The use of corpora to verify how words are used in context can be particularly useful in areas of LSP and terminology. The very possibility of examining terms in texts that deal with the subject matter under discussion makes it easier for the terminologist - and also for the translator - to understand the concepts behind the terms used. However, whereas one can follow the advice of Pearson (1998) for monolingual texts in English, it is less easy to find comparable or parallel texts which are reliable, of an equal level of complexity and which offer the same information value. This paper looks at efforts made to find and use text corpora in English and Portuguese to further terminology work in various areas, the problems that arise and how they can, or cannot, be solved.