Re: "Evolutionary Rate" from Natural Selection

I should have qualified myself. By "genus is a fairly natural level" I meant
that genera tend to be monophyletic even when done before that was a
consideration. That is to say, retrospective analyses on genera named before
we used cladistic techniques seem to show that they are natural enough.
If you will read the other recent posts in this thread, you will see
that this isn't true.

Indeed. By the way, a month or two ago I was in a debate with somebody,
where I asked for modern evidence to support various rootings of trees,
and I was told the people who did the original work hundreds of years
ago were competant and I shouldn't question them. I disagreed with that
"trust the experts" advice, and now I see you (and others in this
thread) making a similar point, that lots of traditional taxonomy is
turning out to be not just slightly mistaken but blatantly wrong, at
least at the genus level (see next below).

Now in fact in birds (other than passerines), families are much more
likely to be monophyletic than genera. Almost all families are clades,
but a large fraction of genera are not.

I see a likely reason for that: The differences between families mostly
represent ancient clade-splits where over time significant
morphological differences have accumulated in the various sub-clades,
to where it's relatively easy to decide which clusters of similar
fossils (or living features) represent undisputable clades, and if the
differences aren't that great then we would place all the various
species within a single family. I.e. definite large amount of
accumulated difference between the two groups equals two different
families, otherwise all one family.

This is especially telling since families tend to be larger than
genera, and so more vulnerable, all things being equal, to
non-monophyly.

I respectfully disagree. Yes, the more species in a group, the more
likely there's a mistake one place or another. But this is compensated
by the effect I described above, that splitting between families
represents a deep division with lots of accumulated differences that in
most cases preclude any mistakes. IMO, in most case the deep division
dominates over the more chances for a mistake. (But see next where this
domination has failed.)

In passerines, families are pretty bad too; we can't yet tell if
they're better than genera, but both are clearly poor guides to
monophyly.

I'm curious what's going on in this case. Is it a case of so very very
many species in the group that somebody has divided the group into more
families than is warranted by the criterion of deep clade-split showing
clearcut morphological differences? Did somebody find that using a deep
split the resultant families had too many species, so they were
arbitrarily broken into more familes with fewer species in each to fix
that "problem"? Or was it a case of so darned many species that experts
didn't have resources to do proper studies to get the detailed data
necessary to identify families correctly, and they chose to do a crude
preliminary bad job rather than wait tens or hundreds of years for
funding to materialize? Otherwise (if neither of those excuses) I don't
understand how it could happen.

Now I do admit that even at high levels there are some mistakes. For
example, phyla are supposed to represent very very deep clade-splits,
where the entire body plan differs from one phylum to another, yet a
month or so ago we were discussing several phyla which have been merged
or split recently. In this case, embryological studies are the key for
making correct decisions, and in some phyla it's difficult to grow the
critters in the lab to study them, so I suppose that's why it's only
very recently that the studies have finally been done to repair the
mistakes. But that's a different issue from the problem with birds. All
birds, all tetrapods, all vertebrates, even all chordates, have similar
embryological development (Haeckel's exaggeration of this similarity
notwithstanding). So such studies aren't needed to classify various birds.
..

.

Relevant Pages

  • Re: "Evolutionary Rate" from Natural Selection
    ... retrospective analyses on genera named before ... I see a likely reason for that: The differences between families mostly ... Yes, the more species in a group, the more ... likely there's a mistake one place or another. ...
    (talk.origins)
  • Re: "Evolutionary Rate" from Natural Selection
    ... > number of species. ... The number of families has no scientific basis. ... two genera in place of one genus that was there before. ... risen to family-level splits. ...
    (talk.origins)
  • Re: This group
    ... start by learning about families and genera. ... I know that Chrysanthemum got chopped into pieces sometime back, with the florists chyrsanthemums going to Dendranthema, and Chrysanthemum being restricted to a few Mediterranean annuals, but I was a bit surprised recently to see crown daisies given as Ismelia carinata - what's left in Chrysanthemum other than corn marigold? ... of those have impacted most gardeners. ...
    (uk.rec.gardening)
  • Re: This group
    ... start by learning about families and genera. ... As for cladists of different religious hues and their interminable wars, ... of those have impacted most gardeners. ...
    (uk.rec.gardening)
  • Re: Naming [Was:Re: Reproductive Selection]
    ... >>same genus or to different genera. ... > limits on how dissimilar two species had to be before being assigned ... You can look at a clade, and say, yes this is a clade. ...
    (talk.origins)