This work describes some grammatical features of the Beja language, spoken in the north-east of the Republic of the Sudan. This language is normally unwritten, and the description is based entirely on material obtained from informants. The theoretical basis of the description is a modified version of the 'Scale and Category' model developed by Dr. M.A.K.Halliday. The model used here, and its relation to the orthodox Scale and Category model, are described briefly in the 'Introduction', and in more detail in the 'Outline', following the chapter on Phonology. The main body of the work is divided into seven chapters: one for each of the five 'units' of Beja grammar - morpheme, word, group, clause, sentence, in that order - one for the phonological realisations of morphemes ('morpho-phonology'), and one for some 'special' sentences, which do not fit into the most common patterns. The six appendices include: lists of regular and irregular verbal Radicals; verbal word-paradigms; an explanation of some transformationally related classes of clause; a few sentences of a story, with grammatical notes; and a glossary of technical terms used in the description. Probably the most controversial features of the work, from the point of view of the orthodox Scale and Category model, are: a. the 'scales' are given a relatively unimportant place; b. 'element' is treated as a category, on a par with !unit', etc.; c. 'elements' are treated, like 'units' and 'classes', as sets to which items belong; d. rather more kinds of class are distinguished; e. the concept of 'depth' is replaced by two concepts: 'complex element' and 'rankshift'; f. 'rankshift' is allowed a more legitimate place; g. the terminology differs in some places from orthodox Scale and Category terminology.