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

Re: No Cretaceous placental mammals?



----- Original Message -----
From: "evelyn sobielski" <koreke77@yahoo.de>
Sent: Friday, June 22, 2007 3:42 PM

I'd like to see - or if, some day, I have the time and
resources and reputation and whatnot, and nobody did
it til then, I'll do it myself - all the isolated
humeri, tarsometatarsi, ... scored and evaluated using
only that single bone from a wide range of taxa living
and prehistoric, and with a matrix that is so
fine-grained that the foot-propelled divers *won't*
cluster. Milk a single element for all the
phylogenetic signal that it holds. Might be that this
is a reasonable way to crack the afiliations of the
assorted crumbs we have laying around from the New
Jersey marls etc; I don't think it has been tried yet
(not for these bird fossils at least).

Don't wait till you have a reputation! _Get it_ _by_ doing such an analysis! I suppose it's enough to just plug them into Livezey & Zusi's matrix (...especially if you also add a few other fossils to this fossil-poor dataset).


Don't make a single-bone matrix. As soon as you put more than two taxa known from more than that bone into your matrix, that would mean to throw away information -- information that would help build the backbone of the tree into which the single-bone taxa can then insert themselves.

----- Original Message -----
From: "evelyn sobielski" <koreke77@yahoo.de>
Sent: Friday, June 22, 2007 3:58 PM

"For example, the &#8216;traditional&#8217; avian
mitochondrial substitution rate of 2% per million
years is frequently adopted without acknowledgement of
the associated uncertainty" should be good for one or
two cases in point (the rate does not hold true e.g.
for most if not all Procellariiformes, there's
something seriously wrong with mol-clocking ratites
but nobody really knows what it is, etc).

There is something seriously wrong with trying to calculate molecular divergence date estimates without at the very least two calibration points, yes. The assumption that a strict molecular clock holds over any range of taxa and time has been empirically disproven so often that I wonder what literature those ornithologists who still believe in "standard rates" read! Outside of ornithology, AFAIK, nobody has been using "standard rates" for _decades_; I was quite shocked last year to learn that some ornithologists still did.