# No subject

Ramella Lorenzo RAMELLA at CJB.UNIGE.CH
Fri Aug 5 16:12:25 CDT 1994

```     Conservatoire=20
et Jardin=20
botaniques=20
de la Ville                      Date   :   05-Ao=FB-1994 13:5=
7 GMT
de Gen=E8ve                      de     :   Lorenzo RAMELLA=
=20
Service:   ENSEIGNEMENT
No t=E9l :   732 69  69 int. =
206

A:  VMSMail User _IN%"longinoj           ( _IN%"longinoj at ELWHA.EVERGR=
EEN.EDU" )
Sujet: RE : geographic data                                          =
=
=20

PROYECT FLORA OF PARAGUAY - GENEVA (Switzerland)
=20
=20
I read your message on geographical data. We are working with a=
=20
specimen-oriented Database on the Flora of Paraguay and other proje=
cts in=20
the CJB (designed by Dr. Catherine Zellweger). Geographical names h=
ave=20
been one of our "hot" problems for a long time. We started using th=
e=20
US-defense gazetteer, but we found out that a lot of place names,=
=20
coordinates etc. did not correspond to reality. For this reason, we=
=20
decided to create our own gazetteer based on the official map of Pa=
raguay=20
(scale 1:1 mio). For that purpose we created our own standards for=
=20
description of locality, in particular such locality known as river=
,=20
it to=20
our standard. In the end we got more or less 1'500 names. For the=
=20
calculation of coordinates, we digitalized the Paraguayan map, foun=
d out=20
the coordinate calculation algorithm, and eventually use the map. W=
e used=20
a digitalization table to calculate each coordinate. I do not want =
to go=20
into details on all the problems of standardisation, verification,=
=20
correction, etc.
=20
But all this work was not sufficient to find out coordinates for th=
e=20
locality description on specimen labels. To solve this problem we=
=20
distinguished the description found on the label and the standard=
=20
locality attributed to one specimen. This means, that for each spec=
imen=20
we have to determine to which standard locality the geographical=
=20
information found on the label corresponds. If there is no standard=
=20
locality, we have to find it from our collection of maps (1:250'000=
,=20
1:50'000), books, etc. The particular problem is the constant chang=
ing=20
nomenclature of place names, in particular with old geographic data=
. In=20
this way we can complete our gazetteer according to our specific ne=
eds=20
(the Paraguayan botanical collection).
=20
Another problem is the accuracy of the geographic data of specimen.=
At=20
this moment, we make a distinction between approximate and precise =
data.=20
But this is not accurate enough. We'd rather define it as a range o=
f=20
precision (1 second, 1 minute, 5-10 min, etc..). We could find some=
=20
interesting consideration in litterature, such as RHOADS (1992) in =
Taxon=20
41(1).
=20