1 Researching the origin of Perugia’s Public Library (1582/1623) before and after Material Evidence in Incunabula

The Biblioteca Augusta of Perugia is one of the earliest public libraries in Italy. Its core foundation is made up of the collection of Prospero Podiani, whose knowledge and interest in the book market led him to create one of the largest collections of sixteenth-century Italy, including some 300 incunabula. Panzanelli’s doctoral research approached documentary evidence pertaining to the collection in order to reconstruct the lively cultural context of the time and space it belonged to. Having transferred the incunabula data into the database Material Evidence in Incunabula , we can now offer a richer and more complex representation of the development and use of the collection, of the books’ survival, loss, and mobility, and of the methodology of the investigation involved.


Introduction
Among the many research topics that the database Material Evidence in Incunabula (MEI) can help to investigate, the reconstruction of historical libraries presently dispersed is certainly one of its most powerful outcomes. 2 Libraries around Europe and the United States merge into MEI the description of the copy specific features of their incunabula, including the recording of former ownership. This way, the disiecta membra of historical collections whose whereabouts are unknown to the libraries that today host fragments of such collections, are being brought together by a search engine. The multiple locations of a dispersed collection are immediately visible in the new MEI thanks to the default facets presentation by 'Present Holding Institution'. To offer just one example, by typing "San Giorgio Maggiore" or "Georgius maior" in the free text search, MEI will display the 78 copies so far recorded as coming from the Benedictine monastery of San Giorgio Maggiore in Venice. They are now preserved in Padua, Biblioteca Universitaria (33 copies), Paris, Bibliothèque nationale de France (8 copies), Chantilly, Musée Condé (2 copies), Manchester, John Rylands Library (2 copies). 3 Individual copies are found in Brescia Queriniana, Locarno Madonna del Sasso, London British Library, Milan Trivulziana, Padua Seminario Vescovile, Treviso Capitolare, Venice Marciana, and one copy in the trade. 4 Also listed are 25 'Historical copies' which we know were part of the books taken to Padua in 1806, after the dissolution of the monastery, and gathered in the ex monastery of Sant'Anna, as recorded by two documents at the University Library of Padua: Catalogo generale, ossia riunione di tutti gli elenchi de' libri scelti dalle biblioteche delle corporazioni regolari concentrate nel già convento di S. Anna di Padova. Catalogo de' libri scelti dalla libraria de' padri benedettini di San Giorgio Maggiore di Venezia (Padova, Biblioteca Universitaria, ms. 2250, pp. 101-152); Catalogo de' libri di qualche pregio separati da quelli di scarto rimasti al Demanio dopo la scelta fatta dai Delegati del cessato Governo Italico per la Pubblica Istruzione (Padova, Biblioteca Universitaria, ms. 2252, pp. 2-22). None of these 25 books has yet been located.
Some are probably lost forever, others, however, surely sit on the shelves of European and American public libraries, unrecognised as coming from the book collection of San Giorgio Maggiore of Venice because not yet catalogued with attention to their copy specific features.
As book in hand cataloguing progresses, and particularly if this is done in MEI, more historical copies will be matched with physical copies now surviving in libraries. Another option is the use of the browser to search the Index of Owners. 5 Let us not delude ourselves into thinking that tracking dispersed libraries is now a simple exercise. The ability of the search engine to capture such data is only the end of a very long process which begins with the willingness of an institution to survey its collections, the 4 MEI offers the possibility of recording copies sold by booksellers or at auction, and of easily tracking the movements of book. For example, a copy of the Livy printed in Venice by Vindelinus de Spira in 1470 and owned by the Benedictines of San Giorgio Maggiore was later owned by Lord Monson (1796-1862) and by his heirs sold through Sotheby in 2013, as attested by the catalogue, Music and Continental Books and Manuscripts, London 5-6 June 2013, lot 209. This copy is recorded in MEI no. 00200404, Holding Institution: Trade Copy (undefined); Shelfmark: Sotheby, London, 2013, lot 209. Its presence at Sotheby's is also given as the last block of provenance of the book. If the volume re-emerges in a public library or in a private collection whose owner agrees to share information in MEI, Holding Institution and Shelfmark will be changed to the current one, but the Sotheby's block of provenance will be there to witness that stage in the life and circulation of this specific copy of Livy. 5 This is a new feature of the new MEI. In line with other large indexes of provenance of incunabula, such as that of A Catalogue of Books Printed in the Fifteenth Century now in the Bodleian Library (=Bod-inc), ed. A. Coates, K. Jensen, C. Dondi, B. Wagner, and H. Dixon, with the assistance of C. White and E. Mathew; blockbooks by N. Palmer, and an inventory of Hebrew incunabula by S. Schaeper, 6 vols (Oxford: Oxford University Press, 2005), and Paul Needham's Index Possessorum Incunabulorum (=IPI) available on the CERL website at http://ipi.cerl.org/cgi-bin/search, personal names present the form, "surname, name, sigla of religious order" (e.g., Hoffmaister, Johannes, OSA); institutions are listed under the name of the city, then of the order, then of the dedication, and followed by the religious sigla (e.g., Venezia, Benedettini, S. Georgius Maior, OSB). It is planned that both indexes will be merged into MEI in the near future. ability of the cataloguer, and the understanding of the value of converging data in an international database like MEI. Essential as it is to get well acquainted with the content of our historical collections, we know that they are but part of a larger body, the one which can still be identified.
Documentary evidence recording the survey of libraries at different times through their history offers essential complementary information on the content and fuller extent of historical collections: it was therefore necessary to devise a system to bring documentary and physical evidence together. 6 The case study in this article will present the MEI solution.
MEI introduced an innovative approach to the recording of provenance: the application of geographical (GeoNames) and temporal indicators applied to every element of provenance, to track the movement of books over space and time during their 500 years of life. Now we are in a position to visualise the movement of thousands of books, and to understand patterns and trends in the use and survival of early printed books. 7 This innovative approach was devised by Cristina Dondi and developed by Alex Jahnke back in 2009 to answer the need to use not only conventional provenance to retrace the history of books (such as former owners' various marks) but physical evidence too: a German binding of the 16 th century, French illumination or Italian manuscript annotations of the 15 th century, or an English purchase note recording a price in the 19 th century. 8 We are pleased to be seeing this concept and model now implemented in similar contexts.
In MEI we have dozens of fields pertaining to provenance. We can single out distinctive copy features (variant, cancellans/cancellandum, parts of text in facsimile, parts of text in manuscript, manuscript signatures, absence of any rubrication, chain-holes); the type of provenance evidence is classified according to inscriptions, coats of arms, genealogical tables, supralibros, exlibris, mottos, emblems, stamps, fire stamps, binding, decoration, 6 See the article and references of Flavia Bruni in this volume, and the articles collected in Virtual visits to lost libraries: reconstruction of and access to dispersed collections. Papers presented on 5 November 2010 at the CERL Seminar hosted by the Royal Library of Denmark, Copenhagen, ed. by I. Boserup and D.J. Shaw, London, 2011 (CERL Papers, vol. 11).
7 Professor Min Chen and Dr Simon Walton of the Oxford e-Research Centre are applying scientific visualisation techniques to our MEI provenance data, as part of the 15cBOOKTRADE project. 8 Alex Jahnke of Data Conversion Group (DCG) at the University of Göttingen. DCG created and supports many of CERL's digital resources. manuscript notes, bibliographical evidence, shelfmark, accession mark, deaccession mark, and seller's mark/note. The method of acquisition can be distinguished as purchase, donation, bequest, exchange, institutional transfer, dedication copy, consignment, requisition, theft.
Each former owner is further defined in his/her gender, status (lay, religious), and profession. Decoration can be searched as illustration coloured in by hand, illumination, ornamental letters, coat of arms, rubrication, partial rubrication, pen trials, illustration stamped in. Manuscript annotations can be studied in the categories of corrections, completions, supplements, extraction of keywords, collation, 9 translation, structuring the text, comments, censorship, reading marks, drawings, corrections/notes by the printer, lecture notes, later rubrication, autographs. All of them can then be further described in their frequency and location in the book.
All MEI fields have been devised with clear historical questions in mind, questions which, until now, we were unable to answer. Now we can begin to address them. Just to offer one example: scholars of books and libraries are aware of the tremendous impact that the secularisation of religious institutions had on the dispersal and formation of libraries.
Historical collections in Europe and the United States would not be what they are today without the vast mobility of books triggered by national policies with regard to religious institutions and their libraries. In MEI we are finally capturing the extent of the phenomenon: to date, some 1,569 copies in MEI have been recorded with the descriptor "institutional transfer".
As mentioned above, in MEI we also devised a simple and effective way to capture "Historical copies", that is copies which we know, from library catalogues or inventories, existed until a certain time in a certain place, but are now lost, either destroyed, or more likely in some library where they have not been identified as referring to that former ownership yet. In the lucky event of the identification of a "Historical copy" with a physical copy, all we have to do is replace "Historical copy" in the "Holding Institution" field, with the name of the library now holding that copy, and provide its shelfmark. At present "Historical copies" from the 9 In its philological sense of comparison with other texts.

libraries of Prospero Podiani of Perugia, of the Benedictines of San Giorgio Maggiore in
Venice, and of Francesco Maria II della Rovere, Duke of Urbino, are being gathered in MEI.

The case study: Perugia's Public Library (Biblioteca Augusta) and its former collection of incunabula
The Biblioteca Augusta of Perugia takes its origin from a private collection formed in the second half of the 16 th century by a citizen of that town, Prospero Podiani. Podiani was born around 1535 to a family of professionals and small landowners from a village in the Perugian countryside, Poggio Aquilone. From there -'de Podio', whence the family name derivesthe notary Paolo di Simone moved to Perugia towards the end of the 15 th century, and started a family that was soon well integrated into the social life of the city. 10 We know very little about Prospero's education, but there is evidence that he deeply valued belonging to a family that had been active in the government of the town and in the university. Personal relationships with scholars from the local university and the newly founded school of Jesuits were also important.
The project of founding a library was conceived by Podiani as a way to turn his passion for books into a job, so as to make the most of his knowledge of the booktrade. The proposal for the library was accepted in 1582, when a long contract was included among the official reports of the city council, to establish the donation and the management of the future public library. 11 Podiani was expected to donate all the books he owned at the time, about 7,500 volumes, as well as all the books that he acquired in the future. Delays occurred in the building of the hall of the library, which was only completed around 1590, when Podiani personally took care of the works. 12 Podiani slowly lost all confidence in the city governors, and consequently refused to deliver the catalogue, and the library with it; he even tried to revoke the donation. Meanwhile, he continued to treat the library as if it was still his own collection. This did not imply that the library was closed; on the contrary, many people had access to the collection, as more than 200 letters and borrowing ledgers testify. The library was actually in use, though not officially open. 13 This situation came to an end in the autumn of 1615: on October 17 the city governors, worried about losing the library, which Podiani had meanwhile moved from the Palazzo pubblico to a monastery, sent an official embassy to visit him so as to have the donation confirmed. One month later, on 16 November 1615, Podiani died.

The Biblioteca Augusta in 1617
A new phase in the history of the Biblioteca Augusta started in June 1617, when Fulvio Mariottelli, a learned and erudite priest, was appointed to rearrange the library, starting with a survey of the books recovered from the monastery and brought back into the Palazzo pubblico. The result of Mariottelli's work was a detailed inventory of the library, which is still preserved in the Biblioteca Augusta as MS 3082. 14 Podiani's name does not appear in the title of the inventory, most probably in revenge for the way the city governors had finally taken possession of the library. Mariottelli, however, who had been on good terms with Podiani, wished to make it clear that the latter was responsible for the creation of that library: as soon as he had delivered the inventory, he published a report on the work he had done and put the name of Podiani in the title. 15 Along with praise of the great value of the 12 Documents testify of the payments that he made in favour of the artists involved in the decoration of the hall: two frescoes representing Parnassus and Christ among the doctors; in both cases Podiani's name appears in the stones in the middle of the frescoes. 13  library as a public good, the report was conceived to explain its arrangement and to give information on the collections as a complement to the inventory.
The inventory consists of unusually detailed descriptions: for each item Mariottelli provided the name of the author, an abridged title, the language, the place and year of printing, and the format. On the other hand, the majority of the books were described before being rearranged on the shelves: the lists of items were therefore determined by the order of the 80 chests that were used to bring the books back to the Palazzo pubblico. Inside the chests, the books were listed according to their format: "Books of the first chest in folio […] in quarto" and so on. 16 Some 7,570 items were listed in 1617 under the supervision of Fulvio Mariottelli, who added his subscription and the official declaration that he took delivery of all the books listed in the document. Three years later ninety items were added consisting of books that had been given back by various people. 17 Altogether the document lists 7,659 items, including descriptions of many items grouped together, such as 260 'pieces of manuscripts' and five chests of unbound books which close the first list (fol. 135r). According to data provided by Mariottelli in his Ragguaglio, the Biblioteca Augusta consisted of about 10,000 volumes, 600 of which were manuscripts.

Before MEI
Given its degree of detail and precision, the manuscript BAP 3082 was chosen as the main source of information on the Biblioteca Augusta. 18 As the starting point for any further indepth analysis, a complete transcription of the document was put into a spreadsheet, so as to have all relevant data in a homogeneous format, as follows:  In view of the scale of the library and of the number of books listed in the inventory, preparatory studies centred on portions of the library, postponing exhaustive identification of all the editions described in the inventory. The attention focused on two aspects: first, how the library was used, that is works, editions, and copies mentioned in letters to and from Podiani, and in borrowing ledgers. Second, the presence of forbidden books as defined in the Roman Indexes published in the 16 th century, that is in 1557, 1559, 1564, and 1596. 20 In both cases, cross-references were arranged between citations of editions found in external sources of information and the descriptions found in the inventory. From a conceptual point of view, the identification of prohibited books was pretty easy, since it relied on the use of the same categories: some or all editions, of some or all works, by a certain author; the identification of copies would follow, when they were still in the library: the citations in the letters and the descriptions in the inventory, and between the descriptions in the inventory and a known edition or, when possible, a copy. One point for the identification of the author; two points for the identification of the author and the work; three points for author, work, and edition; 4 points for author, work, edition, and copy:

After MEI
In MEI the recording of copies is based on a direct examination of the books as well as on external sources: catalogues and bibliographies pertain to the provenance type 'bibliographic evidence'. Moreover, as mentioned before, the system has been conceived in such a way that historical sources, such as the 1617 iinventory of Perugia's library, can be used to record a collection that existed at a certain time.
The system therefore appears to offer the best solution to the challenges that have just been discussed, at least as far as the incunabula are concerned, because MEI only records fifteenthcentury editions. To begin with, all incunabula mentioned in the inventory were described as copies formerly belonging to Podiani and then as part of the foundation collection of the Biblioteca Augusta.
An important result was soon reached: based on the historical source, a set of about 278 copies was identified with certainty as part of the former collection of incunabula of the Biblioteca Augusta. This set was then compared with the result of a census published in 1974, where no more than 60 books were listed. The census was in fact limited to books bearing the ownership note of Podiani (19 copies) or the stamps "August. Perus." and "Prosper. Podian." (41 copies). 23 Of course the 1974 census has been added as another bibliographic reference, since it provides copy-specific information such as further provenance, bindings etc. The identification of copies currently in the library, however, has not been limited to the volumes listed in the census. As a matter of fact, the presence in the Biblioteca Augusta of a copy of a certain edition according to the ISTC record, has been considered as a meaningful pointer towards the identification of an edition of the inventory.
Although at least around 280 descriptions of the inventory seem to refer to incunabula, only copies of editions identified with a degree of certainty have been recordered in MEI, according to the following criteria: 1) "Certain" identification: descriptions that perfectly correspond to known editions: all given elements (author, title, place and year of printing, format) match.
2) "Uncertain" identification: descriptions that do not perfectly match because of one element, usually the format (4° instead of folio or viceversa). This can be considered a minor mistake, since this information in the inventory is not provided in each individual description but at the beginning of a set of books grouped together precisely by format. Moreover, it may be assumed that the author of the description did not work according to current standards.
3) 'Bibliographical merger': descriptions that do not match any known edition because they actually refer to books bound together: in Fig. 3 a scheme has been provided to show a case of a Sammelband perfectly matching an inventory description.
Beyond these three categories, descriptions apparently referring to incunabula (about 25 items) have been discarded. Possibly they might be identified if a direct examination of the copies were to reveal the reason for the variance between the description and the copy in hand. In a case such as the Sammelband shown in Fig. 3, the identification has been facilitated by the preservation of the original arrangement of the copies. This is not always Didier Marcotte has published some articles announcing the discovery of this copy, with important information on Laetus's geographical works. 25 As Marcotte has pointed out, the discovery brings to light what had been thought lost. 26 Further information may possibly emerge concerning the handwriting which filled the margins of the Cosmographia, which has still to be identified, as well as the scribe who copied the manuscript. As soon as new data are found, they will be recorded in MEI, where the history of this particular copy will be finally reconstructed by gathering together information scattered in historical catalogues, bibliographies and scholarly articles of all sorts. In doing so, each piece of information will be confirmed, refuted or clarified by the others.

Conclusion
In the opinion of the authors, the value of the database in supporting research on dispersed collections resides in its elasticity in hosting and accommodating information of various nature, bibliographical, physical, and documentary. And also in capturing and converting into historical evidence different levels of certainty, a sure inscription, but also manuscript annotations or a decoration style assigned palaeographically or art-historically to a certain area and period, an old shelfmark which can be meaningless in itself, but meaningful when set in a comparative environment. Finally, the possibility of continuously adding to a record to reflect the evolving life of the book. Our latest addition to the ongoing reconstruction of the Podiani collection comes from the trade: the electronic catalogue of the London bookseller Peter Harrington displays a copy of the 1499 Etymologicum magnum graecum. 28 Wherever the new location of the book, it is now captured as MEI 00200830. One of the objectives of the 15cBOOKTRADE Project 29 is the creation of visualisation software -already in its test version -which will map the complex history of these books, alone and as part of large groups, historical copies together with physical copies, for a better understanding of our