. Absolutely yes.
>
> All the differences you point out are quantitative differences. In some
> cases, they're biologically important; in other cases, not. But there's a
> reason we can refer to all these things as "branching systems" -- their
> strong qualitative similarity. It's branching, not stripes, spots or
> something else. Now, there could well be several ways to get branching, but
> this just adds another layer to our classification.
>
> What many contemporary biologists don't realize is that qualitative is
> more important than quantitative. Think of it this way. Suppose you have a
> dataset that exhibits cycles. You have one dynamical model of the system
> that produces cycles of the wrong amplitude and frequency and another that
> produces a flat line. The second model may give a larger R^2, but the first
> model is the better one. It has captured something important about the data
> (the fact that it cycles) that the second model completely missed.
>
> It's a bit ironic that my example comes from math (qualitative dynamics).
> The old anatomists would have understood. So, likely, would some
> contemporary developmental biologists.
>
> Jane Shevtsov
>
> --
> -------------
> Jane Shevtsov, Ph.D.
> Mathematical Biology Curriculum Writer, UCLA
> co-founder, www.worldbeyondborders.org
>
> “Those who say it cannot be done should not interfere with those who are
> doing it.” --attributed to Robert Heinlein, George Bernard Shaw and others
>

Reply via email to