Paleonet: Biostratigraphic datums - Signor-Lipps
Jere H. Lipps
jlipps at berkeley.edu
Wed Apr 18 19:55:37 BST 2007
The Signor-Lipps Effect (I wish now I had insisted on senior
authorship!!) was described by us in 1982 in:
Signor, Philip W., III, and Jere H. Lipps. Sampling bias, gradual
extinction patterns and catastrophes in the fossil
record. Geological Society of America Special
Papers. 190:291-296. (I am sorry but I do not have pdf's or
reprints left of this paper.)
Dave Raup later dubbed the effect of the gradual extinction patterns
"the Signor-Lipps Effect", although I don't know where.
I with other people have continued to look at this effect. It
indicates that the true ranges of species are seldom accurate for two
major reasons: taphonomy and population dynamics. These two result
in lost intervals in the ranges of species--in other words, species
are seldom present over their entire evolutionary histories in any
section and they may come and go in a section. Taphonomic
considerations--preservation, diagenetic destruction, failure to find
specimens, etc--have been considered to be the chief cause of the
effect, and much effort has gone into more thorough searches of
sections with terminations or gaps in species ranges. The second
cause, population dynamics, may be even more important, however, and
there's no way to determine this unless the species are encountered
again up section. Two major components of this are
important--migration in and out of the area and population declines
to very low numbers. We did a study of foraminiferal species ranges
in a Miocene section that was sampled about every 1.5 m. At every
sample horizon, species ranges showed a gradual decline up to that
point in the section, even when the species reappeared later. We
attribute this to the population dynamics--species come and go or
decline in numbers in space and time.
My feeling about Martin's question is to use the actual occurrences
in the section. If the species is known to occur higher in other
sections, then that can be considered. Otherwise, it is gone, and no
amount of additional hunting may reveal it again. We should also
plot biostratigraphic data as occurrences in sections rather than
ranges in which species with gaps in occurrences are assumed to be
continuous in that section. Likely, they were not present during the
gaps but migrated or died out locally and returned later. We are
probably missing something in the environmental interpretations by
ignoring these gaps. Marty Buzas and Steve Culver have looked at
this for foraminifera on the Atlantic and Gulf coasts within formations.
This also brings us to the sampling problems that face
micropaleontologists. We usually assume that 300 specimens is
enough to examine. This number came from the sedimentologic
literature where people were trying to characterize the mineralogy of
the sediment. Since there's seldom very many different kinds of
minerals (~5) in sand, 300 grains works. But when we encounter 80
or more (what if there's 350 species present??) species in different
proportions, then 300 is insufficient. Instead, micropaleontologists
and probably others too, should use rarefraction techniques in which
specimens vs species are counted until the curves level off, at which
point we can assume that the major species have been found (but
probably not all). This could mean 1000 or more specimens need to
be identified--every sample will be different. I recall one sample
set where I knew a species occurred (Al Loeblich found one but he
could find a micro-needle in a haystack), but it took about 10,000
specimens to find one of it. It was rare, indeed, and even
rarefraction would not have revealed it, unless I had more patience
than I demonstrated that day.
Jere
At 09:32 AM 4/18/2007, you wrote:
>For discussions of this subject, see Walsh (1998; 2000), as well as
>the papers by Signor and Lipps and Lipps and Signor that Peter
>Sheehan just alluded to (I can't find the references at the moment).
>
>Walsh, S. L. 1998. Fossil datum terms, paleobiological event terms,
>paleontostratigraphy, chronostratigraphy, and the definition of
>land-mammal "age" boundaries. Journal of Vertebrate Paleontology 18:150-179.
>
>Walsh, S. L. 2000. Eubiostratigraphic units, quasibiostratigraphic
>units, and "assemblage zones." Journal of Vertebrate Paleontology 20:761-775.
>
>Steve Walsh
>Dept. of Paleontology
>San Diego Natural History Museum
>1788 El Prado
>San Diego, CA 92101 USA
>slwalsh at sdnhm.org
>
>
>
>
>-----Original Message-----
>From: paleonet-bounces+slwalsh=sdnhm.org at nhm.ac.uk on behalf of Sheehan, Peter
>Sent: Wed 4/18/2007 9:20 AM
>To: PaleoNet
>Subject: Re: Paleonet: Biostratigraphic datums - a question
>
>It also depends on how common the fossil is in the sample
>interval--if it is found in only a quarter of the samples in the
>fossils range the actual top will most likely be farther up than
>another fossil found in every sample in the fossils range. Strange
>that no one has mentioned the extensive literature on the Signor-Lipps effect.
>Peter Sheehan
>
>-----Original Message-----
>From: paleonet-bounces+sheehan=mpm.edu at nhm.ac.uk
>[mailto:paleonet-bounces+sheehan=mpm.edu at nhm.ac.uk]On Behalf Of Martin J. Head
>Sent: Tuesday, April 17, 2007 6:46 PM
>To: PaleoNet
>Subject: Re: Paleonet: Biostratigraphic datums - a question
>
>
>Dear Tony,
>
>Many thanks for your comments, which I certainly appreciate. As a
>matter of clarification, I should just say that this is not a
>methodology I have yet practiced, but we are considering it as a
>means of expressing uncertainty in observed HOs (and LOs). The
>datum I have in mind would express the age as a mid-point between
>the two bracketing samples, and provide an error (+/-) based on the
>ages of those two samples. This seems a legitimate way of
>expressing uncertainty, but perhaps it is extending the data too
>far. I should also point out that the methodology would be applied
>to a DSDP hole for which there is a very precise age model, such
>that sample spacing becomes the main source of error (which is why
>we are trying to express it somehow).
>
>I can't say I've ever seen this methodology applied to
>biostratigraphic HOs/LOs etc., but I think it is common to construct
>a zonal boundary at the midpoint between two samples, and average
>the age of the samples to calculate the age of the boundary. In a
>sense, this is an analogous situation.
>
>Again, thoughts on this approach would be greatly appreciated.
>
>Martin
>
>
>On 17 Apr 2007, at 19:04, Tony D'Agostino wrote:
>
>
>In 27 years of industry biostratigraphy I have never seen or heard
>of usage described by Martin Head (below). In the oil business that
>would be what we refer to as "made up data". A HO datum (please no
>Don Imus comments) is routinely adjusted to the top depth of a
>sample interval when working with cuttings material or any kind of
>sample that is a composite over a stratigraphic interval.
>Occassionally I've seen workers assign a datum to a midpoint of such
>a sample interval, but I've never encountered, let alone used the
>method described by Dr. Head. Why would one interpolate a depth for
>a HO datum at some point that is not sampled or examined, providing
>a basis for observations? Best practices dictate that you assign the
>datum to some depth within or at the end points of the sample
>interval. At least you know with certainty, if not precision, where
>that the fossil was observed to occur. Leave the "interpolation" to
>the seismic interpreters.
>
>Tony D'Agostino
>20746 Prince Creek Drive
>Katy, Tx. 77450
>281-646-1660 adagostino at houston.rr.com
>
>"The limits of a tyrant are determined by the endurance of those
>that oppose him" Frederick Douglass
>
>-----Original Message-----
>From: paleonet-bounces+adagostino=houston.rr.com at nhm.ac.uk [
>mailto:paleonet-bounces+adagostino= houston.rr.com at nhm.ac.uk]On
>Behalf Of Martin J. Head
>Sent: Tuesday, April 17, 2007 5:27 PM
>To: PaleoNet
>Subject: Paleonet: Biostratigraphic datums - a question
>
>
>Dear Paleonetters,
>
>I have a question concerning the nomenclature of a particular
>biostratigraphic datum.
>
>Although highest occurrences often refer to the highest sample in
>which a taxon is found in a particular section, the real highest
>occurrence will likely be at some interpolated point (normally
>placed at the midpoint) between the highest observed occurrence and
>the next sample up. The difference in position between the observed
>HO and the inferred HO will depend on the sample spacing.
>
>Is there a term for this inferred/interpolated HO?
>
>I'd wondered if HOD (highest occurrence datum) might be the
>appropriate term, but I believe this refers to a three-dimensional
>surface of HOs whereas my "inferred HO" could refer to a single
>stratigraphic section (i.e a point).
>
>I hope the answer is not too obvious! Thanks for any help or ideas.
>
>Martin
>
>
>------------------------------------------------------------------------
>Martin J. Head
>Professor
>Department of Earth Sciences
>BROCK UNIVERSITY
>500 Glenridge Avenue
>St. Catharines, Ontario L2S 3A1
>CANADA
>Tel 905-688-5550 ext. 5216
>Fax 905-682-9020
>Email mjhead at brocku.ca
>www.brocku.ca/earthsciences/people/mhead.php
>
>
>
>_______________________________________________
>Paleonet mailing list
>Paleonet at nhm.ac.uk
>http://mailman.nhm.ac.uk/mailman/listinfo/paleonet
>
>
>
>
>
>
>
>
>------------------------------------------------------------------------
>
>Martin J. Head
>
>Professor
>
>Department of Earth Sciences
>
>BROCK UNIVERSITY
>
>500 Glenridge Avenue
>
>St. Catharines, Ontario L2S 3A1
>
>CANADA
>
>Tel 905-688-5550 ext. 5216
>
>Fax 905-682-9020
>
>Email mjhead at brocku.ca
>
>www.brocku.ca/earthsciences/people/mhead.php
>
>
>
>
>
>
>_______________________________________________
>Paleonet mailing list
>Paleonet at nhm.ac.uk
>http://mailman.nhm.ac.uk/mailman/listinfo/paleonet
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.nhm.ac.uk/pipermail/paleonet/attachments/20070418/41e6c13e/attachment-0001.htm
More information about the Paleonet
mailing list