Answers to domain corpora request

Thanks to everyone who answer my request for open-source domain corpora. 
Leonel Ruiz and Stella Tagnin pointed me to corpora in Spanish and 
Brazilian Portuguese. For English, Ylva Berglund mentioned OPUS (an open 
source parallel corpus). From the text mining front, big textual 
collections of Bio-Medical full-text articles are now available, as 
pointed out by Paul Buitelaar (http://muchmore.dfki.de/resources1.htm) 
and Kevin Cohen (http://www.biomedcentral.com/info/about/datamining/ 
[8,000 plus articles in xml]), among other data collections. Also, the 
Linux Documentation Project provides a quite big, typological 
homogeneous collection.
Unfortunately, big textual collections from other disciplines are more 
difficult to obtain in dowloadable form.  I am now compiling a 300 
article collection from Sociology journals, in case anyone is also 
interested in cross-genre comparatives  and lexical acquisition.

Carlos Rodríguez
National Autonomous University, Mexico