Statistics and the fossil record [was: Re: EARLY EVOLUTION OF 'BIRDS']

["Jonathan R. Wagner" <znc14@TTACS.TTU.EDU>]

        I must admit that I am neither a statistician, nor an avid follower
of the statistics/fossilization/phylogeny arguments which have been
occurring on the list. My brief perusal of these debates has lead me to the
conclusion that participants may not be giving sufficient consideration to
the nature of the terrestrial stratigraphic record as it relates to
vertebrate fossil deposits. I cannot sit here and give you a basic
historical geology lecture, nor can I give you the paleontology and
stratigraphy lectures many people on the list azpparently need. I highly
encourage everyone involved in this debate to avail themselves of geological
resources before continuing this discussion. It is vital to your
understanding of vertebrate paleontology that you do so.

        To help you along, I'll give a brief summary of some of the
highpoints of the problem:
        1) Rock does not equal time: A rock body (of the type we are
interested in) is neither a snapshot nor a long-exposure photograph of a
particular period of time. It is a series of net deposits accumulated over
one or many periods of time, and may include significant hiatuses.
        2) Sedimentary rocks do not form everywhere: This is especially true
of terrestrial settings. Sediments are deposited in sedimentary basins, and
many, if not most places on the continents are not sedimentary basins.
Further, the distribution of sedimentary basins in time and space is by no
means equitable. Terrestrial deposits are certainly less common than marine
deposits (including here coastal deposits).
        3) Sedimentary rocks do not last forever: They can be weathered and
eroded, leaving significant gaps the record of a basin, a region, or the world.
        5) Fossilization is rare: We all know this. Bones are sedimentary
particles, their presence or absence in a sedimentary environment prior to
fossilization is based on a complex series of factors, as are their chances
of preservation. These factors seem to be overall decidedly against
preseveration.
        6) Fossilization potential is variable: Not all environments
preserve fossils equally well, in terms of abundance and completeness.
Generally, it seems safe to say that fossilization potential in terrestrial
environments just isn't all that good, but is better in some than others.
        7) Fossil accumulations are not always a snaphot of a fauna:
Preservation is not necessarily evenly distributed about a fauna. A deposit
may approximate the fauna quite well, or not at all. It cannot be assumed a
priori to represent the extent of faunal diversity in the depositional
environment.
        8) World faunal diversity is variable in space and time: No
environment carries a complete record of world faunal diversity. Indeed, we
should hardly expect any environment to even approach being a good sample of
terrestrial diversity, except perhaps at the grossest taxonomic levels
        
        As a first-order approximation, the above does not do justice to the
extent of the problems with the terrestrial vertebrate fossil record,
especially problems in applying statistical methods to it. Fossilization
does not produce a random sample of the world fauna; fossils are
non-randomly distributed with respect to space, time, morphology, and
phylogeny. Your sample is therefore inappropriate for many statistical methods.
        
        Wagner
--------------------------------------------------------------------------------
     Jonathan R. Wagner, Dept. of Geosciences, TTU, Lubbock, TX 79409-1053
 "Only those whose life is short can truly believe that love is forever"-Lorien