[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Nest predation

>I had thought that parental care was more reliable because of extant
>phylogenetic bracketing (since crocs and birds do it nonavians did it
>too). Because you couldn't bracket with nest parasitism it would seem a
>less useful trait.

Again, it depends entirely on the scope at which you're working.  At the
level of Archosauria, I agree - parental care is more unambiguously
optimized than is nest parasitism.

But this is a different issue from utility in phylogeny reconstruction.
One can never, ever, use a feature reconstructed via EPB as primary
evidence in phylogeny reconstruction.  EPB is done post hoc.  And at the
level of Archosauria, parental care is more or less universal among living
members (with some deeply-nested exceptions), and nest parasitism is a rare
feature that might, in fact, preserve a phylogenetic signal if considered
in a total evidence environment - as such, parasitism would be "more
informative" in that parental care, by itself, tells us nothing about
within-group relationships.

>> Second - which is likelier to result in phylogenetic noise during a
>> cladistic analysis, one that is lost and gained many times or one that is
>> gained a handful of times?
>Don't you mean "...one that is lost and gained many times or one that is
>*lost* and gained a handful of times."

Not necessarily.  Based on what I know of avian phylogeny, nest parasitism
was not lost - it appears to have been gained independently multiple times,
though a loss is possible depending on how one resolves the cuckoos.

Still, your point is well-taken.  The rate of evolution of a character
(i.e. number of transformations, in either direction) will reflect its
informativeness at a given phylogenetic level.  (How one determines this a
priori for morphological characters is a huge problem, though.)

>Yes, the first makes more noise but if you make the assumption that the
>second is therefore more reliable, you are more likely to listen more
>carefully to it, thus effectively amplifying it!

Not in a strict, unweighted parsimony analysis, by definition.  Such
analyses - which are what most vertebrate paleontologists use - make no
such assumptions a priori.  There are approaches with molecular data that
do (maximum likelihood), but there are often sound thermodynamic reasons
for making them, and with modern programs one can modify these assumptions
and see how the resulting tree differs depending on how they are modified.

>> How do you measure evolutionary malleability, in the absence of a phylogeny?
>Right.  Except that they are all under construction.

Or, like anything else in science, tentative as long as disparate signals
are present.

>> >I think my difficulty with the particular trait under discussion is that
>> >it seems rather obviously ephemeral...and being ephemeral (if it
>> >is) we can move to adaptationist hypotheses such as: given certain
>> >preexisting conditions trait X is likely to evolve at some frequency.
>> ***only within the crown group.***
>No.  I mean universally.  This could be true for Hymenoptera or
>Archosauria or any group that has developed the following
>preconditions: host species has large investment in parental care; host
>cannot discriminate parasite; parasite has higher reproductive success
>parasitizing than immediate ancestor had not parasitizing; species are
>both oviparous.

But you are basing your model on living neognath birds, which is why it is
inapplicable to anything outside that group.  This is the basis for EPB
(or, more generally, DELTRANS optimization) - in the absence of hard
information, we restrict the distribution of character states to their most
recent plausible common ancestor.

We know nest parasitism has arisen multiple times within Neognathae.  We
are unaware of any such instances in ratites or tinamous, and crocs don't
do it.  Therefore, we cannot extrapolate the model to anything in between
(i.e. nonavian dinosaurs), as the evolution of this trait appears
restricted to one subset of the entire group.

>But then systematics only deal with the states of existence and not how
>they got to be that way.  But while cladistics can inform hypotheses of
>origination (for example, my hypothesis of parental investment being
>selected to increase embryonic rate depended upon phylogenetic
>relationships between archosaria, birds, crocs, and dinos), I suspect the
>reverse is not true.  Adaptationist hypotheses cannot and should not be
>used to construct phylogenies.

Of course not - but they must be based on a sound phylogenetic hypothesis
first.  I recommend Harvey and Pagel's *The Comparative Method in
Evolutionary Biology* (Oxford Univ. Press) for a more cogent discussion on
this matter.

>Chris, I suppose it's obvious that I am fairly ungrounded in this
>discussion.  I appreciate your teacherly attitude.  I have purchased
>_Encyclopedia of the Dinosaurs_ and am dragging my way through a
>discussion of systematics.

I recommend a basic primer on phylogeny reconstruction and parsimony
analysis - the Wiley et al. Compleat Cladist (KU press) is a good place to
start.  The chapter on phylogenetics in the Encyc. of Dinosaurs is OK, but
I find CC to be more clear.  (It's also much cheaper!)


Christopher Brochu

Postdoctoral Research Scientist
Department of Geology
Field Museum of Natural History
Lake Shore Drive at Roosevelt Road
Chicago, IL  60605  USA

phone:  312-922-9410, ext. 469
fax:  312-922-9566