Re: Extinction (was: Religion)

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

I really should just shut up, but there are some things I can't  
resist commenting on in Stan Friesen's latest post.

Just because Williams paper is more recent than Sheehan et al.'s  
doesn't mean it's "outdated" and therefore irrelevant. Quite the  
contrary: Sheehan's paper at least makes an attempt to deal with the  
problem statistically, whereas Williams' includes no original data  
whatsoever, much less an analysis (whatever else you might say about  
it).

I honestly believe that the now-infamous "2-3 m gap" is irrelevant to  
this debate. I already went over the argument that this interval  
represents no more than 40,000 years, unless there is a major hiatus  
within it that nobody has detected (as far as I know), and Friesen  
agrees with this figure. If we are talking about dinosaur extinctions  
in the context of the Maastrichtian world, we should remember that  
this interval includes less than 0.4% of the record for the  
Maastrichtian. Dinosaurs do NOT have a good record, and having EITHER  
a sampling- or taphonomy-produced gap of this length is absolutely  
trivial. Just look at the Cenozoic mammal record in North America,  
which is far, far better understood than the dinosaur record of any  
continent in any time interval: there are about 3000 published  
localities of Paleocene to middle Pleistocene age (my own data,  
unpublished), which is an interval of about 64.5 m.y. This works out  
to an average of one locality per 21,000 years with PERFECTLY uniform  
sampling. Of course, the sampling is far from uniform; the large  
majority of these localities include one or two mammalian taxa out of  
the several hundred that would have existed at any one time; and  
mammal localities are a lot more common than dinosaur localities on a  
formation-by-formation basis. Do you see anyone arguing about "gaps"  
in the mammal record?

As far as I know the age of the Mexican tsunami deposit has not been  
called into question (the Brazos River section is another issue).

The end Triassic IS one of the Big 5, the others being the end  
Ordovician, late Devonian, Permo-Triassic, and K-T. The original  
reference on this is Raup and Sepkoski (1982: Science 215,1501). I  
suspect this is just a lapse on Friesen's part, unless, of course,  
there is a problem with R & S' analysis (entirely possible but it  
sure looked good to me).

I think Friesen and I agree that the "world gone to hell" model may  
work for the Permo-Triassic, although I don't think it's quite  
necessary for the K-T (not that the late Maastrichtian was the nicest  
time in Earth history). However, the timing of the glaciations is way  
off from the boundary (see Erwin's recent book). I agree that sea  
level change may have made things tough at the K-T, but it's not  
clear to me how cause and effect worked here - there was an abrupt  
and major climate change at the K-T (not BEFORE the boundary, at or  
after the boundary), and that could have had an affect on sea level.  
Maybe we can get some geologists/paleoclimatologists/
paleoceanographers to comment on this.

The article cutting the Hell Creek into three units and showing there  
were no differences in the dinosaur faunas in those three units was  
indeed the same as the Sheehan et al. article I have been talking  
about.

The Hell Creek doesn't include a complete K-T boundary section in  
EVERY exposure. If you look at Archibald's strat sections (1982:  
UCPGS 122) you will see that the situation is extremely complex, with  
an awful lot of lateral variation. Many of the "boundary" sections  
are badly messed up by channel complexes that cut down into or  
completely obliterate the boundary. Furthermore, the main evidence  
for the location of the boundary in the first place is the location  
of the "lowest" coal, wherever that may be. Archibald was a little  
more thoughtful about it than that, but I suspect there are still  
major problems trying to show the top 2 meters are exposed in a lot  
of places. This is a very short span of section! I think somebody  
should start looking at hiatuses in the dinosaur record LOWER in the  
Hell Creek. I would be very surprised if there were fewer than a  
dozen gaps of at least 2 m in the section, given that it must be a  
few hundred meters in total. The point here is that EVEN if  
undersampling (the basic point of the Signor-Lipps effect argument)  
doesn't account for the 2-3 m "gap," taphonomic vagaries certainly  
could. Similar (and actually much, much longer) gaps are found  
everywhere in the terrestrial vertebrate record. Since Friesen has  
already made a good point about this being a phenomenon entirely  
local to the Hell Creek, I think we should call a truce on it.

Too bad there aren't any non-avian dinosaurs around we can ask, but I  
suspect they weren't too good at going 500 or even 5 years without  
substantial standing vegetation to feed on (apart from ferns and  
scrubby angiosperms). Smaller animals like multis presumably would  
have less trouble with this particular problem.

The record for extinctions other than the K-T is far, far worse in  
all respects (excluding the end Pleistocene, which is restricted  
almost entirely to mammals and birds). I am including the other four  
of the Big 5 here; the lack of data extends to biostratigraphy,  
absolute dates, mechanisms like flood basalts, glaciations, impact  
craters, you name it. If the K-T impact story is unconvincing, I  
guess we should just give up completely trying to explain large-scale  
patterns in the fossil record (of course, that would put me out of a  
job).

A really huge volcanic eruption could change atmospheric chemisty and  
therefore climate in a big way, but to be honest this seems to me  
like a much less efficient way to cause massive extinctions in the  
oceans than slamming a gigantic asteroid into a carbonate shelf. Of  
course, this depends on what you are putting in the atmosphere, how  
much of it, how fast, and how it effects ocean chemistry, and  
possibly these things could be worked out. I will try to get ahold of  
the Rampino refs - are any of the in Geology? That would make things  
easier.

Checking the Williams article again - he repeats the critical data on  
dinsaur tooth abundances - I see that the drop in dinosaur abundance  
starts in the Bug Creek Anthills fauna. This is the same fauna where  
we first see (to my mind) completely unambiguous Paleocene mammals  
mixed up with presumably reworked Cretaceous mammals. Note the  
monotonic decrease in the density of dinosaur remains going up into  
higher and higher channels, exactly as predicted by a taphonomic  
"moving window."

Whew, maybe I should get some work done today...