Data localities

Robert Mesibov mesibov at SOUTHCOM.COM.AU
Wed Oct 23 22:36:35 CDT 2002

I've encountered all the problems in this Taxacom thread while trying to
"upgrade" locality information on old museum labels. Worse, I still meet
field workers who record localities _today_ as constructs of words, without
accompanying numerical location data (latitude/longitude or UTM).

You can understand why verbal locations are popular. We use them all the
time in everyday speech ("on the south shore of Clear Lake"), and they can
apparently be made fairly exact: "3.5 km west of the Little Creek bridge on
Mt Thompson Road". The fundamental problem with verbal locations is that
they work by linking an unnamed locality to an already named one, i.e. to a
landmark. If the landmark disappears, moves, changes its name or loses its
name, the link is lost. Names are simply not reliable enough for spatially
referencing a specimen locality. The best spatial reference is the
latitude/longitude system, which is universal, has outlasted thousands of
names and will outlast many more. The UTM system (the basis for Australian
and many other map grids) is just as good, because it can readily be
converted to lat/long.

"Verbal location" enthusiasts tend to go glassy-eyed when numbers are
mentioned. I gently steer such people in the direction of GPS units. I also
point out all the problems already noted in this Taxacom thread, plus 3 more:

(1.) Verbal locations are almost unavoidably vague.

A verbal location typically gives to a locality a name which covers a much
larger area. For example, many plant and animal specimens in Tasmanian
collections are labelled "Mt Wellington." The verbal location "Mt
Wellington" covers about 25 sq. km of very diverse topography and
vegetation. Many aquatic specimens are labelled "Great Lake", whose current
area is something like 180 sq. km. These verbal locations satisfied the
collectors, yet in almost every case the specimen probably came from a spot
with an area less than one hectare, i.e. 1/100 th of a square kilometre.

Spatial blurring of this kind can be avoided by careful use of distances
and directions in a verbal location, but in an area with few place names,
this kind of qualification can be tedious. A recent invertebrate survey
looked at a site "About 2 km E of Big Heathy Plain by Ben Nevis, NE
Tasmania, about 200 m upstream from the NE-to-S-flowing bend in a mainly
S-flowing tributary of the unnamed creek between Farrells and Memory
Creek." That awkward but accurate construction narrows it down to about 5
ha. The site was actually defined by the collector as 55GEQ557176, a 1 ha
UTM grid square.

(2.) Verbal locations don't have universally accepted set-up rules. Here
are three clearly specified verbal locations in eastern Tasmania:

2 km W of Watch House Bay, S of Little Swanport
1 km S of the junction of Mitchelmores Creek and Little Swanport River
1 km E of the junction of River Road and Swanston Road, N of the Buckland
Military Training Area.

Without looking at a map, I doubt if even a local resident would know that
these three actually refer to the same place (a low, unnamed hill within
the 1 km UTM square 55GEP7411).

(3.) Verbal localities are strangely attracted by named places. The
specimen locality is 5 miles south of Hicksville, but the label just says
"Hicksville" because that's the nearest named place in the gazetteer. I'd
like to think that this strange attraction is something that _used_ to
happen, but doesn't nowadays, only I know different.

If only label-writers would keep in mind the people who read those labels:
us poor, ignorant taxonomists and museum curators, who only know what we read.

Dr Robert Mesibov
Honorary Research Associate
Queen Victoria Museum and Art Gallery
Home contact: PO Box 101, Penguin, Tasmania, Australia 7316
(03) 64371195; 61 03 64371195

More information about the Taxacom mailing list