arrangement I have worked out with great labour; yet not I, but the grace of God with me. I entreat you therefore, reader, do not contemn my work as something rude and barbarous.' The most striking feature of the dictionary is its etymology. Almost every word is supplied with a derivation, often very far-fetched. Thus glisco is derived from 'glykis, quod est dulcis; que enim dulcia sunt desiderare solemus': gliscere therefore is equivalent to desiderare, crescere, pinguescere and several other words. After this we are not surprised at the following account of a dormouse. 'Glis a glisco: quoddam genus murium quod multum dormit. Et dicitur sic quod sompnus facit glires pingues et crescere.' Here is another piece of natural history. 'Irundo ab aer dicitur: quia non residens sed in aere capiens cibos edat, quasi in aere edens.' There is simplicity in the following: 'Nix a nubes, quia a nube venit.' Again: 'Ouis ab offero vel obluo: quia antiquitus in inicio non tauri sed oues in sacrificio mactarentur. Priscianus vero dicit quod descendit a Greco ... oys.' Besides his philology the good Dominican was also a theologian; and when he comes to the words upon which his world was built, he cannot dismiss them as lightly as the snow. So Antichristus has two columns, that is to say a folio page: confiteor 11/2, conscientia 21/4, ordo 21/2, virgo two columns. Much light is thrown on Balbi's work by the dictionary of his predecessor, Huguitio of Pisa, Bishop of Ferrara (d. 1210). The title of this, _Liber deriuationum_, indicates its character. Instead of the alphabetical principle the words are arranged according to their etymology; all that are assigned to a given root being grouped together. This made it necessary, or at any rate desirable, to find a derivation for every word; and with ingenuity to aid this was done as far as possible. Besides derivatives even compounds came under the simple root; and in consequence it must have been extremely difficult to find a word unless one already knew a good deal about it. It is no wonder that the book was never printed; although it occurs frequently in the catalogues of mediaeval libraries.