University of Virginia Library

While many Italian scholars have made substantial contributions in this century to the history of the book, there has only been one great twentieth-century Italian analytical bibliographer, marchese Roberto Ridolfi (1899-1991). Ridolfi was more or less a contemporary of Fredson Bowers (1905-1991) but, though he shared Bowers's versatility and intense concentration on the matter in hand, his personality, interests and life-style were very different, more similar in some ways to the earlier, British, generation of leaders in the development of textual bibliography. Ridolfi came from an ancient Florentine family which had played a prominent part for centuries in Florentine and Italian political and cultural life. Apart from service in the First World War, and a brief period in the 1950s, to which I shall return, when he held an appointment at the University of Florence, Ridolfi was never gainfully employed. He lived the life of a gentleman at the family villa of La Baronta, on the hills to the south of Florence. But, like Greg, who likewise had sufficient private means to make it unnecessary for him to seek employment, Ridolfi filled his life with scholarly endeavour. Florence, its history and its culture were central to, indeed, in some senses, his only interests. In this he was typical of many Italians, whose allegiance to Italy is often less strong, or less apparent, than their attachment to the history, culture and well-being of their own ancient locality.

Ridolfi's only formal experience of higher education was a course in chemistry at the University of Pisa, followed at the insistence of his father, who hoped to exploit some thermal springs on the family's land. This course probably helped Ridolfi more than an Italian university education in letters might have done to appreciate the high level of skill necessary for good quality craftsmanship, and to understand the techniques of type-casting and paper-making. He had, anyway, quite a practical cast of mind. But his vocation was humanistic, not scientific. In an

autobiographical volume Ridolfi recounts how as a boy he used to wander in the grounds of La Baronta and on the hillside beyond, reciting aloud verses from his favorite poets.[1] His first passion was for lepidoptery. When this faded with the years (and with the disintegration through neglect and the passage of time of his collection of butterflies and other insects), his interests moved to history. Florentine history, naturally, and specifically, family history, since his first subject of study was the figure of a famous Renaissance ancestor, cardinal Niccoló Ridolfi (1501-1550), diplomat, bibliophile and patron of letters. Even though he never published the biography of Niccoló "in nine chapters, with an appendix of documents," which he wrote after three years of work,[2] the research he undertook for this project was crucial to his future career. It revealed to him the riches which were to be found in the largely uncatalogued and often disastrously housed private archives of the ancient Florentine families, to which his family connections gave him unique access. As he later wrote of those years:

Ridolfi's first major contribution to scholarship, which occupied him during the second half of the twenties and the early thirties, was as a describer and cataloguer of some of these archives. A proposal for the recording and conservation of private Italian archives, published in

1928, attracted the attention of Mussolini, and was even the subject of a leading article in The Times.[4] The major publication of this phase of his career, a volume of guides and inventories to the private archives of Florence, appeared in 1934, and Ridolfi had assembled the material for three further volumes. But political events (the publishers, the Florentine house of Olschki, were Jewish friends, and Ridolfi was always courageously outspoken in their defence, making no secret of his opposition to Fascism) prevented the continuation of the series, and the material for the remaining volumes later became a victim of the brief occupation of La Baronta by German paratroopers in 1944.

But Ridolfi's work in the private Florentine archives had already set him on the next stage of his career as an historian, which, as he himself notes, lasted more or less for the rest of his active life. Among the riches he had discovered by chance in the course of his work were important new materials regarding two of the greatest figures of Renaissance Florence. One was the Ferrarese Girolamo Savonarola, prior of the Dominican friary of San Marco from 1491 until his death at the stake in 1498, whose tempestuous career as preacher and prophet in the last two decades of the fifteenth century profoundly affected the political and cultural life of the city. The other was the politician and writer Francesco Guicciardini, formulator and executor for many years of Papal policy in Italy, and one of the creators of modern historiography. The next phase of Ridolfi's activity as an historian was devoted to illustrating the lives, and publishing or republishing the works, of these great figures, to whom he later added the third great Florentine of the period, Niccoló Machiavelli. Of the numerous learned publications to which this historical activity gave rise special mention should be made here of Ridolfi's biographies of this trio, published between 1952 and 1960 in a remarkable burst of activity, given Ridolfi's other commitments at the time. These were translated into English and other languages and brought the fruits of his scholarship before an international reading public.

I have sketched Ridolfi's career as an historian, and in so doing anticipated somewhat the passage of time, because it is significant for an understanding of his activities as a bibliographer. It was his work on Savonarola that first made him enter the domain of analytical bibliography. The friar's fiery preaching and his involvement in the return to a more popular form of republican government after the expulsion of the Medici in 1494 gave rise to a flood of printing in Florence, much of it, including editions of single sermons produced soon after their delivery,

undated and even unsigned. To bring some order into this chaos required the careful and accurate analysis of surviving copies. As had been the case with his archival and historical work, Ridolfi received no training as a bibliographer, but was entirely self-taught. He relied, I think, on the lessons to be learned from the Catalogue of Books Printed in the XVth Century Now in the British Museum (1908- ) and from Konrad Haebler's Handbuch der Inkunabelkunde (1925). But more than that, he drew also on his previous experiences as an historian. This is Ridolfi's great strength, and modernity, as a bibliographer. He was never worried by doubts about whether analytical bibliography was a science or an art, never bothered about its place in liberal studies, or its function as a "handmaid" of literature, to use Greg's term.[5] To him, a book, in its materiality, was simply an historical document, and analytical bibliography a branch of historical research. Ridolfi's commitment as an historian also helps us to understand the occasional nature of his bibliographical activities. His major contributions to analytical bibliography were concentrated into a single decade, more or less coinciding with the 1950s. For him analytical bibliography was a research tool to be used when necessary, not a lifelong commitment.

Bibliographical problems, as I have said, first entered Ridolfi's ken when his interest was aroused in the activities and works of Savonarola in the late 1920s and early 1930s. They first surface in his splendid edition of Savonarola's letters (1933) and in his contributions to the first volume of the bibliography of Savonarola's works by his uncle, prince Piero Ginori Conti, published in 1939 (another work never completed, because of the war). But the real stimulus came from his decision in 1944, immediately after the end of hostilities in Tuscany, to accept the invitation of Aldo Olschki, son of Leo S. Olschki, to become editor of the leading Italian bibliographical journal, "La Bibliofilia," founded by Leo in 1899, and published ever since by the family publishing house. His editorship of this journal, which was to last for thirty-eight years, exposed Ridolfi to frequent and regular contact with the world of Italian and European bibliographical scholarship, and encouraged him to think again about the bibliographical problems which he had encountered in his work on Savonarola. As he wrote in a characteristic passage of the Memorie:

The first fruits of this period of study and reflection on bibliographical matters, the article "Proposta di ricerche sulle stampe e sugli stampatori del Quattrocento" ("A Proposal for Research on Fifteenth-Century Printing and Printers"), appeared in La Bibliofilia in 1949.

As the title of this article reminds us, Ridolfi's interests as a bibliographer centred on incunables, and on Florentine incunables at that. This is simply the result of his own experiences as an historian, and emphasizes the fact that, however strong his interest in the book, both manuscript and printed, his bibliographical studies were contingent on his historical interests.[7] A further reflection of his personal experiences can be seen in the emphasis placed in this study on the attribution and dating of unsigned editions. The attribution and dating of unsigned incunables constitute, in fact, the context of all Ridolfi's contributions to bibliographical method. But it would be incorrect to say that they also define their limitations, because his methodological proposals are often valid for other periods of hand printing.

Within this context, the purpose of the "Proposta" is to stress the need for a multidisciplinary approach. First of all, he comments, those wishing to date incunables ignore at their peril the evidence offered by the contents of the works in question, which can sometimes provide unequivocal termini a quo or ad quem. This comment may not seem very original to those coming to bibliography from the study of literature,

but for bibliographers with other backgrounds its advice is timely.[8]

Another line of enquiry for the bibliographical scholar suggested in the "Proposta" is archival research, particularly in legal archives. Here we see the experience of the great authority on the private archives of Florence. But the comment also reveals the broad view of the historian, who conceives of bibliographical studies not as a series of watertight compartments, but as a continuum involving various activities, which together provide the material for the history of printing, itself a part of the history of the book.

The one specifically bibliographical proposal contained in this article is for a study of the phenomenon of offset ("contrastampa" in Italian). The brief reference to offset in the "Proposta" provoked so much interest that Ridolfi was moved to enlarge his comments in the form of an article wholly on the subject, which appeared in the next number of La Bibliofilia. This article constitutes the only attempt to have appeared in Italian, and indeed, as far as I am aware, in any language, to describe the nature of offset and the uses to which it can be put in bibliographical analysis.[9] The article's subtitle, "Nuovi sussidi per l'attribuzione e la datazione dei paleotipi" ("New Aids for the Attribution and Dating of Incunables") reiterates the context of all Ridolfi's bibliographical work. Because of his concentration on attribution and dating, Ridolfi focuses attention on those cases in which the offset comes from a different edition, dismissing as of little bibliographical interest cases where it

comes from another page of the same edition. In this judgment, however, he is unduly influenced by his special interests. Offset from pages of the same edition can sometimes offer important evidence of such things as the imposition of preliminary matter, the existence of otherwise unrecorded cancels, and the order of sheets through the press. And, in a footnote to his article, Ridolfi himself uses "internal" offset to excellent purpose, as we shall see.

Ridolfi distinguishes two degrees of offset. The first, "that which more properly deserves the name, occurs when a sheet which has just come off the press, or on which the ink is still fresh, comes into contact with another and is offset on it, even if little or no pressure is exerted," the second, "when a page on which ink is no longer wet, but on the other hand not yet completely dry, leaves an impression on another page with which it is in prolonged and continuous contact, as for example when the sheets are bound together in a volume."[10] Instances of the first degree of offset, Ridolfi continues, are those of most use to bibliographers, since they provide evidence of the simultaneous, or almost simultaneous, printing of the pages concerned; moreover, since the ink is still fresh, the offset impression is usually quite easy to read. Examples of the second degree of offset are much less compelling evidence, because of the longer time-lag between printing and offsetting, which Ridolfi estimates as anything between 48 hours and a month; in any case, they are frequently illegible. The latter is the really telling comment here. "Slow" or "tardy" offset, to use Ridolfi's epithets ("lenta," "tardiva"), provided its source can be identified, can still furnish useful chronological and attributional evidence.

In the second half of the article Ridolfi discusses three cases, two of them illustrated with photographs, in which light is thrown on undated and/or unattributed Florentine incunables by the presence of offset, two being instances of his first degree of offset, and one of "slow" or "tardy" offset. A general point emerges from these cases, one which has been mentioned by Ridolfi earlier in his text. In the small, popular editions of few gatherings in quarto format, which form the subject of these examples, offset from another edition occurs exclusively on the first and last pages of a sheet. Ridolfi reasonably concludes that these sheets were stored already folded. In a long footnote (p. 137, n. 2),

devoted to showing that some useful evidence can be offered by "internal" offset, deriving from sheets of the same edition, Ridolfi shows conclusively, from the sequence of offsets on the first and last pages of each gathering, that folded sheets of one rather different edition, the Prediche di Frate Hieronymo da Ferrara [Florence: Bartolomeo dei Libri, Lorenzo Morgiani, Francesco Buonaccorsi], ad istanza di ser Lorenzo Vivoli, 8 Feb. 1496/7, a substantial Chancery folio gathered mainly in eights and containing Savonarola's Lenten sermons for 1496, were stored already assembled into copies. This detail may be related to the fact that the edition in question was shared by three printers; nonetheless, it illustrates the sort of unprejudiced reasoning Ridolfi brought to the interpretation of bibliographical evidence.[11]

We come now to the period of Ridolfi's other major contribution to Italian bibliographical studies. This was occasioned by the tenure of a post at the University of Florence, where he taught "bibliologia" (analytical bibliography) from 1953 to 1957. This appointment provided further stimulus for his reflections on the methodology of bibliographical research. During the years 1954 to 1956 some of the material which he had prepared for his lecture courses was published, suitably adapted, in a series of articles in La Bibliofilia, under the general title of "Nuovi contributi alla storia della stampa nel secolo XV" ("New Contributions to the History of Printing in the Fifteenth Century"). Naturally, the printing in question was Florentine; the articles were later gathered and published, with some additional material, in the volume La stampa in Firenze nel secolo XV ("Printing in Florence in the Fifteenth Century"). These articles are notable for the gradual development of what was to become fully fledged only in his little book of 1957, Le filigrane dei paleotipi ("The Watermarks of Incunables"), namely, an understanding of the way in which the dating and attribution of incunables can be assisted by paper study. In the "Proposta" of 1949 there is no reference at all to watermarks. Obviously, when writing that article Ridolfi had not yet realised their potential in bibliographical analysis. This was partly due, as he himself explains in Le filigrane dei paleotipi (pp. 11-12), to

the influence of Konrad Haebler, who in his Handbuch der Inkunabelkunde had expressed substantial scepticism about their usefulness as bibliographical evidence.[12] It is typical of Ridolfi's confident and independent approach to bibliographical analysis that he soon realised this scepticism was unfounded, or at least not justified in the particular cases he was considering. For the vast majority of Florentine incunables had been printed on paper made in the mills at the small centre of Colle di Valdelsa, about 25 miles south of Florence, and something was already known about paper manufactured at Colle; indeed, in the 1950s, when Ridolfi was working on watermarks, some of the old mill structures still existed. Ridolfi did not, like Allan Stevenson at Basel in 1961, climb into the drying lofts of these old mills and commune with the pigeons, but he did use the studies already published, imperfect though they were, on the history of the paper industry at Colle and on the watermarks they contained, and this information, powerfully supplemented by his own observations, enabled him to forge what for Italian bibliographers, and for almost every other bibliographical scholar living at that time, was a new instrument of research.[13]

Ridolfi's developing understanding of the bibliographical use of paper study can be followed in two of the "Nuovi contributi." The first, "Lo Stampatore del Mesue e l'introduzione della stampa in Firenze" ("The Printer of Mesue and the Introduction of Printing in Florence"), was devoted to a theme dear to Ridolfi's Florentine heart, namely that "printing was not brought to Florence . . . by the `printer of Mesue,' one of those many foreign printers who introduced the art to almost every other Italian city, but had in reality been `re-invented' by a Florentine goldsmith, Bernardo Cennini."[14] Among the evidence used to demonstrate this thesis, watermarks did not provide the most significant nor,

to tell the truth, the most convincing proofs, but we can already see in his comments on the watermark evidence the first steps along the path of understanding which would take Ridolfi, with his usual rapidity, in a couple of years to a more or less complete grasp of the possibilities of paper study for bibliographical research.

In the first place, Ridolfi, while fully acknowledging the importance of Briquet's magisterial collection of watermark tracings, quickly realised its inadequacies for bibliographical work. He also fully understood the reasons: Briquet's vast repertory was based overwhelmingly on the watermarks found in archival documents, not books; and his use of tracings made the representations of watermarks in his manual too approximate. Ridolfi from the very first used only photographs as illustrations of the watermarks he was discussing. Unfortunately, he said nothing at all in any of his relevant works about the methods used to obtain his sometimes excellent and always serviceable results.[15] To return to his article on the printer of Mesue, what is even more interesting is to see, in a long note on p. 8 of the study, the beginnings of an explicit understanding of what is only implied in his criticism of Briquet's tracings as too approximate, namely, the presence and significance of small differences in watermarks, an understanding which is the beginning of wisdom in paper research. In this note Ridolfi writes:

Ridolfi was obviously very close, when he wrote these words, to realising that watermarks are twins; indeed, when he came to prepare this article for inclusion in the 1958 volume La stampa in Firenze nel secolo XV, he re-wrote the note to give explicit expression to the concept (without, there or anywhere else, resorting to the metaphor made famous by Allan Stevenson in his article of 1951).[17]

A further step forward is evident in the third of the "Nuovi contributi," "G.W.3851: Antonio Miscomini, non Compagnia del Drago," published in La Bibliofilia in 1955. With this article we are back on familiar territory, the dating of incunables sine notis. The subject of this third "contributo" is the Novella di Tancredi principe di Salerno by the Florentine Girolamo Benivieni (1453-1542), attributed in the Gesamtkatalog der Wiegendrucke (III, p. 630) to the Compagnia del Drago at Florence, and dated c. 1497/8. For the identification of the true printer, Antonio Miscomini, Ridolfi relied on typographical evidence; for a preliminary dating, which allowed him to place publication of the edition in the 1480s, he turned to considerations of content. But for a precise dating it was on paper evidence that he relied:

Those readers interested in the workings of coincidence will not fail to have noticed that Ridolfi must have been looking at the Colle di Valdelsa crossed hammers in Florentine incunables, and realising the significance

of their "slight changes" for his research, at more or less the same time that Allan Stevenson, in the Pierpont Morgan and other American libraries, was becoming familiar with the ebb and flow of the Basel bull's head and cross-on-mounts watermarks, which were to form the basis of his dazzling demonstration of the date and place of printing of the Missale speciale.[19] While postponing to a later stage a comparison between the two scholars, let me assuage curiosity by saying now that there is no evidence of contact between them; indeed, it seems that in their activity as paper scholars they were totally unaware of each other's existence.[20]

To return to the passage quoted above, it is clear that in consequence of his observations of watermarks in Florentine incunables, Ridolfi had succeeded in imposing a certain order on what seemed, when he was working on the first "contributo," a confusing succession of marks having more or less the same motif. This step forward resulted from his having carried out "systematic studies" of the Tuscan watermarks in Florentine incunables, as a result of which he had been able to form his own watermark catalogue, limited in scope to local marks, but much more relevant than anything he could get from Briquet. With this material, he had come to realise that when a couple of moulds became worn out with use, the papermakers of Colle would replace them with another pair in which the same watermark motif was used, but sufficiently differentiated to enable the paper made with this particular pair of moulds to be identified. Thus, the proliferation of similar but not identical marks, which had led Haebler and others to reject the study

of watermarks as a tool for bibliographical research, turned out to be the very feature which made them useful.

Ridolfi's mature thinking on watermark evidence is contained in the little booklet of 1957, Le filigrane dei paleotipi. Here he gathered the more or less scattered observations of the "Nuovi contributi" into a coherent exposition. He begins by stating that what makes watermark evidence a valid tool in the dating of unsigned editions is the fact that incunable printers did not keep large stocks of paper, but bought (or were given by those commissioning their work) the quantity of paper necessary for the printing of each book, or group of books. This understanding is a fundamental step forward. As usual, Ridolfi gives no indication of the source of his insight; we must assume that it is the result of his own observations, both of the paper supply in the books themselves (they were, after all, his main teachers, as he himself says), and of the few documentary sources available to him, such as the diary of the printing establishment in the Dominican convent of S. Jacopo at Ripoli, on the outskirts of Florence.[21] As he puts it:

From these considerations Ridolfi deduces the following principle: "one can be absolutely sure that the paper used in an edition is more or less contemporary with the date printed at the end of that edition."[23]

Here, however, in attempting to distance himself from the calculations of Briquet, which are based on the far slower consumption of paper in the manuscript period, Ridolfi has exaggerated, or better, has drawn too sweeping a conclusion from the material with which he was familiar, that is, mainly popular editions comprising few sheets, and produced by local printers. His principle is frequently violated, most obviously in the case of editions whose printing for some reason or another is protracted over a period of years, and for which some of the paper has been bought before printing began. In addition, possibly because many of the editions he was dealing with were relatively small, Ridolfi makes no explicit distinction between runs of paper with the same watermark, which are likely to belong to a single purchase of paper, and single sheets with other watermarks, which are likely to be the remains of purchases made on other occasions, and thus not necessarily "more or less contemporary with the date printed at the end of that edition."[24] Indeed, his principle has to be applied with care even in the cases he had in mind; in this context, it would be interesting to re-examine the specific instances of dating by paper evidence brought forward in his publications of the 1950s, to see if they stand up to a more sophisticated analysis. It is perhaps significant that Stevenson, who probably knew more than Ridolfi about the practices of the Renaissance paper industry, was more cautious in his claims regarding the contemporaneity of paper manufacture and paper use, declaring that stocks of paper used in printing were "ordinarily . . . secured within one, two, or three years after manufacture. . . ."[25] Nonetheless, Ridolfi's principle shows that he was fully aware of the main fact of life concerning paper supply in early printing, namely, that printers and publishers normally could not afford to lay in stocks of paper against potential future consumption, but bought or were supplied with paper for specific jobs, as and when required.

It is natural, given Ridolfi's interests, that he should immediately draw a corollary from his principle, which is "that an edition sine anno is to be assigned to the same period as a dated edition in which an identical watermark can be observed."[26] This leads him to the central point of his exposition, an explanation of the reasons for, and the modes of, the similarities and differences between watermarks of the same motif and design:

This is Ridolfi's most mature expression of his understanding of watermark diversity. He describes three sources of differentiation in watermarks having the same motif: 1) watermarks are twins, because the moulds used in their manufacture are twins, but they are dissimilar, not identical, ones; 2) a watermark with the same motif can be used in the moulds of another vat, but it will be distinguished "by some detail of the design or by its size;" 3) when worn-out moulds are replaced, the mark will be repeated, "but with some difference in design or size, or with the addition of some particular design detail." There is also a fourth source of differentiation, previously mentioned by Ridolfi, and it is far from being the least important for his method; this is "the deformations produced by usage in the delicate design of wire thread," from which it follows that "even the images produced by the same mould vary in the course of time in ways which are more or less visible."[28] In other words, a watermark decays in the course of its relatively short life, it presents itself to the bibliographer in a series of successive states, which offer him another element capable of establishing a chronological succession. This sequence of states of the same watermark is all the more precious in that in practice it is often difficult to distinguish between Ridolfi's second and third type of differentiation. Indeed, it can sometimes be difficult to distinguish, on the basis of design alone, between all three types of differentiation, as twins can differ quite substantially in size. The fact is that the only sure way of identifying twins is to find them in a substantial run of paper.

Ridolfi has now arrived at his conclusion:

Quod erat demonstrandum. Ridolfi concludes the first part of his booklet, which constitutes undoubtedly the most acute contribution to analytical bibliography made by an Italian scholar, with a proposal, speedily dubbed utopistic by colleagues (as indeed it was, at the time), for the compilation of a general catalogue of incunable watermarks, preferably photographic, organized by place of printing, and based on the examination, not just of every edition, but of all the accessible copies of every edition. The second part of Le filigrane comprises a discussion of one particular mark, a Greek cross in a circle, which was used in a mill at Colle di Valdelsa for the last twenty-five years of the fifteenth century (the dates come from the inspection by Ridolfi of printed editions, but the identification of the place of origin of the mark is derived, characteristically, from a document). Ridolfi charts the successive differences in the mark, of which he claims to have identified about fifty versions, deriving from at least twenty-four pairs of moulds, and adds: "Among these versions, those observed in dated editions, or in editions which can be confidently dated, allow us to date all the editions sine anno in which the marks occur."[30] The booklet ends with a "little catalogue" of editions containing this mark, followed by a series of twenty-three photographs, illustrating thirty-eight marks.

The name of Allan Stevenson has already occurred several times in these pages. It is clear that in their work on paper research he and Ridolfi arrived more or less simultaneously at exactly the same conclusions about the usefulness of watermark evidence for dating editions sine anno; yet each seems to have been completely unaware of the other's existence.[31] The different life-styles and formae mentis of the renowned Florentine aristocrat and scholar, and the more-or-less failed American college professor, hit the eye.[32] When one comes to compare their works,

however, the advantage clearly lies with the less obviously successful and gifted Stevenson. For Ridolfi, analytical bibliography was one among many scholarly interests, pursued with great flair, but also with an aristocratic sprezzatura, straight out of Castiglione's Book of the Courtier, which led him to be brief in his exposition, and reticent about the techniques and methods employed. It is also true that in Ridolfi's work these techniques and methods, including paper research, were means to an end, not ends in themselves. Perhaps in different circumstances he would have developed further his brilliant deductions about the way offset and watermark evidence can assist the bibliographical scholar, and would have given the same penetrating attention to other aspects of bibliographical analysis; but, as we shall see, fate conspired to place obstacles in his way, and after his brief experience as a teacher of analytical bibliography, he turned back to his historical studies. His edition of Machiavelli's Mandragola, published in 1965, was based on a recently discovered manuscript, not on the printed editions; Ridolfi's discussions of these editions, and the other bibliographical articles he published in the 1960s, contain no new insights into the methodology of bibliographical research. Stevenson was more limited in his scholarly interests than Ridolfi, but also more single-minded. He devoted all his intellectual energies to paper research, and wanted nothing more than to pass on his knowledge to his readers. The point comes over forcefully when we compare the 400 pages of The Problem of the Missale Speciale with the 45 pages of Ridolfi's little work. Stevenson is incomparably more informative than Ridolfi, both on paper research and on paper production. In The Problem of the Missale Speciale, as well as a long discussion of the likely time-span between the date of paper manufacture and of its use in printed books, and a description of how watermarks are made, we have a series of practical and useful pieces of advice for the bibliographer, about such things as the vital distinction between runs and remnants, the existence of sewing dots (those invaluable aids to distinguishing incunable watermarks ), the practice of placing twin watermarks in opposite halves of the mould (certainly a feature of Venetian paper in the Renaissance, as Stevenson knew it to be of Basel paper), and the gradual sideways movement of marks in the mould, if not attached to a chain- line. None of this information can be obtained from Ridolfi's contributions, which, for all their brilliance, are disappointingly silent, perhaps deliberately, about techniques and methods. Stevenson, on the other hand, does his best to tell all. For anyone wanting to learn how to begin and how to prosper as a paper scholar, Stevenson is the man.

No-one has yet written the history of Ridolfi's course at the

University of Florence. Even the dates are uncertain. According to Maria Jole Minicucci, it ran from 1953 to 1957, but one who had been a student in the first year of the course recently put its start as 1952. It was never well attended. Indeed, in one year there was only a single student.[33] Perhaps this was the reason, or one of the reasons, why in 1957 the post held by Ridolfi was abolished. With it went the Centro per lo Studio dei Paleotipi ("Centre for the Study of Incunables"), which Ridolfi had set up in the university, and of which his booklet, Le filigrane dei paleotipi, was the first and only publication. A few years later, in 1963, Ridolfi re-founded the Centre in the Biblioteca Nazionale, Florence, with Maria Jole Minicucci, a member of the library staff, as collaborator, but its activities were cut short, this time for good, by the disastrous flood of November 1966, which seriously affected the library buildings. The only one of its projects to survive this final setback was a typewritten index, compiled by Minicucci, to the 2460 illustrations of type published between 1907 and 1939 by the Gesellschaft für Typenkunde des XV Jahrhunderts in parallel with Konrad Haebler's Typenrepertorium der Wiegendrucke (Halle 1905-24). This index, which integrates and corrects the previous partial indexes of Madsen and Juchhoff, has been deposited with Ridolfi's library in premises belonging to the Cassa di Risparmio di Firenze.[34]

The termination of his appointment at the University of Florence and the subsequent vicissitudes of the Centre for the Study of Incunables more or less put an end to Ridolfi's activities as an analytical bibliographer. If man had not done so, nature would soon have stepped in, because as he advanced in years, Ridolfi's eyesight began to fail; by the time of his death he was completely blind. In his lifetime his remarkable if limited contribution to analytical bibliography was little understood in Italy, where until recently this particular branch of the discipline of book history has not been much cultivated.[35] The originality of

Le filigrane dei paleotipi was not appreciated, either in Italy or elsewhere. The book did receive a friendly review in The Library from Victor Scholderer, the Italian specialist of the British Museum Library and a friend of Ridolfi's, but it is typical of the imperfect appreciation at that time, in all but a very few bibliographers, of the importance of paper research that Scholderer calls attention, not so much to the first, methodological, part of Le filigrane, but to the discussion of the crossed hammers mark from Colle in the second part. I suspect this is why Ridolfi's book is listed in the Bibliography of the Stevenson edition of Briquet's Les filigranes (1968), not in one of the sections "Evidential value of watermarks" or "Methodology," but in the geographical section "Watermarks in Early Printed Books—Italy," with the comment: "Mainly concerned with Colle wms."[36] A proper appreciation of Ridolfi's stature has come only in the 1990s, with the maturing of a new interest in analytical bibliography among Italian book historians and textual critics, coinciding as it has done with the reassessment, occasioned by his death in 1991, of Ridolfi's multifarious contributions to Italian culture.[37] In addition to the specific lessons contained in his bibliographical studies outlined above, present and future generations of Italian bibliographers can learn from Ridolfi's activities as an historian a more general lesson. Analytical bibliography is part of an historical discipline, and while it must begin with the consideration of the printed book as a material object, its results have little value unless immersed in the context of the cultural, economic, political and social forces at work in the production of printed books.