BESL is a complete database of the English lexicon. It consists of over 230,000 lemmas, over 350,000 word forms, 60,000 proper nouns, 3,000 abbreviations, and 58,000 multi-word compound nouns. Each headword is provided with a full listing of all inflected forms and other morphological variation. Every word form is marked for part of speech (using Penn TreeBank notation). Most single-word forms include a representation of IPA pronunciation. BESL covers both British and American English, and other spelling variants, with cross-references between corresponding forms. Each lemma is graded on a scale between 1 and 9 to indicate frequency, based on corpus evidence. Lemmas are also classified by domain, where appropriate (e.g. Computing, Religion). Obscene or offensive lemmas are marked using a 2-grade system. Proper name lemmas in BESL include personal names, surnames, place names, and brand names. BESL is provided in XML.