Some Observations on the 1532 Edition
of Ludovico Ariosto's Orlando Furioso
by
Conor Fahy
The 1532 edition of Ariosto's Orlando furioso was the
last
edition of this great work to be printed in the author's lifetime. It has
always been regarded as of prime importance for the text of the poem,
because it was printed under the author's eyes at Ferrara, where he lived,
and, compared with earlier editions, contained numerous variants, as well
as extensive new material. It thus comprised a new version of the work,
intended to supersede that published for the first time in 1516, also at
Ferrara, and reprinted there with minor changes in 1521, and then in a host
of reprints, apparently unauthorized, which appeared in Milan, Venice and
Florence between 1524 and 1531.[1]
Only in the last half-century, under the influence of modern textual and
linguistic criticism, and of structuralist and post-structuralist critical theory,
which sees each version of a work as a discrete entity, of equal interest to,
and indeed value as, any other version,
have scholars begun to take an interest in the 1516 edition of the
Furioso and its text. This modern scholarly interest apart,
once
the 1532 edition, with its numerous stylistic and linguistic changes and its
six additional cantos, had appeared, it immediately became and has
remained the basis of the reading text of the work.
During these four hundred and fifty years of critical interest, an
awareness has grown among Italian scholars that copies of the 1532 edition
of the Furioso differ from each other in ways significant
enough
to affect the work of a critical editor. However, it was not until the critical
edition of the Furioso published by Santorre Debenedetti in
1928 that a satisfactory account was given of these variations, and their
nature explained. Though trained in a scholarly tradition almost entirely
preoccupied with problems of manuscript transmission, and working before
the appearance of the first manual of textual bibliography in any language,
Debenedetti accurately identified the two sources of internal variants in the
1532 Furioso—authorial press-corrections and a
cancelled
sheet—and, hampered though he was by the
centrifugal organization of the Italian library system and the inadequate
cataloguing of sixteenth-century books in Italian libraries, and, of course,
by working in the pre-microfilm era, he nonetheless succeeded in locating
and collating all but two of the copies of the edition accessible to him in
Italy at that time. Having re-traversed much of the ground covered by
Debenedetti and become intimately acquainted with his editorial work, I
have no hesitation in describing his edition as one of the masterpieces of
modern Italian textual criticism. The editor's learning, intelligence,
sensitivity and integrity shine from every page. Inevitably, however, his
work has its shortcomings from the bibliographical point of view. In the
first place, he did not understand the structure of the book, a quarto in
eights; more fundamentally, though less surprisingly, he did not arrive at
the bibliographical concept of
state, but considered each
press-variant in bibliographical isolation from
all others, and without reference to its bibliographical context; finally, he
left uncollated more than half the surviving copies of the 1532
Furioso.
With these considerations in mind, I began a couple of years ago to
re-examine the 1532 Furioso. The results, though still
incomplete, have been extensive and, in some ways, surprising. A complete
statement of my findings will be published in Italy in due course in book
form; meanwhile, the present paper discusses certain points which seem to
me to be of general interest.[2]
My discussion begins with a general proposition concerning the
Furioso, that the first edition of 1516 and the definitive
edition
of 1532 were both financed by the author. The evidence for this is
circumstantial, as no contracts between Ariosto and the two printers in
question, Giovanni Mazocco for the 1516 edition, and Francesco Rosso for
the 1532 edition, have yet been found, but none the less its cumulative
effect is convincing. It was first presented more than fifty years ago by
Michele Catalano, author of a richly documented biography of the poet,
whose arguments in this context, though occasionally unacceptable on points
of detail, are generally valid.[3] We
begin with a letter from Ariosto's master and the dedicatee of the
Furioso, Cardinal Ippolito d'Este, brother of the Duke of
Ferrara, addressed to the Marquis of Mantua, dated 17 September 1515,
requesting permission for the transport across Mantuan territory free of tax
from
Salò, on the shores of lake Garda, of 1000 reams of paper, "as I
am
about to arrange for the printing of a book
by my servant M. Ludovico Ariosto".
[4] There is no doubt that this is a
reference
to the forthcoming first edition of the
Furioso. The letter,
which
is preserved in the Mantuan archives, is in Ariosto's handwriting, and I
agree with Catalano and others that the Cardinal had probably provided
Ariosto with no more than a signature to help him obtain exemption from
customs dues for his paper. Even if he also furnished financial support, it
is clear from other epistolary evidence that he left detailed arrangements for
the production of the edition to his "servant". It was the author, for
example, and not the Cardinal, nor the printer, who arranged for privileges
to be obtained from the Pope, the Doge of Venice and others.
[5] On 7 May, 1516, just over a
fortnight
after the completion of the printing of the first edition of the
Furioso,
[6] a Mantuan
courtier, Ippolito Calandra, wrote to
the son and heir of the Marquis of that city, who was in France, telling him
that two days previously Ariosto had arrived from Ferrara with a case of
books containing his new work, the
Orlando furioso. He had
presented copies to the young man's father, mother and uncle; the rest he
was proposing to sell. Calandra added that when the poet put the copies up
for sale, he would buy one, get it bound and send it to France for his young
master.
[7] Whether Ariosto sold the
copies himself, or entrusted their sale to a local bookseller, it is clear that
they belonged to him. This is confirmed by another letter, this time from
Ariosto himself, written to Mario Equicola, secretary to the Marquis of
Mantua and a fellow author, dated 8 November 1520. In this letter the poet
acknowledged the receipt of a sum of money resulting from the sale of
copies of the
Furioso in Verona (with the comment, which
authors of all periods will sympathise with, that it
was less than he expected), and asked his friend to arrange for the Verona
bookseller to return to him any unsold copies, because none seemed
available anywhere else in Italy, and he could not satisfy the daily requests
he received.
[8]
In the light of this letter, we can understand why there was a new
edition of the work in 1521, though so far no similar evidence has come to
light of the author's direct involvement in its publication.[9] With the 1532 edition, however,
we are
back on familiar territory. We have letters to the Doge of Venice and to the
Ferrarese ambassador at the Milanese court requesting privileges for the
new version of the poem, and a letter to the Duke of Mantua (the young
man for whom Calandra had bought a copy of the 1516 edition, who had
succeeded his father in 1519, and been elevated to the rank of Duke by the
emperor Charles V in 1530) asking for free transport across Mantuan
territory of more paper from Salò. This time the letter is written by
the
poet in his own name, and the amount of paper specified is 400 reams,
which, if it is the quantity actually bought and used, would give a print-run
for the 1532 Furioso of about 3000 copies.[10] The colophon of the 1532 edition
is dated 1 October. Within three months Ariosto was seriously ill. He never
recovered his health, and died on 6 July 1533. It is not surprising, then,
that this time we have no evidence of his having been personally involved
in the sale of his work. But we have ample evidence that this edition, too,
was his property: the most striking, and the best known, is a letter from his
brother Galasso written a few days after the poet's death, in which he
complains that because of the poet's illness and death three quarters of the
edition still remains in the hands of his heirs, unsold.[11]
I have dwelt on the evidence for the personal involvement of Ariosto
in the Ferrarese editions of his masterpiece because it undoubtedly helps to
explain the interesting features of the 1532 edition, as they emerge from the
interpretation of the evidence provided by the collation of the surviving
copies. The three features which I wish to discuss in this paper are: the
press-corrections; the cancelled sheet; and the existence of perfect copies
of the edition, consisting only of sheets in the corrected state.
The press-corrections
In the Nota at the end of the third volume of
Debenedetti's 1928 edition of the Furioso the editor lists and
discusses thirty-seven variants which have emerged from a collation of the
eleven copies examined.[12] This list
only contains variants relevant to the constitution of the text. In
Debenedetti's working material, there is an unpublished list of thirty other
variants, many of which are not relevant to the constitution of the text, e.g.
they concern the correction (or the inadvertent creation) of printing errors.
At the time of writing, I have collated ten of the eleven copies utilized by
Debenedetti (the present whereabouts of copy no. 4 in the list in Appendix
A is unknown) and the copies numbered 12 to 23 in the same list. The
number of press variants of which I currently have note is 282. Of the new
variants some fifty per cent are of Debenedetti's second type, those not
relevant to the constitution of the text; the rest are
stylistic or contenutistical variants which, like the thirty-seven variants in
Debenedetti's Nota, cannot reasonably be attributed to anyone
other than the author. The extent of the phenomenon can perhaps best be
shown by relating it to the structure of the volume. As has been mentioned,
the 1532 Furioso is a quarto in eights, with thirty-one
gatherings, and so sixty-two sheets. My 282 variants are regularly
distributed from beginning to end of the volume, and occur in ninety-one
of the 124 formes needed to print it. This figure is the more impressive if
one bears in mind that the sample on which I have been working is
small—less than one per cent of the print-run, if the figure of 3000
is
correct—and if I confess frankly that my collations have been
selective
rather than comprehensive.[13]
It is reasonable to conclude that in the printing of this edition
press-correction was normal, and that the collation of a bigger sample
would lead to the discovery of press-variants in many, perhaps all, of the
thirty-three formes at present without variants. For this, however, we must
probably await the discovery of the whereabouts of further paper copies of
the edition.
Only seven of the sixty-two sheets which comprise the volume have
no press-variants. Of the fifty-five sheets in which variants have so far been
found thirty-six have variant states of both formes. Thus, it was certainly
not the norm in the printing of this volume for the second forme of a sheet
to receive its final corrections while the first forme was on the press.
Presumably, a final proof of the second forme was not normally taken until
perfecting had actually begun. But there is another aspect of the
press-correcting which deserves comment. In seventeen of the thirty-six
sheets with variant states in both formes, the uncorrected states of both
formes (and consequently also the corrected states) only occur together, that
is, in the same copy or copies. The number of copies involved is, in one
case, one copy; in five cases, two copies; in four cases, three copies; in one
case, four copies; in two cases, five copies; in three cases, six copies; and
in one case, seven copies.
Twelve of the sheets involved belong to the second half of the volume. The
phenomenon assumes even greater proportions if we include "near-miss"
cases, in which there is a difference of only one or two copies in the group
of those with uncorrected (or corrected) states of both formes. This gives
a further twelve sheets, again concentrated in the second half of the volume.
Indeed, after gathering N there is only one sheet where there is not some
close connection between the list of copies with uncorrected (or corrected)
states for the two formes of a sheet. A much less extensive instance of this
phenomenon was discussed many years ago in connection with
seventeenth-century editions of Massinger's plays, and was taken to
demonstrate an "orderly regularity in perfecting".[14] This assumes, of course, that all
the
edition of a sheet is printed on one side before perfecting begins. In the
printing of the 1532 Furioso, the cases listed above involving
a
certain number of copies provide a substantial obstacle to the acceptance of
this explanation. The presence of the uncorrected state of both formes of
some sheets in a sizeable minority of the surviving copies suggests that
there was a delay in proofing these formes, and it is stretching credence a
bit far to require that the same length of delay which occurred in
performing this operation on the first forme should also occur later, in the
printing of the second forme. More consistent with the evidence in these
cases is the hypothesis that both formes were corrected at the same time,
after some of the edition had been printed and perfected. As for the
"near-miss" cases, the obstacle which they present to the acceptance of this
explanation can be overcome on the hypothesis that on some occasions the
correction of the forme on the press was delayed until after printing had
re-commenced. If the 1532
Furioso really did have a
print-run
of 3000, then the printing of a sheet would have been an operation barely
possible to perform on one press in two full days of continuous work, let
alone in one; in cases where the printing of a sheet did not start at the
beginning of the working day, the operation would have spilt over into a
third day. As a working hypothesis, I assume that the instances of copies
bearing uncorrected states of both formes of a sheet indicate the gradual
establishment, during the course of the printing of this edition, of the
practice of not making the final corrections to the formes until after the end
of the first day's printing, during which a portion of the edition of the sheet
being worked had been printed
and perfected.
[15]
The cancelled sheet
The regularity with which press-variants occur throughout the edition
and, in particular, the nature of some of those which are clearly attributable
to the author, involving words, phrases, and even whole lines, and, in one
case, two consecutive lines, lead one to reflect on the reason for their
presence. Given the quality of the Orlando furioso, its
importance to the author, and the circumstances of its printing, I believe
that they represent, not the main, but the final act of correction carried out
by the printer and the author on the text of this edition. They are, as it
were, the tip of the iceberg of correction; the bulk of the correction is
invisible, hidden to modern eyes, because it was carried out on proof-sheets
which have not survived; indeed, I would suggest that the surviving
press-corrections are due, not only to eleventh-hour changes of mind, but
also to the fact that the corrections carried out on the proofs were
sometimes so extensive as to make it
difficult for all of them to be correctly transferred to the type under the sort
of pressure normal in early printing.[16]
It would be nice to be able to produce, in support of this contention,
one of the proof-sheets of the 1532 Furioso, discovered in
those
treasurehouses of rare finds, the libraries of Italy. I cannot do that, but I
hope to persuade readers that I can do nearly as well, by demonstrating that
the cancellandum of inner A is to be understood as
representing,
more or less, the text of that sheet as it appeared in the proofs.
It was once again Debenedetti who first realised that the numerous
variants to be found between copies in ff. A3-6—he lists
approximately
one hundred—were attributable to two different settings of type.[17] He argued
that the readings in the first setting (Type I), while not identical with those
of the corresponding stanzas of the 1521 edition (which edition, duly
emended, served as printers' copy for the 1532 edition), nonetheless
represented linguistic choices closer to those found in that edition than to
those of the definitive edition; these latter were, however, completely in
accord with those of the second setting (Type II). His argument is
particularly significant in the linguistic context of sixteenth-century Italy.
Ariosto, as a Ferrarese, spoke a dialect very different, phonologically,
morphologically and lexically, from the Florentine which was the basis of
the literary language, and the story of the three authorised editions of the
Furioso is,
inter alia, that of a substantial
linguistic
journey, marked by the inconsistencies of genius but whose general
direction is quite clear, towards the adoption of a non-dialect standard based
on the language of the Tuscan classics,
as interpreted by the great Venetian writer and grammarian Pietro Bembo.
Bembo's fundamental work on the subject,
Prose della volgar
lingua, appeared in 1525, between the publication of the second and
third authorised editions of the
Furioso. The language of the
1532
Furioso, considerably more "regular" than that of the
1521
edition, is consistent with itself, from start to finish of the
poem—except
in Type I of ff. A3-6, where there is a substantial regression towards the
language of the 1521 version.
[18]
Debenedetti's explanation of this accident—for such it certainly
is—was that some pages of the 1521 text used as printers' copy had
been
marked with provisional corrections, which, due to the author's
absentmindedness ("era distrattissimo!", Debenedetti adds), he had never
revised. As with Debenedetti's consideration of individual variants, this
explanation engages the textual situation without paying any attention to its
bibliographical context, but whereas for the press-variants Debenedetti's
great sensitivity triumphantly carries the day (only one of his thirty-seven
textual choices turns out to be contradicted by the bibliographical evidence),
in the case of the cancelled sheet his lack of bibliographical awareness leads
him into an untenable position. The 1521 and 1532 editions, though
identical in structure—they are both quartos in eights—do not
have the
same disposition of text on the page: in the earlier edition each page has
two columns of four stanzas, while in
the definitive edition the text is arranged in two columns of five stanzas.
Thus, none of the sheets of one edition contains exactly the same text as the
corresponding sheet of the other edition, even in the first gathering, where
there is no new material. In the 1521 edition the portion of text contained
in inner A of the 1532 edition (Canto I, st.18-Canto II, st.14)
begins at the bottom of the first column of a3r and carries on through to the
top of a8r. Nor is the material of this portion a discrete textual unit: as
anyone can see by consulting the text either in the original or in translation,
the portion begins and ends
in medias res.
Obviously, Type I of inner A in the 1532 edition prints its not yet
linguistically or stylistically mature text because that is what the compositor
found in his exemplar, which, as has already been said, was a marked-up
copy of the 1521 edition.[19] The
crucial question is, what was the relation of the text in the relevant portion
of the exemplar to the text it carried elsewhere? Debenedetti's explanation
requires that, while for the rest of the work the exemplar should bear a text
which, except for the press variants, (which, apart from one
instance—the outer forme of outer C—comprise only a handful
for
each forme), had already reached the stage of linguistic and stylistic
maturity witnessed by the definitive edition, for the seventy-eight stanzas
in question, which comprise neither a discrete bibliographical nor a discrete
textual unit in the exemplar, it should have a text which failed to reach this
standard by such a long way that an average of
nearly fifty corrections per forme had to be introduced in a new setting of
type to bring it up to the same level as the rest of the work. In human
affairs, nothing is impossible; for the historian, however, in his task of
reconstructing the past out of the inadequate documentation which remains,
it is the probable which counts. In the circumstances, what Debenedetti
suggested is improbable. More probable is that the text represented by Type
I of sheet inner A reveals the linguistic and stylistic condition of the
exemplar not only in that passage but in the surrounding text as well; the
reasonable assumption is, in other words, that all the text of
the
exemplar, at least in the early part of the work, presented this same
provisional linguistic and stylistic quality. The reason why this provisional
quality does not now show elsewhere in the 1532 edition, except in the
press-variants, can only be that it was removed by extensive corrections
carried out on proofs before printing
began. Type I of sheet inner A thus represents the text of the exemplar
before the author corrected the proofs, Type II indicates the nature and the
order of the corrections he introduced into the proof-sheets.
The fact that three of the four surviving vellum copies of the edition,
which were presumably sent to Ariosto's patrons and influential friends as
soon as printing was over, contain Type I of inner A suggests that Type II
was printed after all the rest of the edition, after, indeed, an unknown
number of copies had already left the printing-house.[20] What little can be gleaned
from an examination of Type II does not contradict this suggestion. Though
the running-titles of the 1532
Furioso are in general
uninformative, and allow no firm deductions to be made concerning the
pattern of printing (the verso running-title throughout the volume consists
of the single word "CANTO", and the recto running-title, which contains
the canto number—e.g. PRIMO, SECONDO, etc. up to
forty-six—changes on average every one and a half sheets), A6r of
Type
II has the unique error "SCONDO", which shows that the inner forme of
Type II was printed with a skeleton different from that used for the
corresponding forme of Type I.
[21]
One can only speculate how Type I came to be printed. Perhaps the
author was prevented by other commitments from correcting proofs for this
sheet. More likely, I think, is the hypothesis that he made corrections, but
that they failed to find their way into the forme. We are presumably at the
very beginning of the printing of this edition; the printing house may not
have yet settled into a proper rhythm for this job.[22] It is significant that, in addition to
the
cancel, the one sheet with numerous press variants—outer
C—also
belongs to the early gatherings of the volume. The high number of variants
in its outer forme, twenty-three in all, suggests that here, too, the author's
corrections had not been incorporated into the forme before printing began;
in this case, however, the omission was noticed early in the press run (four
of the copies examined have the uncorrected state, nineteen, including the
vellum copies, the corrected state) and so did
not require a cancel.[23]
In postulating an Ur-text of the definitive edition of the
Furioso entrusted to the printers' copy and requiring heavy
correction to bring it to its final form, I am not suggesting anything which
clashes with what we know of Ariosto's artistic practices or of his
preparatory work on this edition. Like most writers, he was always
dissatisfied with what he had written; the letter from his brother Galasso,
already quoted, shows that even after the publication of the definitive
edition he was thinking of further amendments. As for the preparation of
the text of the definitive edition, he had certainly begun work by early
1528, the date of his request for a privilege from the Venetian authorities.
In the swiftly changing climate of opinion on linguistic matters in
sixteenth-century Italy, four years was a long time: solutions adopted early
in the correcting process could easily seem in need of substantial
modification when the work went to press.
[24]
One has to admit, however, that if Type II of sheet inner A was a
cancel, intended to correct an oversight which had left some early pages of
the text linguistically and stylistically disfigured, it was a very unsuccessful
one. Of the twenty-four surviving copies listed in Appendix A only three
contain the cancellans. I have no satisfactory explanation of
why
this should be so. I assume that it is connected with the fact that the author,
who owned the edition and probably took responsibility for its distribution,
as he seems to have done with the first edition, was afflicted with a fatal
illness within three months of the completion of the printing.[25]
Perfect copies of the 1532 Furioso
When I embarked on this study, I expected that the numerous
uncollated copies of the 1532 Furioso would reveal
unrecorded
states of some formes, which would allow me to make a contribution,
perhaps an extensive one, to the constitution of the critical text of the work.
These copies do indeed contain such states, with many textual variants, as
has been said, yet I am unable to propose a single change to the text
established by Debenedetti, because his text, in every case, already has the
correct reading in these places.
To understand how this is so, we must go back to Debenedetti's work
on this edition. Having considered the thirty-seven variants affecting the
reading of the text, and decided, impressionistically but with remarkable
accuracy, which reading was to be preferred, Debenedetti then looked at the
way those readings were disposed in the eleven copies he had consulted. He
found that two copies (nos 9 and 10), though they differed in one variant
for which he
was uncertain of the correct reading, agreed with each other everywhere
else, and always had the correct reading. Both copies, furthermore, had
Type II of sheet inner A. Happily unaware of the view expressed by
McKerrow in his
Introduction, that "it is quite unscientific to
speak of a more or less corrected
copy of a book",
[26] Debenedetti came to the conclusion
that
these two copies were superior to any other, and followed their readings in
his text. In this he was helped by the fortunate chance that copy no. 10 was
readily available to him, having been deposited by its then owner in the
Biblioteca Nazionale of Debenedetti's home town, Turin, for his
convenience.
[27] Working from a much
wider knowledge of the press-variants of the 1532
Furioso,
I
can only confirm Debenedetti's findings. While the other paper copies have
a varying number of formes, ranging from nine to thirty, in their
uncorrected state,
copy no. 10 has none at all: all ninety-one formes with press-variants are
present in their corrected state. The "perfection" of copy no. 9 is only
slightly less: the only difference between it and copy no. 10 is the variant
referred to by Debenedetti, which represents the second round of
press-corrections on the inner forme of inner G. Given the number and
percentage of formes involved, it is not possible to attribute the "perfection"
of these copies to the workings of chance.
Debenedetti described copy no. 10, somewhat ambiguously, as
"large", and as "the only copy with margins intact".[28] Thanks to the courtesy of the
present
owner, I have been able to examine the copy recently, and have formed the
opinion that it is printed on larger paper than the other surviving paper
copies. This opinion is based on a comparison between the measurements
of the margins in copy no. 10 and those of the spaces between the edges of
the type-page and the edges of the sheets in copy no. 21, a complete copy
in sheets of the 1532 Furioso.[29] We have only to posit the practice
of
running the large paper copy or copies through the press last of all in the
printing of each forme to obtain a satisfactory explanation of why copy no.
10 always has the corrected state.
The situation is not as clear in the case of copy no. 9, whose
dimensions are approximately the same as those of the majority of other
paper copies.
It is possible, of course, that it is a large paper copy, like copy no. 10,
which has been heavily cropped. However, in many sheets a countermark
is visible in one of the corners.
[30]
There is also the fact of the variant reading in the inner forme of inner G.
which differs from the reading of copy no. 10; the reading of copy no. 9
was taken off the forme in the first part of the run, to judge from the
readings of surviving copies (six in agreement with copy no. 9, fifteen with
the reading of copy no. 10). Inner G of copy no. 9 certainly did not go
through the press at the end of the run, and is thus probably not a large
paper sheet. If copy no. 9 is not a large paper copy—and that is the
direction in which the evidence at present seems to point—then the
author
and the printer were able and willing, when making up complete copies, to
differentiate between normal-sized copies of each sheet which had corrected
states and those which had not, and to form
copies on normal-sized paper of the whole work containing only corrected
states.
[31] It is surely significant that
both copy no. 9 and copy no. 10 also have the
cancellans of
sheet inner A.
[32]
Ariosto was one of the greatest poets of the European Renaissance,
and the Orlando furioso was his life's work. He published it
in
his home town, Ferrara, while in the employment of the Ducal family, who
took a keen interest in the poem. He had the motivation and the means to
ensure that the printing of his masterpiece was carried out under his close
and continual supervision. The evidence suggests, as I have argued in the
preceding pages, that, when publishing the definitive edition of his great
work, he availed himself of these circumstances to receive and heavily
correct proof-sheets, either of each forme, or of both formes of a sheet
together, and that he arranged with the printer to have some copies made
up consisting only of sheets in the corrected state. These are not unnatural
requirements on the part of an author; on the contrary, for one as great as
Ariosto, and in his enviable position, it would have been surprising if he
had been content with less.[33]
APPENDIX A COPIES OF THE 1532 EDITION OF THE
ORLANDO FURIOSO
[34]
Copies utilized by Debenedetti for the 1928
edition
- 1. Ferrara, Biblioteca Comunale Ariostea [S.16.1.16] Type
I
- 2. Ferrara, Biblioteca Comunale Ariostea [S.16.1.17] Type
I
- 3. Florence, Biblioteca Nazionale Centrale [Banco Rari 160]
Type I
- 4. Biblioteca Melziana [private collection; subsequently belonged
to Marchese di Soragna; present whereabouts unknown] Type I
- 5. Parma, Biblioteca Palatina [GG.II.197] Type I
- 6. Venice, Biblioteca Marciana [Rari 440] Type I
- 7. Milan, Biblioteca Trivulziana [Triv.G.101] Type I
- 8. Vicenza, Biblioteca Comunale Bertoliana; printed on vellum
[Bacheca 18.2.1.(raro 1)] Type I
- 9. Bologna, Biblioteca Universitaria [Raro C.5] Type II
- 10. Biblioteca Melziana [private collection; now the property of
Professor Cesare Segre, Milan] Type II
- 11. Vatican City, Biblioteca Apostolica Vaticana; printed on
vellum; presentation copy for Cardinal Ippolito II d'Este [Barb.lat.3942]
Type II
-
Other copies
- 12. London, British Library; printed on vellum; Grenville copy
[G.11068] Type I
- 13. London, British Library [C.20.c.11] Type I
- 14. Manchester, John Rylands University Library; Spencer copy,
formerly owned by G. A. Barotti [10195] Type I
- 15. Vatican City, Biblioteca Apostolica Vaticana [Rossiano 4686]
Type I
- 16. Vatican City, Biblioteca Apostolica Vaticana [Ferraioli
IV.4086] Type I
- 17. Oxford, Bodleian Library [4.A21.Art.]
Type I
- 18. Oxford, Codrington Library, All Souls College [mm.10.13]
Type I
- 19. Frankfurt-am-Main, Stadt- und Universitätsbibliothek [IL
1930/307 Nr.1] Type I
- 20. Florence, Biblioteca Nazionale Centrale [Palat.2.7.3.15]
Type I
- 21. Verona, Biblioteca Capitolare [R.VIII.4] Type I
- 22. Cambridge (Mass.), Houghton Library, Harvard University
[*IC5.Ar434.5160.1532] Type I
- 23. New York, J. Pierpont Morgan Library; printed on vellum;
Charlemont copy [PML 800] Type I
- 24. [formerly Cambridge (Mass.), Houghton Library, Harvard
University; sold at auction in London in 1962; purchased by Carlo Alberto
Chiesa of Milan, who sold it to a Milanese collector, now dead; present
whereabouts unknown] Type I
Notes