UNTIL RECENTLY, William's florilegium of extracts from classical and patristic sources, the Polyhistor, attracted little attention. Now, however, that florilegia are seen to be illuminating cultural documents, not merely quarries for rare texts or useful readings, the Polyhistor has been edited, and some of the extracts in it discussed.L. D. Reynolds, for instance, showed that William was the first person after late antiquity to know the complete correspondence of the younger Seneca, which had circulated in two parts. William included excerpts from both parts in the Polyhistor, and his principle of selection was characteristically individual: he was more attracted by Seneca's wit than by his moralistic or edificatory utterances. This is characteristic of the content of the Polyhistor as a whole, which hardly reflects its stated aim. It is a collection of ‘dicta et facta memorabilia’, considerations of style and moral instruction being secondary though by no means absent.

In this chapter I wish to draw attention to another set of extracts in the Polyhistor, those from the Noctes Atticae of Aulus Gellius. Gellius was not a common author in early Europe. In the ninth century his work was known, in part, to Lupus of Ferrières, but in the twelfth century a sizeable group of (mostly English) literati was familiar with it, and it passed into several well-known florilegia. Surviving manuscripts of the Noctes Atticae confirm the impression that between these centuries Gellius seems to have suffered almost complete oblivion. Like Seneca's letters, the Noctes circulated in two sections after late antiquity: books i–vii, and ix–xx, the contents of book viii having been lost early on. Lupus knew only the later section; William, a generation earlier than John of Salisbury, knew both parts of the text, though whether complete or in the form of excerpts is less easy to determine. William is to our knowledge the first man since late antiquity to know the complete text of Seneca's letters; in the case of Gellius only the compiler of the florilegium known as TY may have known both parts of the textual tradition before William. William's extracts, moreover, are numerous, of reasonable length, normally verbatim and textually valuable.