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

Extinction (was: Religion)



From: John Alroy <jack@homebrew.uchicago.edu>
 > 
 > 1) Empirical claim: the evidence really IS strong, at least with  
 > respect to the theory that a mass extinction did occur. The type of  
 > evidence I am used to looking at is statistical evidence for or  
 > against changes in extinction rates. ...

I know of nobody who is seriously denying this any more.
In fact the K-T extinction is one of the Big 5 (not the
biggest though).

 >  I really  
 > feel that this is unambiguous and calls for some kind of a mechanism,

Well, it certainly would be nice to have one, but the evidence may
not be there to get one, especially if extinctions are complex
events with complex causes.
 
 > and in this case the impact theory fits the geological data quite  
 > nicely.

THIS is where we start to disagree.

Some groups, for instance dinosaurs, brachiopods and so on, show a
stepped decline starting *before* the boundary layer.

There are known major impact sites that do NOT correspond to
known extinctions.

Between these two, I think that substantial doubt can be cast
on any *simple* model of an impact as *the* cause of the extinctions.
[An impact as a *contributory* cause is a different matter].

Now, in this case there are alternatives to a pure impact model
that are reasonable.  These are multi-causal models in which
flood basalt volcanism plays a major role, and other factors,
such as impacts, oceanic anoxia, ice ages and so on provide
additional contributory input that determines the *magnitude*
of the extinction.
 
 > This may be a flawed outline, but I think you see my point: I feel  
 > that if you just look at the KIND of arguments going back and forth  
 > about the K-T, they fall into phase d and not into phase c.

Nope, alternative models are still being developed and evaluated.
If this were phase d, that would not be so.

Also, your analysis leaves out the sequence of phases in
dealing with theories that are initially attractive but fail
to work in the long run.  The early phases will be *very*
similar, or even identical.  I think that it is too early to
tell which way this is going.

There is also a third pathway, where there is insufficient
basis for distinguishing the competing theories.  In that
case opinion swings back and forth in what amounts to fads.
  
 > Also, you could make a case that the anti-impacters really are  
 > advocating a volcanic theory for the extinction, e.g., with regard to  
 > the shocked quartz, iridium (and other trace element) anomalies,  
 > tektikes, Chicxulub crater, "tsunami" deposits, and extinction  
 > itself.

Nope.  I would say that the evidence for an impact at the K-T
boundary is fairly solid, though not utterly conclusive.  However,
the evidence for a major episode of volcanism surrounding the
boundary is fully conclusive.

Add to this the lowering of the shelf seas, the cooling climate,
and so on, and you get alot of factors all conspiring to produce
environmental stress over a long period of time.

By the way, the Chicxulub "crater" may not be one, or at least
not a meteoric one, and the last I heard most of the supposed
tsunami sites were questionable.

Has there been new material on this in the last 3-4 months?

If so, could you post the citations so I can read it?

 >  But then you have to recognize that there are TWO anti-impact  
 > theories at work here - volcanic extinction, and gradual extinction;  
 > that these theories are mutually incompatible;

Huh?  The volcanic theory is that the extinctions took place
over a period of about a million years or so.  Or by gradual
do you mean the 5-15 million year range?  If so, the evidence
for the extinctions being *that* gradual is weak - in fact the
evidence is contrary to that scale of change.

So, where is the conflict?

Or do you have the idea that the "volcanic theory" is about
a single major volcanic explosion replacing the impact??

Besides, why do the extinctions have to have a *single* cause?
Even *if* volcanism in India was the primary cause, this does
not mean that other contributing factors may not have been
operating  over a longer time period.

 > and that the SAME  
 > people are advocating both of these theories at the same time. The  
 > important point is that claims about this or that being "gradual" or  
 > "volcanic" seem to me like point-by-point attacks on the evidence  
 > that are not really intended to support an alternative theory.

Oh, let's see, evidence that all of the significant extinctions
events in the Mesozoic are correlated with flood basalts is
not *positive** evidence??

Evidence from careful surveys that dinosaur bones are *substantially*
rarer in the last 2-3 meters below the K-T boundary is "merely" a
refutation of abrupt extinction??  (In fact dinosaurs are so
rare in this upper zone it used to be though there were *no*
dinosaur bones in it).
 > 
 > A couple of specific points in response to earlier posts.
 > 
 > Stan Friesen: "the abundance
 > of dinosaur remains drops sharply a few meters below the siderophile
 > layer (aka boundary clay)."
 >      Again, please see Sheehan et al. (1991: Science 254,835) for  
 > the evidence against a gradual extinction pattern for dinosaurs. In  
 > situ dinosaurs are found "within 60 cm of the boundary" EVEN  
 > according to Williams (1994: J Paleont 68,183), who argues against  
 > the catastrophic model.

Please, the existance of bones within inches of the boundary is
NOT the issue.  It is the fact that there *so* *few* of them.

Williams results are, to my mind, pretty convincing.  He even uses
*Sheehan's* data in reaching his conclusions.  The existance 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.  Citing an *older* study by
Sheehan is not going to change that.

The *only* issue now is whether this is a purely local phenomenon
of the Lance/Hell Creek, or is found elsewhere in the world.

If it is local, then it is not relevent to the global extinctions.
If it is found elsewhere, then it eliminates all abrupt extinction
models.

Note, 2-3 meters is approximately 100,000 to 250,000 years
in that part of the world.  This places the *entire* reduction
of abundance within the span of the Deccan volcanism in India.

 > This is less than an arm's length.

So?  How *many* bones are found at that distance?
If the extinctions were abrupt, there should be as many or *more*
bones in the last meter as in the lower levels.

This is *not* found, not by an order of magnitude.

 > I am  
 > getting a little frustrated that certain people (definitely not Ralph  
 > Chapman) don't seem to be aware of 1) the extreme rarity of in situ  
 > dinosaur remains, 2) the Signor-Lipps effect , 3) the fact that  
 > taphonomic regimes are wildly variable and not fully understood in  
 > terrestrial sediments, and 4) the fact that we are talking about 60  
 > cm, not 60 m or 600 m. Most "good" stratigraphic data for fossils are  
 > on the scale of meters, not centimeters.

I am certainly aware of these things.  They do *not* alter the
statistical distribution of dinosaur bones and teeth.

True, the last occurances of individual species in the fossil
record will be smeared out in the fossil record even for an
abrupt extinction.  But Williams takes that into account.
[For one thing, by lumping all dinosaurs together, you pretty
much reduce the Signor-Lipps effect to minor variation].

And, yes, in general stratigraphic data are usually only
on the scale of emters for large animals like dinosaurs, but
the *intensive* sampling of the last Hell Creek Formation
cited by Williams means that this is *not* true for this one
specific case.  For once we have adequate data on the centimeter
scale for *total* *bone* *counts*.
 > 
 > Holtz: "Evidence of impact DOES NOT EQUAL evidence of extinction...  
 > The evidence for impact causing the extinction is, of course, the  
 > weakest,
 > and entirely circumstantial."
 >      Any argument "A caused B" has to show that A happened; that B  
 > happened; that A happened before B and in the same place; and that A  
 > and B are associated with each other beyond the expectations of  
 > chance alone (and maybe other things!). Therefore, by definition the  
 > evidence for an impact and for catastrophic extinctions must be much  
 > stronger than the "evidence" for a causal connection. However, I  
 > would like to point out that all of my A and B conditions seem to be  
 > met in this case

Except the last - "beyond the expectations of chance alone".

That is far from demonstrated.  You need solid (not merely
suggestive) evidence for impacts in association with other
major extinctions to accomplish that.  That is not now present.

The other extinctions that have siderophile enhancements show
a terrestrial elemental profile, not an ET one.

Siderophile peaks with terrestrial profiles have been found
that are *not* associated with extinctions.

Thus, merely showing a siderophile peak is not sufficient
evidence of an impact.  You need at the very least either
shocked quartz or an impact site.

 > (if we can agree on the argon dates and the  
 > Chicxulub crater being a crater and the extinction being real).

We cannot agree on the Chcxulub site being a crater,
and even if it is, this does not constitute sufficient
evidence that the bolide that made it caused the extinctions.

One coincidental bolide impact during a major extinction
event already in progress is *quite* within the range of
possiblities, or even probabilities, given the frequency of
lesser bolide impacts, *and* the frequency of extinction
events.

 > If  
 > this is "weak" or "circumstantial," what causal theory isn't?

Ones in which all known instances of a particular effect have been
shown to be correlated with the postulated cause.
 > 
 > Stan Friesen (earlier post): "The pattern of disappearance of many  
 > groups, including that
 > of the dinosaurs... does not fit that model..."
 >      I'm always surprised that my paleontological colleagues put  
 > so much emphasis on "predictions" based on speculation about the  
 > physiology, ecology, and life history of extinct and ancient forms  
 > like multituberculates and dinosaurs, when 1) the extinction data  
 > themselves are heatedly debated (what "percent" of multis really went  
 > extinct? nobody seems to agree);

This is not what I was talking about.
I was talking about the "barren zone" in the Hell Creek Formation.

This should no longer be a matter of debate, unless somebody
in the last weeks has come up with new data contradicting
Williams' results.

 > 2) the "predictions" are extremely  
 > debatable;

Well, the prediction of *abrupt* extinctions (< 10,000 years)
*is* required by the impact theory.  I do nto see this as
debatable, or even debated.

No matter how severe the impact, the environmental effects would
not last more than a few millenia, and the residual ecological
adjustments, with associated additional extinctions, could not
last all that much longer after that.

Showing that the extinctions started more than 100,000 years
prior to the impact is certainly sufficient to rule out the
imapct as the *sole* cause.

 > 3) the statistical analyses of these "predictions" are  
 > inadequate;

What, pray tell, is wrong with Williams' analysis?

 > The important point is that a really unusual
 > mass extinction DID happen.

So did 4 other "unusual" mass extinctions in the last 600 million
years.  And before that we cannot tell how many there were, as we
cannot identifiy microorganisms well enough from fossils to tell
if an extinction occured.

And then there were about 5-8 *lesser* mass extinctions between the
P-Tr and the K-T extinctions.

Mass extinctions do not seem all that unusual to me.

 > Another important point is that the marine invertebrate workers have  
 > found an almost completely non-selective extinction pattern for  
 > molluscs. This analysis suffers from few or none of the above  
 > problems with the vertebrate analyses, and a climate-driven extincton  
 > model fails completely to account for these data.

True, but I suspect that the effects of mass volcanism would
be similar to those of an impact, but over a longer time period.
This *would* explain thsi "non-selective" extinction.

By the way, in what way was it non-selective?

Did they address vagrant/opportunistic versus "climax" species
selectivity or just environmental selectivity?  In the other
groups I am aware of, the "climax" species were selectively hit,
and the opportunistic species weathered the process comparatively
well.
 > 
 > "Coincidences *do* happen, and bolide impacts are actually fairly
 > common on a geological time scale."
 >      See above sociological argument. Note that in terms of size,  
 > the Chicxulub crater is at the far end of the spectrum for "normal"  
 > bolide impacts.

*If* it is an impact feature.
 > 
 > "Well, last I heard, the supposed tsunami beds were highly
 > questionable - they were apparently both earlier than the
 > K-T boundary, and probably not tsunami beds."
 >      One place to look is the recent account of the Snowbird  
 > conference in Science (I don't have the reference right here, it's a  
 > recent issue). Participants in the conference went together to  
 > examine a "tsunami" deposit and came away agreeing that a very  
 > high-energy mechanism was needed to create it. The minority of  
 > nay-sayers were previously committed to the anti-impact school and  
 > EVEN SO simply said they needed more evidence before they could make  
 > up their minds. There was no question about the dating of the  
 > deposit.

OK, I will try to check the recent issues of Science on this.
[*Which* "tsunami" beds - the West Indies or the Texas?]

Remember, I already agree an impact probably occured, I just see
evidence that contradicts the idea that it was the *primary*
cause of the extinctions.

Actually, I was quite surpirised to hear that the tsunami beds
were doubtful, as they seemed quite well analysed.  I will be
quite willing to acknowledge they are real.

By the way, has the *dating* been re-verified?

At least for the Texas "tsunami" beds, I have heard (but not
checked for myself) that they predate the K-T boundary.
 > 
 > "New evidence has come to light of intact latest Cretaceous
 > sediments *inside* the supposed impact structure."
 >      I posted on this before. The single published paper on this  
 > supposed evidence (Meyerhoff, in Geology) provides no documentation  
 > for the age of the "fossils" mentioned and is based on a log kept  
 > during the drilling of the core more than 25 years ago. The authors  
 > seem unaware of the possibility of reworking or misidentifications.

They may be.  But I am not.

That is why, in an earlier article I said these results need
to be independently confirmed.

Until they are carefully evaluated, they *are* sufficient
evidence to cast *doubt* on the identity of the Chicxulub
feature.

Note, I fully expect a major paper or two on this issue within the
year.  The people who think it is an impact crater are going to
be rushing out there to gather new evidence even as we speak.
When they have finished, and have published their results,
then we will have good reason to say one way or the other what
the Chicxulub feature really is.

Right now all we can say is: we don't know.
 
 > "the abundance of teeth drops drastically several meters
 > *below* the boundary clay - at about the same level as dinosaur
 > bones become seriously scarce, the so-called barren zone."
 >      This is based on material from WITHIN deep channel deposits  
 > in the Hell Creek. These channels appear to be earliest Paleocene  
 > (based on the mammal fauna; see D. Lofgren thesis), and the fall in  
 > abundance is perfectly in accord with the taphonomic "moving window"  
 > model of reworking.

Not all of them.  The lowest channels have Lancian mammals in them,
not Puercan, and still show a drop in numbers of dinosaur teeth.
 > 
 > 
 > "Yep, so what?  All that shows is that the marine extinctions
 > were relatively abrupt...
 > I am *not* going to repost my summary of the evidence that
 > extinctions in general are *multicausal*, and usually associated
 > with flood basalt style volcanism."
 >      Multicausal models that casually dismiss evidence in this way  
 > are dangerously unparsimonious.

What evidence am I casually dismissing?
As far as I know I am using *all* of it!

Remember - I am allowing that the impact was one of those multiple
causes.  Abrupt extinction of some marine organisms is
perfectly within the range of my model.

Note, there are also some marine organisms that do *not*
show abrupt extinctions, but rather stepped extenctions.
[For instance brachiopods].

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

The peace of God be with you.