Re: Morgan Tears 3.

H. M. Hubey (
30 Oct 1995 22:01:05 -0500

David Froehlich <> writes:

>I hope you are not serious. The last common ancestor of elephants and
>seals is some sort of small insectivourous mammal, probably in the
>Cretaceous. The last serious contender for aquatic proboscidean is
>Moeritherium or Numidotherium from the Eocene of North Africa. All
>subsequent proboscideans are much more terrestrial.

Just out of curiosity, does any doubt ever enter into your
mind when you say this?

I mean, let's examine the facts. WE have some bones, and some of them
resemble others. Some are large, and some are small. We look at the
whole bunch and then try to trace out some kind of a family tree
based on not much more than gazing at the bones, and imagining some
getting larger and some smaller, some bones stretching and chaning
morphology, etc.

The last time I even imagined a problem close to this was a matching
or a clustering problem. Given n objects with a single measurement
for each object try to put them in some kind of an order of
"relatedness" based on some kind of difference or distance.

IF these were statistical problems what kind of level of confidence
would you place on these statements? Or is this "science" still
some distance from even realizing the complexity of the problem
and the possibility of errors creeping into the bone similarity
matches? The more I think, the more curious I get. Are complete
skeletons of some of these putative ancestors and their
close relations available in all cases, or is the middle filled
in via lots of imagination and hope.


Regards, Mark