|Apparently, I am not making my point. My guess is that what substitutes
|for rules is merely that an index fossil must fit the paradigm. For
|example, one characteristic is that a useful index fossil must have
|seemed to have lived for a short period of time.
Your guess is incorrect. As Jim has described, index fossils are
not strictly delimited. They are the most precise end of a gradient of
usefulness for subdivision of the Earth's stratigraphy. They are
recognized because their first or last appearance occurs in a regionally
or globally *consistent* order with respect to the stratigraphy (i.e.
lower or higher) and other fossils when they are "in situ".
|If it is found with fossil species [including other index fossils]
|thought to have existed for over 150 million years, it would not be used
|as an index fossil. This sounds like cyclic to me reasoning to me.
Yes, it would not be used as an index fossil, because it is does
not provide a "useful" (i.e. *precise* enough) subdivision. No, it is not
cyclic reasoning. It isn't even close. Taxa with longer ranges (e.g.,
the entire group of dinosaurs) *still* occur in a consistent stratigraphic
position relative to other fossils, when in place. For example, they
consistently post-date (i.e. are above) the occurrence of all trilobite
species. This is a broad distinction which only separates Paleozoic from
Mesozoic rocks, but it is still consistent world-wide, and individual
species of trilobites or dinosaurs within that interval might be used to
establish finer subdivisions. If so, they could be called "index
fossils". In practice, there are species with wider geographic
distribution (e.g., global) and more common than dinosaurs or trilobites,
so these others are far more likely to be used for biostratigraphic
subdivision. Planktonic organisms turn out to be particularly useful.
Think about the way they disperse and it makes sense.
|There are no independant proofs that a particular index fossil lived for
|merely 10 million years.
Putting aside the issue of "proof" for the moment, yes, there is
independent evidence, in the form of radiometric dates (which also
indicate the expected order of occurrence), and there is ample other
independent evidence that a particular index fossil occurred during a
particular interval, and is unlikely to extend a significant distance
higher or lower. There is also ample independent evidence for the
correlation of time intervals from one area to the other on the basis of
fossils. For example, volcanic ash beds with precise, distinct chemical
compositions can often be traced laterally over large distances. None of
these are dependent upon, for example, evolutionary theory, contrary to
the claims of some "young Earth" creationists. In fact, there is little
in biostratigraphy which depends upon evolutionary theory. John Smith was
subdividing the rocks of England using fossils decades before Darwin
proposed evolutionary theory.
Fossils are not used to establish the "10 million years". They
are only used to establish the subdivision of the Earth's stratigraphy.
That subdivision can then be calibrated by other methods (e.g.,
radiometric dating, which is the most precise one). The subdivision of
the Earth's stratigraphy by fossils is based on the recognition of unique
events -- the first appearance and last appearance of distinctve species.
It does not matter whether these events are caused by evolution by natural
selection, "hydrodynamic sorting" in a "global flood", or directly by God
Himself. The procedure and principles are the same. Ninenteenth-century
"flood geologists" used basically the same principles for biostratigraphic
subdivision as geologists/paleontologists do today. Evolutionary theory
may help some of the finest subdivisions, but really is not a necessity
for most, or a significant bias on what is accomplished in
biostratigraphy, particularly at the broad scale usually considered.
The occurrence of fossils in a particular order is a testable
hypothesis which can never be "proven". Science does not deal with issues
of definitive "proving". The hypothesis can only be tested by more
sampling. The basic stratigraphic order of fossils has been documented
and its interpretation tested since at least the late 18th century.
"Index fossils" merely reflect the increased precision and geographic
scope of subdivision since that date (for example, there are many ammonite
species used to subdivide the Late Cretaceous into more than 30
intervals). To a first approximation, the broad sequence you see
documented in the geological literature has been very, very thoroughly
tested. There *are* continual refinements to the stratigraphic
distribution of fossils as sampling continues, but these are usually found
in the distribution of rare species (which is precisely why they are not
recommended for use in biostratigraphy), or they amount to very small
adjustments a non-specialist would hardly notice or care about.
|Thus any rules to prove that the fossil species existed for only 10
|million years would be obviously weak.
It is not usually possible to say a species "existed for only 10
million years". Numerical calibration of an interval of occurrence
invariably has uncertainty associated with it because of instrumentational
uncertainty and stratigraphic uncertainty associated with the numerical
dates (usually radiometric) used to constrain the numerical age. A time
of occurrence would usually be quoted as, for example, 10+-2 million
years. In popular literature or summaries, such uncertainties are often
dropped, unfortunately.
If index fossils can subdivide an inteval estimated to be 10+-2
million years into (for example) 10 index fossil zones, then the
uncertainties on the numeric ages bracketting the top and bottom of all
ten zones will swamp the estimates for individual zones. For example, the
duration of the zones could be estimated at 1+-2 million years (the
statistics are a bit debatable). Looks silly, doesn't it? But this is
why biostratigraphic subdivision is still so highly valued -- it is still
usually higher resolution than radiometric dating. The numerical date
would seem to suggest the interval could be 3 million or zero duration
:-), but this is just an artifact of insufficient sampling and precision
to resolve such fine subdivisions using numerical dating techniques. This
large uncertainty does *not* reflect uncertainty about the *relative* age
of the interval, because the numerical duration of a fossil interval in
"millions of years" is a completely different question from the issue of
its relative stratigraphic position.
|What say you?
What have you looked at from the conventional literature? In
introductory geology texts, index fossils are usually mentioned, but not
discussed in any detail. The following book has a pretty approachable
discussion of the procedure for biostratigraphy and other techniques used
for stratigraphy:
Blatt, H.; Berry, W.B.N.; and Brande, S., 1991. Principles of
Stratigraphic Analysis. Blackwell Scientific Publications: Boston,
p.1-512. ISBN 0-86542-069-6.
If you can not get this one, or if you want more details, I can
recommend others; and if you have any questions, do not hesitate to ask.
I regularly use biostratigraphy in my own work. There are circumstances
in which it could be used in a circular fashion, but in practice these can
occur only when trying to distinguish very fine subdivisions, and are
still independently testable. I can explain this if you want, but it does
not affect the broad sequence usually documented in more popular
literature.
-Andrew
macrae@geo.ucalgary.ca
home page: http://www.geo.ucalgary.ca/~macrae