Varro wrote three books De bibliothecis, according to a list by Jerome (Ep. 33.2 = testimonium 23 GRF Funaioli). The work may have had something to do with his commission to build a massive public library for Julius Caesar (Suet. Iul. 44.2), though Caesar was assassinated before the library could be built. It may also have some connection to Rome's first public library, which Asinius Pollio added to the Atrium of Liberty in the 30s b.c. Pollio, after all, gave a portrait to Varro alone among living authors (Plin. HN 7.115). The known fragments are few.

There are places where Varro is quoted on book-related matters, but without an attribution to his De bibliothecis. One example is a note on the history of papyrus in Pliny's Natural History (13.69–70). Pliny reported that Varro had written about the origins and history of papyrus and other writing materials (palm leaves, bark, lead, linen, wax). Varro apparently wrote that papyrus was only discovered after Alexander the Great had founded Alexandria, and that parchment was invented at Pergamum in response to a papyrus embargo from Egypt, which had been instituted when the Ptolemies became jealous of the rival library. Such a history of papyrus could easily have fit in the De bibliothecis, but it could just as well have fit in some of Varro's voluminous other writings.

It is sometimes supposed that all of Isidore of Seville's Origines 6.3, 6.5 and 6.9–14 come from Varro, with Suetonius as an intermediary.Footnote ² This is certainly possible. Isidore's treatment of books and libraries in his Origines included chapters on writing materials (wax at Orig. 6.9, papyrus at 6.10, parchment at 6.11, various others at 6.14). But we should keep in mind that the only certain connection between Isidore's Book 6 and Varro's De bibliothecis is in subject matter (i.e. books and libraries). The role of Suetonius as intermediary is based on the fact that Isidore cited Suetonius elsewhere, and on the fact that Isidore made use of Suetonius once in regard to a book-related matter.Footnote ³ Given these tenuous connections, the temptation to use Isidore to reconstruct Varro's De bibliothecis should be resisted.

There are only two securely attested fragments, both preserved in Charisius, who wrote his Ars grammatica around a.d. 362.Footnote ⁴ One simply remarks that in the second book of Varro's De bibliothecis he used the form uectigaliorum rather than uectigalium (fr. 53), and the other that he used the term gluten in the phrase ‘he fixed it up with glue and citron-wood’ (glutine et citro refecit,Footnote ⁵ fr. 54).

The reading is not very satisfactory. The gluten suggests that Varro is talking about repairing a scroll.Footnote ⁶ Yet citron-wood is not known to have been used in the manufacture or repair of book-rolls. The quotation from Varro is cited in two places in Charisius, and the editio princeps (1532) emended the first of these to cinere (the one at p. 110 Barwick), an emendation that Fabricius used for the second (at p. 167 Barwick) in his edition of 1551.Footnote ⁷ Presumably they believed Varro was referring to the use of ash in ink. Both Keil and Barwick rejected this emendation, so they must have found the reading citro defensible. Seneca (Tranq. 9.6) makes reference to citron-wood bookcases (armaria), so they may have accepted the reading on the grounds that Varro was writing about repairing bookcases. It is true that gluten could be used in wood-working (see Lucr. 6.1069–71), but it seems unlikely that citron-wood would be used along with gluten as a part of a repair. Citron-wood was an expensive luxury item, often mentioned alongside gold or ivory.Footnote ⁸ In fact, moralizing rhetoric at times used the combination of citron-wood with gold or ivory to suggest an over-the-top extravaganceFootnote ⁹ – and this is what we see in Seneca. He is not referring to any actual citron-wood bookcases, much less suggesting that these were a common item. To the contrary. He asks (Tranq. 9.6): quid habes cur ignoscas homini armaria <e> citro atque ebore captanti? ‘What indulgence should you show for the kind of man who would try to acquire bookcases made from citron-wood and ivory?’ Seneca uses the idea of citron-wood bookcases to categorize a certain kind of individual: the kind who would spend a fortune on the accoutrements of literature but not spend time on literature itself.

A much better reading would be cedro. Cedar was most definitely used in book-roll production and repair. Pliny the Elder writes that cedar had fungicidal and pesticidal properties.Footnote ¹⁰ Papyrus was frequently treated with cedar oil for this reason. As Vitruvius writes (2.9.13): ‘When other things, like books, are treated with cedar-oil, they are not damaged by worms and rot’ (ex cedro oleum ... quo reliquae res cum sunt unctae, uti etiam libri, a tineis et carie non laeduntur).Footnote ¹¹ The darker colour of cedar-coated papyrus was also aesthetically prized.Footnote ¹² Modern finds of ancient papyrus attest to the presence of cedar, especially for scrolls containing literary works.Footnote ¹³ In his TLL entry for citrum (TLL 3.1207), faced with the fragment of Varro in question, Stadler suggested that in this case citrum perhaps referred to cedar, yet such a usage would be entirely unparalleled. Rather, this is a mistake that should be corrected.

The similarity of sound and semantic field between cedro and citro would make the corruption an understandable one, especially if the copyist was not knowledgeable about the manufacture and repair of papyrus. Therefore, the fragment of Varro's De bibliothecis reading glutine et citro refecit (fr. 54 GRF Funaioli) should be emended to glutine et cedro refecit.