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

Re: Extinction (was: Religion)

 > Here we go again. I hope everybody is enjoying this (but I doubt  
 > it!). This is in response to Stan Friesen's latest missive.
 > Stepped extinctions: I thought we had gone over the dinosaur evidence  
 > enough by now, but I guess not. I've referenced the Sheehan et al.  
 > (1991) paper twice and heard no actual discussion of the data, just  
 > repetition of the claim that dinosaurs started to go extinct before  
 > the K-T, as if this was some kind of widely-known fact (even Ralph  
 > Chapman said something along these lines!!).

One - unless there is something in Sheehan's paper that is not
addressed in Williams' paper, it is outdated.

Two - I am not talking about individual taxa being counted, I am
talking about total bone counts of *all* species of dinosaur
at once.

Three - I am not saying this pattern is (yet) generally demonstrated,
it is demonstrated (so far as I know) only in the Hell Creek Formation.
It is possible to do so there because of the intensive  searching that
has been done in that formation.

 >  My own experience with  
 > the Cenozoic mammal record, which is MUCH better understood than the  
 > dinosaur record and involves many more valid taxa, is that  
 > paleontologists routinely assume that little squiggles in their  
 > diversity curves correspond to "step-wise extinction events" when  
 > these squiggles are due to artifacts created by sampling or by the  
 > arbitrary, coarse time scales that are used or by the failure to test  
 > null hypotheses explicitly.

All this applies to the usual methods - but *not* to the "barren
zone" datum for the Hell Creek Formation, which persists despite
intensive attempts to eliminate it by detailed, foot by foot,

The datum is *not* individual last occurances - it is not even
last occurances at all.  It is purely a substantial reduction
in *frequency* of bones per volume of sediment, irrespective
of taxon.

 > As for brachiopods, I have no idea what data are being discussed  
 > here, but since brachiopods took a huge hit at the Permo-Triassic  
 > boundary and never recovered (see the diversity data in "Phanerozoic  
 > Diversity Patterns" [1985: J. Valentine, ed.], for example), I  
 > suspect Stan is confusing something. Certainly molluscs (clams,  
 > snails, cephalopods) show a huge and decidedly non-stepwise  
 > extinction at the boundary.

I will have to cross check this.

The study I have in mind was a study of a Brazos River section
that also supposedly showed a tsunami layer.  In the article
the tsunami layer was taken as marking the K-T boundary.

For whatever shellfish class this was about, there was a decided
stepwise loss over some vertical distance.  The article discussed
the issue of the discordance between last observed occurances and
actual extinction date.  As I remember it, the conclusion was that
the "pre-boundary" losses exceeded what was expected for sampling

I later saw mention of a reference (which I did not read) which
purported to show that the supposed tsunami layer was *earlier*
than the K-T boundary and that it was not even a tsunami layer.

Now, supposing the revised dating is actually correct, this
actually *reinforces* the spread of extinctions, since the drop
offs now occur even *further* before the boundary.

[I have the reference at home, and I will try to remember to bring
it tomorrow or the day after].

 > "Harmless" impact sites: I know of no Phanerozoic impact crater as  
 > large as the Chicxulub crater. Can anyone cite a counter-example?  
 > Based on Raup (1992: Paleobiology 18,80), the closest runner up is  
 > the Manicouagan crater, the age of which is uncertain and could well  
 > fall at the end-Triassic mass extinction (one of the "big five").

No, the end-Triassic is NOT one of the Big 5.

The other three are all Paleozoic.
 > Alternate mechanisms: Ice ages? This must be an error: the late  
 > Cretaceous is a notorious "greenhouse" interval.

I wasn't talking about the K-T extinctions here - I was talking more
generally.  In particular, I was refering to the P-Tr extinctions.

[Though it is interesting to note that there was a substantial
drop in temperature near the K-T boundary].

For the K-T extinctions the contributing factors were probably
the lowering sea levels (thus increasing seasonality), and the impact.

 > What is  
 > the evidence for ocean anoxia?

Again, this refers to *other* extinction events.

Almost all of the lesser extinctions between the P-Tr and the K-T
are associated with ocean anoxia (actually an elevated oxygen
compensation depth).

 > but this would be no surprise the essence of a  
 > "multi-causal" model is to find a separate argument for every piece  
 > of evidence.

Not so.  At least not for me.  I look for factors that reinforce
one another.

Most of the extinctions involving ocean anoxia were primarily
marine anyway.

 > I agree that in principle alternatives to the impact model for the  
 > mass extinction could be developed. In practice, I'm not aware of a  
 > well-articulated, published development of such arguments.

Hmm, I recommend "Extinction Events in Earth History"
edited by E.G. Kauffman and O.H. Walker. especially the
article by Leary and Rampino.

[Actually I think the Brazos tsunamite article is in there as
well, but I am not sure].

 > The  
 > Keller/Officer/etc. camp are trying to show that there was no mass  
 > extinction, not that some alternate model can explain the event. The  
 > anti-extinction camp is just plain wrong

So do I.

Are they *really* still at it?
That seems silly to me.
 > Tsunami deposits: ...
 >  "tsunami" deposit in northeast Mexico,

Ah, this sounds like it may be in the same sedimentary sequence as
the Brazos deposits.  It may actually "rescue" the Brazos
deposits as a tsunamite.  Good, I always did feel uncomfortable
with the alternatives there.

Now, did they verify the correspondance with the K-T boundary?
[The other charge against the Brazos datum was that it predated
the K-T boundary].

[Note, I actually would *prefer* the Brazos data be as originally
presented - but I try not to let my preferences unduly influence
me in such issues].
 > The contradiction between the volcanic theory and the "extreme"  
 > gradual theory is that the former is trying to explain a mass  
 > extinction and the latter denies that one ever took place.

Ah, well, I have never been talking about that version,
and neither has anybody else on this forum.

As far as I know there are *very* few who take that position
any more.  Certainly Williams' data is wholly inconsistant
with that.

Almost all of the references I have seen practically take
the existance of the mass extinction for granted.  Even the
most extreme position I have seen recently would only extend
the mass extinction over the span of the Late Maastrichtian,
a period of around 4 million years.

And the article (?Sheehan's) which broke the Hell Creek up into
swatches and compared the abundance of dinosaurs in each third
(after doing a detailed grad-student sweep of the formation)
pretty much demolished any possibility of the extinctions operating
on *that* scale.

The issue remaining is the 100k year scale versus the 10k year scale
required by the pure impact hypothesis.

I consider that anybody taking any other position *is* ignoring
the data, big time.

 > There is,  
 > of course, an important distinction between the gradualist and  
 > "semi-gradualist" positions, as Friesen indicates; the latter believe  
 > that the extinctions did occur but took about 10^5 - 10^6 years, or  
 > more. My own view is that the evidence is clearly in favor of much  
 > shorter time scales, 10^4 years or more, but we should argue about  
 > this by looking to particular data sets.

Actually, the evidence I have seen is that different group
died out on different time scales.  Some may well have died
out on 10^4 scales.

 > My point about positive evidence and arguments is that the  
 > anti-impact/catastrophe camp has more and more focused its efforts on  
 > attacking particular evidence for the impact or for a rapid  
 > extinction. Of course there could be some who argue in a positive way  
 > for an alternative, volcanism-driven model - again, where are the  
 > references?

Try looking at almost anything by Rampino.
 > I'm amazed that the "2-3 m" issue won't go away. Why are you so  
 > willing to ignore sampling effects and variation in data?

Because we are talking about a case where a line of graduate
students spaced a few feet apart walked across the formation
parallel to the upper boundary noting every little piece of
bone they found.

This is most *definately* a sufficient sample to detect trends
in bone abundance.

The last two or three meters is a *distant* outlier from the other
part of the formation.  This difference *far* exceeds mere sampling
error.  We are talking here about an order of magnitude (or
maybe two orders) difference in abundance.

 > Furthermore, what is the EVIDENCE for this "gap"? It consists  
 > entirely of a claim made by proponents of gradual mass extinction  
 > (e.g., Archibald, Clemens, Williams) that such a gap exists. Even if  
 > there WAS real, statistically meaningful evidence for this "gap," it  
 > would show nothing apart from the existence of variation in  
 > taphonomic regimes, which no serious paleontologist would question.

Huh?  Given the spatial heterogeneity of the Hell Creek, I
do not see how this pattern could persist over the whole
formation (several hundred square miles of exposed rock).

If it were merely taphonomic it would be a local affair.

 > I'm sorry, but I'm getting frustrated with comments like "Williams  
 > results are... pretty convincing... The existence of a *large* drop  
 > in abundance of dinosaur bones (and teeth) in the last few meters of  
 > the Hell Creek Formation is now a thoroughly established fact." THERE  
 > ARE NO DATA IN WILLIAMS, much less a statistical analysis! Sheehan et  
 > al.'s data showed NO CHANGE IN ABUNDANCE and involved a proper  
 > statistical analysis.

At an *insufficiently* *detailed* level.  The barren zone is
too small too have shown up in the tripartate division of the
Hell Creek that he used.

I noticed that the first time I heard about it in detail.
A drop in the last tenth or twentieth of the formation is
*not* going to show up in a comparison of thirds.

Sheehan simply does *not* *address* the issue on this scale.
I consider such coarse graining to be wholly inconclusive.

 >  Take another look at the paper. What would it  
 > now take to "prove" a rapid extinction of dinosaurs?

1. An analysis as detailed as Sheehan's at the 3 meter resolution
level that showed the last partition did *not* have a reduced


2. A showing that the barren zone is found *only* in the Hell Creek
and is thus a Western Norht America regional phenomenon.

Either one would be fairly strong.

 >  Doesn't anyone realize that  
 > arguing over "2-3 m" is the paleontological equivalent of debating  
 > the last 2-3 seconds of an NBA playoff?

Perhaps, but it since it distinguishes between the possibility
of a single catastrophic event causing the extinctions and the
impossibility of that, it is a necessary argument.

It is equivalent to arguing whether the ball was in the air when
the gun whent off.  That *is* significant.

 >  The supposed 2-3 gap  
 > would translate, therefore, to no more than 43,000 years, but this is  
 > a very high overestimate because 1) the Two Medicine, despite being  
 > nearby the Hell Creek and not much older (Campanian), was deposited  
 > in a proximal foreland basin - sedimentation rates in the Hell Creek  
 > were probably half of this (Rogers pers. comm.); and 2) this figure  
 > completely ignores variation in sedimentation rate and is an average  
 > based on 372 m (!).

Hmm, could be.  As I remember it, the Hell Creek is about the
same depth, and covers the Late Maastrichtian - about 4 million
years.  Proportionately, this makes the gap about 30,000 to
40,000 years.

Still, if the impact were the cause, the drop in abundance
should happen *at* the boundary clay, not below it.

Interestingly, the rate cited implies a *very* short duration
for the ecological effects of the impact, as the fern spore
spike is measured in centimeters (perhaps only one cm).  This
implies that some sort of relatively mature vegetation was
re-established in about 500 years!  Perhaps less.

 > I am not going to argue  
 > particulars about the evidence for Chicxulub being an impact crater  
 > because I feel that this has been dealt with quite adequately by bona  
 > fide geologists (I'm a biologist).

Nope, I'm afraid it *hasn't* been.

The article was only published a few months ago, and has not
been responded to *at *all*, by anybody.

I'm afraid we are going to have to wait to see where this one falls.
[Actually, I am still leaning towards it being the impact site,
as the supporting data do seem strong - but I am reserving judgement].

 > Causal theories: "all known instances of a particular effect have
 > been shown to be correlated with the postulated cause." This is not  
 > terribly generous - once again, I feel like I am being asked for a  
 > videotape of the K-T. I suppose I could ask for the same criterion to  
 > be applied to the volcanism story, but why bother?

No, I am asking you to compare the record for the K-T extinction
with that for other extinctions.

That is how one establishes cause and effect.

*One* correspondence is almost never sufficient by itself.
[Unless you can show that extinctions remained at the background
level right up until the moment of the impact, you do not even have
a fully verified correspondence].

 > Non-selectivity: Check the paper by Raup and Jablonski (1993: Science  
 > 260,971), which I believe I already have cited. There is no  
 > geographic pattern to the extinction of molluscs, NOR is there any  
 > relationship between various ecological characteristics (body size,  
 > infaunal/epifaunal, etc.) and extinction probabilities. This DOES  
 > match with a catastrophic event because only a catastrophic event  
 > could so thoroughly change "the rules of the game"; a drawn-out  
 > volcanic episode doesn't.

I think it does.
 > I don't have a reference for the Texas "tsunami" beds. Can somebody  
 > else supply us with one?
 > Lancian mammals: The "lowest" channels certainly do have Lancian  
 > mammals in them, namely, reworked Lancian mammals. This, like the  
 > dinosaurs, has been gone over endlessly by the VP'ers and all I can  
 > say at this point is that I have never in my life seen such an  
 > unambiguous case of a single pattern (mass extinction) being obscured  
 > in a very minor way (a few meters) by sampling effects (reworking and  
 > Signor-Lipps).

Lancian mammals reworked and essentially no unreworked Puercan ones??
That seems odd to me.    The *middle* channels have a mixture
of Lancian and Puercan mammals, I believe the lower ones are almost
entirely Lancian.  (I will have to double check this as well).

swf@elsegundoca.ncr.com         sarima@netcom.com

The peace of God be with you.