Robert Raven R.Raven at MAILBOX.UQ.OZ.AU
Sat Jun 10 09:19:34 CDT 1995

We have three levels of confidence in our existing database of spiders.
bear in mind that 80% of what we identify is unnamed.
First the taxonomiic table has a lvel of error given only by whether the
id is to family,genus,tribe, etc.
Second, the actual specimen has a "cf" value which operates again as a
text field and notes only genus uncertain, species uncertain, etc. So
that the printout would show something like Homo?? or Homo cf sapiens.
Third level is the most contentious because I refuse to budge for IT
managers who get our data. It is the level of accuracy in the locality
field. Some of our records, like most museums, date back to times when
place names were different and/or more encompassing. The "new" GIS folk
want an actual numeric measure in minutes & seconds. We cannot do that
because often our only place name is an irregular shaped object, eg a wedge
shaped island and its error dimensions are not consistent over the area.
Hence, I insist that the best we can do is to use error identifiers like,
unit (i.e. the object given), after that we use which method was used in
obtaining the coordinates. That may be a state or federal gazeteer, it
may be a calculated point (eg. 10miles NW of Bayshore; which is taken on
line of sight), it may be donor (the donor gave us actual cooridinates,
very rare), or we researched history of collecting of an individual to
establish the point. I find the attention to fine detail here rather
irritating because typically some of th animals we are talking about may
move several minutes, if not a degree in a day.
Not any easy task...Good luck
I am presently trying to design an interactive form that will allow
ecologists who want spiders identified to establish fairly closely how
much it will cost for various levels of error, eg. I do it all ("0%
error) to you do it all (80% error). Granting agencies here are funding
only about 80% of the requested fund on each grant, at best and taxonomy
is what gets dropped out of these proposals because I think it isn;t well
enough documented what they get.
The other project that bears upon your point is to design a taxonomic
database that is used to enhance the lists that we nomimally are required
to provide. These lists will contain data on how useful the taxon is for
biodiversity questions, e.g. an orb-weaving spider is widespread and
quickly recolonises burnt areas whereas a trapdoor is often more
localised and affected by trench digging.
The database summarises all the terribly useful information lying around
in the brains of taxonomists that rarely gets published because it is so

Robert J Raven
Museum Scientist (Arachnology)
Queensland Museum, Australia

More information about the Taxacom mailing list