On Wed, 18 Jun 2003 10:21:25 -0400, "H.M. Hubey"
<hubeyh@...> wrote:

>Ok, I see what you mean now. When comparing forms without looking into
>history one can easily make mistakes. The thing is that I think the errors
>are symmetrically distributed (e.g. like the Gaussian) so that given enough
>data they will cancel out.

Oh my god.


How to reconstruct a Brontosaurus, the Hubey way:

- Take as many bones as you can lay your hands on (doesn't matter from
which animal or which part of it)

- Randomly glue them together.

Since a Brontosaurus has Gaussian distribution (in the immortal words of
Anne Elk (Miss): "All brontosauruses are thin at one end, much thicker in
the middle and then thin again at the far end"), given enough bones, the
errors will cancel out.

Next week's recipe: Tuna salad à la Sumérienne.


=======================
Miguel Carrasquer Vidal
mcv@...