Re: Extinction (was: Religion)

[John Alroy <jack@homebrew.uchicago.edu>]

Sheehan et al.'s three-way partition did seem too coarse to me and I  
was bothered by this at the time (the data were presented at a  
meeting I attended before the publication came out and I remember  
being skeptical when I saw the talk). I would like to see someone  
reanalyze their data, maybe this will happen.

I will repeat one last time that 1) there are no published DATA  
supporting the "2-3 m gap" theory; 2) the number of relevant sections  
is rather small, regardless of the amount of Hell Creek outcrop in  
the general area; and 3) locating the boundary in those sections is  
difficult. Furthermore, I haven't seen any discussion of the  
"taphonomic control taxon" method for confirming the "gap." This  
would work by finding a highly fossiliferous locality (or localities)  
in the "gap" that is NOT a channel locality (see Bug Creek discussion  
below if you are wondering why channels are just too complicated to  
be relevant to all this) and DOES include large terrestrial  
vertebrates.

The Cambrian extinction probably was above background (i.e., a real  
mass extinction) but I believe it wasn't as bad as the end Triassic.

I seem to remember that the late Permian glaciations were way, way  
off from the boundary, not just a little bit, but this is moot  
because I think we all agree that the Permian event is still a  
mystery.

My claim about an abrupt climate change immediately after the K-T is  
based on analyses of leaf floras, which show dramatic physiognomic  
changes right after the boundary indicating a much cooler and drier  
climate (this is based on work by Wolfe and Johnson).

According to Swisher et al. (1993: Can J Earth Sci 30,1981), the K-T  
boundary is exactly coincident with the IrZ-Coal, which exactly marks  
the Hell Creek-Tullock boundary and includes the iridium spike. I am  
not sure why, then, Friesen believes that this traditional definition  
has been changed. Perhaps he is going from the Berkeley school  
argument that the Puercan really starts in the Cretaceous, but I  
think even those guys don't believe this any more in the wake of  
Lofgren's thesis work on the problem. Furthermore, this would put the  
boundary either at the boundary or in the Tullock, not in the Hell  
Creek. It's much more likely that Friesen is going from the argument  
that the assorted Bug Creek channels are Paleocene but, technically  
speaking, are in the Hell Creek. However, these are channels that cut  
down from the top of the formation and clearly post-date it; given  
this the age of the channels doesn't bear on the fact that the  
formational contact is at the boundary where it is not destroyed by  
channels.

I won't comment on the volcanism model until I read some of the  
pro-volcanism literature.

>From what Friesen says the Hansen and Upshaw mollusc paper is  
equivocal. Personally, I think two extinctions in 50 cm of section  
point to a catastrophic event no matter how you cut it. For example,  
this could result from two "events" that were both caused by the  
aftermath of a single impact - perhaps a physical disturbance event  
(tsunami?) followed by a general ecological collapse (elimination of  
the plankton due to chemical changes in the seawater? or whatever).  
But this is entirely besides the point because I'm not at all  
impressed by the evidence for an initial extinction phase (as  
reported by Friesen): 15 out of 55 species having their ranges  
truncated 50 cm below the boundary sounds to me like a local  
paleoecological or taphonomic effect, probably in combination with  
the Signor-Lipps effect (rare species having truncated ranges that  
"smear" an extinction downwards). Hansen is in the Keller camp and I  
would treat his interpretations very cautiously.

My latest update of the Bug Creek Anthills list includes 27 species,  
of which the following five are Puercan:

 Stygimys kuszmauli
 Gypsonictops petersoni
 Oxyprimus erikseni
 Mimatuta morgoth
 Procerberus formicarum

The following three species are definitely known in the Cretaceous  
and are known in other Paleocene localities:

 Mesodma sp. (species level taxonomy is unclear, this could be an
   exclusively Cretaceous species)
 Cimexomys minor
 Protungulatum donnae
 

Here I am assuming that the Long Fall Horizon and Harbicht Hill are  
in the Puercan. So it is fair to say that Bug Creek is dominated by  
Lancian species, and it's conceivable that a few of these species are  
misidentified or actually did originate in the Cretaceous, or that  
Fox is right that Long Fall is Cretaceous, in which case several of  
the above species are as well. So it seems to me that there are two  
fair interpretations:

1) Bug Creek really is Cretaceous;
2) the channel itself is earliest Paleocene, but the fauna is largely  
dominated by reworked Cretaceous material, plus a few unreworked  
Paleocene species and possibly a couple of survivors.

I favor interpretation 2 because it seems reasonable that a huge  
diversification wouldn't start for, say, at least 10,000 - 50,000  
years after the boundary. Five new species vs. three survivors in  
that amount of time is pretty impressive. The mammals DO experience a  
gigantic diversification in the early Paleocene, presumably because  
of "ecological release" (the evidence for this is overwhelming, we  
can talk about it if you want). However, a logistic growth model  
predicts a "lag phase" before the diversification kicks in.
In any event, I'm not sure it matters which scenario is correct  
because the difference between BCA and Ken's Saddle (the only  
unambiguously Cretaceous channel in this data set) in terms of the  
concentration of dinosaur material is only of a factor of two. This  
doesn't strike me as all that impressive given the huge amount of  
variation among localities in the density of fossils everywhere in  
the record. More importantly, the stratigraphic distance between  
these two localities (as shown in Williams' redrafted figure) is on  
the order of centimeters, not enough even to show which channel is  
younger!! I do agree with Friesen that we shouldn't treat the  
dinosaur abundance data as unequivocally showing a gradual decrease.  
Even though they are compatible with the "moving window" taphonomic  
model, there are only five samples with dinosaurs to start with.