(main menu)
CAMPBELL
articles
CHARTS
menu
SITEMAP
site is
ABOUT
NEW
|
|
(main menu) |
CAMPBELL articles |
CHARTS menu |
|
SITEMAP |
site is ABOUT |
NEW |
|
1. British Isles and the Baltic
2. West Africa and the Cape Verde Islands
3. Corner scales, illuminated initial 'G', and borders
4. Scales and distances
5. Comparing west African place names from samotamat to cauo de sancta maria
| |
|
The intention of this note is to describe and analyse what appear to be the significant features of the visual appearance and hydrographic content found on the charts of Grazioso Benincasa. [For a list of those, and the works by his successors and imitators discussed in this note, see the Microsoft Word format Table 1: British Isles and the Baltic.] Where these findings identify conventions that are evidently unique to him, their recurrence in works that (now at least) lack a signature or date increases the probability that they are his work. Conversely, the absence of such features on charts in his style or with similar content points to the use of his models by copyists. Where the analysis identifies a chronology of development, this can indicate a terminus post quem for such copies.
The prompting for this investigation was the recent emergence, in Rotterdam, of an atlas clearly inspired in some way by Benincasa's work. In addition, the small number of supposedly pre-1501 works that have appeared since publication of my 'Census' in 1986, includes a Benincasa-type chart in Turin. How should the questions of authorship and dating be approached? What was the context against which such decisions should be reached?
Grazioso Benincasa is celebrated for several reasons. He was the most prolific of the pre-1501 chartmakers [or perhaps this was simply because his works, mostly atlases, had a better chance of survival than loose charts]. His well-preserved atlases have a crisp elegance about them and his reputation was kept alive by his successors, notably his own son Andrea, by Conte di Ottomanno Freducci (fl. 1497-1539 - though this may be two people) and by the latter's son Angelo (fl. 1547-56). Astengo (2007, p.217) notes that the work of the Freduccis was sometimes itself copied by Giorgio Sideri II (Callapoda). Collectively, these cartographic descendants perpetuated the appearance and content of Benincasa's charts beyond 1550 (see Corradino Astengo, in The History of Cartography, vol.3:2 (2007) pp.220-2, reproducing a sheet from an atlas of 1539) so that for almost a century elements of his style can be seen in the work of his imitators. In the case of the British Isles, this meant that the outlines he inherited, which were then transmitted via his successors, remained on printed maps until about 1550.
In 1468, Benincasa was the first portolan chartmaker to include the west African coast beyond Cape Verde as far as Liberia. He also progressively documented the Cape Verde Islands. The lack of surviving Portuguese charts before the 1480s gives that special significance. However, this note suggests new interpretations of those developments and transfers much of the credit to Petrus Roselli. Indeed, Benincasa does not seem to have thought the Portuguese discoveries very important. How far the west African coast extended on his subsequent charts depended entirely on the space he was prepared to make available on the atlas sheet concerned.
While Benincasa was not innovative in many respects, the toponymic analysis shows that that he did introduce more new names than his contemporaries. All were found on his early charts, up to 1465. Benincasa's previous career as a ship captain and his authorship of a detailed written portolano of the Adriatic, Aegean & Black Sea (1435-45) [transcribed by Kretschmer (1909, pp.358-420) and also by Biondi (1998), who includes a full reproduction of the original MS in the Ancona archives] indicates that he must have had considerable knowledge of the coastal place-names, for those areas at least. However, the 23 innovations attributable to him in the Microsoft Word table of 'Significant' Names do not fall disproportionately within the region covered by his portolano [see The addition of 'Significant Names' to the 31 sections of coastline (tables) (a Microsoft Word document)]. [That figure has been expanded to 38 innovations in the replacement toponymic listing, now in the form of an Excel spreadsheet - Listing and analysis of portolan chart toponyms along the continuous coastline from Dunkirk to Mogador... - sort on Column G.] {This sentence added 24 May 2015}
Perhaps any other new names that might have been in his portolano were picked up by chartmakers working before Benincasa started drawing charts in the 1460s [this point has not been checked]. It should be noted, though, that the toponymic analysis did not consider the occasional rare (possibly unique) names found on his charts. For more on that, see Innovative Names, and specifically the Toulon-Genoa analysis, in Table D of Rare names, toponymic gaps and reintroductions (a Microsoft Word document)]. {Alterations made to this paragraph 18 September 2011, & then entirely rewritten 11 February 2012}
This extended note follows an examination of those Benincasa charts and atlases that are readily accessible through reproduction. This means there is a comprehensive coverage for the period up to 1469 (via the indispensable DVD accompanying Pujades (2007)) but only intermittently thereafter. Given the prohibitive expense of obtaining high quality scans, this analysis would not have been contemplated had the Pujades corpus not been available. I wish to acknowledge, in the warmest terms, my deep debt to him and to those others who helped make it happen.
Perhaps one day the Pujades corpus will be extended onwards, at least to 1500 for signed works. There are 53 Benincasa charts and atlas sheets up to 1469 and 34 thereafter (as well as three now undated, but reliably attributed charts). [To the second figure, 34, should now be added the six sheets in the 1474 atlas re-discovered in the Ukraine National Library in Kiev, in May 2015]. Equally important, perhaps, would be to provide general access to those unsigned (and hence undated) works that have been loosely assigned to the late 15th or early 16th century. My present count of these is about eight and if, as seems likely, they prove mostly to be Venetian work (possibly officially produced) their almost total absence from past portolan chart studies risks perpetuating an imbalance in our understanding of the various regional contributions.
It is also a requirement for any informed consideration of the nature and authorship of the 'pseudo-Benincasa' works (discussed below under Benincasa, his successors and imitators) that their outlines, ornamentation and handwriting can be studied via high resolution images. Because of the uncertainties surrounding them, unsigned charts are far less likely to have been reproduced in print, which compounds the difficulty. Once they can be subjected to close examination (remotely on a computer screen) it may prove to be the case that some of those works assigned roughly to the period around 1500 may need to be moved earlier - as happened when Pujades examined a few of the productions loosely assigned to the 15th century - but it is more likely that they will be pushed firmly into the post-1500 period.
The date on Egerton MS 2855, as it now appears, is MCCCC . XXIII (i.e. literally 1423). However it seems probable that it was originally MCCCCLXXIII (i.e. 1473, the date normally given to the work). Beneath the large dividing full-stop there is a faint L. The most likely explanation is that an attempt was made to hide the L so as to make the atlas 50 years older. Tampering with the dates of early maps is far from unknown, particularly in the period before much had been published about them, leaving purchasers in ignorance. This work's provenance can be traced back to the Venetian monk Giacinto Placido Zurla (commentator on the Fra Mauro map) in 1818, subsequently passing through other hands before being bought by the British Museum in 1905.
Numerous factors corroborate a 1473 date, i.e. that the missing L was originally present. Benincasa would not have written a date with a full-stop in the middle; the other 1473 atlas (in Bologna) has the date written just as this must have been originally; Benincasa's known period of activity was between 1461 and 1482; the west African names were those introduced in 1468 with, significantly, a few minor peculiarities otherwise found only on the 1474 atlas; and none of the features recorded in the various Benincasa Tables raises questions about a 1473 date. Any doubts that have been expressed about the date should therefore be laid to rest.
Some features that are occasionally attributed to Benincasa prove to be borrowings from early practitioners. This applies particularly to the British Isles, where several visually distinctive features can be traced back to the 14th century. The channel separating England from Scotland (with its elaborate 'hinge' at the mid-point) was pre-figured by Vesconte and re-appears (if faintly) on the 1403 Francesco Beccari chart, and unmistakably on the 1426 chart by the latter's son Battista. Interestingly, Giacomo Ziroldi (in his first surviving work of 1422) prefaces the names of England and Scotland with ya (island), even if the divide was not depicted until 1426 (in the form of a pair of rivers as broad as the Thames). {This sentence added 10 February 2017}.
Indeed, Benincasa was firmly in the Beccarian tradition - presumably copying from a chart by Francesco's son, Battista, such as the one from 1435, or possibly one by the priest Bartolomeo Pareto (whose single surviving work, a chart dated 1455) was drawn close to the start of Benincasa's chartmaking career (around 1460). It is certainly significant that the father and son Beccaris and Pareto worked in Genoa, from which Benincasa signed his two oldest surviving charts. However, the 'Colour & Shape Analysis', on which see below, reveals that Benincasa took stylistic conventions from others besides the Beccaris.
Benincasa's habitual inscription ('Lacus fortunatus ubi sunt insule que dicuntur y sce beate ccclxvii'), beside the 'lake' (i.e. large bay) let into the west coast of Ireland (colourfully dotted with islands), is a close copy of Battista Beccari's. For some reason, however, Benincasa invariably noted 367 islands, rather than the figure of 368, found from the early 14th century onwards and the one used by Beccari, i e. ccclxviii turns into ccclxvii. See an illustration (for whose quality I apologise) here. [Though Vesconte used a different figure, 358, from the time he introduced an earlier form of this legend in about 1321] {This sentence added 24 February 2014}.
The larger figure (368) was found not just on the work of the Beccaris but also on anonymous Venetian atlases assigned by Pujades to the second quarter of the 15th century: his A 31 Vatican (Vat.Lat.9015) ['Census' 158 and for scans the Medea Chart Database] which has 'y. fortunate insule sante ccclxviii beate', and his A 30 Parma (II,32,1624) ['Census' 103 and for scans Medea Chart Database] which gives 'Insulle ccclxviij beate'. The full 367-island inscription above (setting aside the 367/368 issue) has not been noted on charts other than the signed work of the Beccaris, Grazioso Benincasa, his son Andrea (atlas of 1476 and charts of 1490 and 1508) and the chart in the Florence Archives (CN.9) that is almost certainly by Grazioso [see Table 1: British Isles and the Baltic and for a scan Medea Chart Database.]
As far as copyists are concerned (see below under Works formerly attributed to Benincasa or copies of his productions), only three of those charts that extend as far as the British Isles (of those it has been possible to examine via reproduction) include the inscription beside this feature: the Rotterdam atlas, the Turin chart (though this appears to have additional words) and the 1497 Freducci chart in Wolfenbüttel. In each case they repeat Benincasa's inscription, including his figure of 367 islands rather than the traditional 368 used by Beccari. This presumably indicates that their source was Benincasa and not some other copyist of Beccari. The 1537 Conte di Ottomanno Freducci atlas in the Hispanic Society, New York gives a figure of 372 (Sider, 1992, fig.8) - presumably a mis-reading of the last half of the figure as lxxii instead of Benincasa's lxvii. Giorgio Sideri (Callapoda), Benincasa's final disciple, continued with the 367 islands inscription between 1537 and 1563, after which (up to 1570) the available space was filled instead with a large coast of arms. {This sentence added 19 March 2014}.
The Czech text in Kupčík (2004) draws comparison between the three-sheet atlas in Rovigo (Concordiano 486 - 'Census' 108 and for scans the Medea Chart Database) and the work of the Benincasas. On p.85 he transcribes the 'IRlanda' and 'Lacus fortunatus' legends from the split-sheet chart (his Pl.14). Unfortunately the 'Lacus fortunatus' figure could not be read further than ccc. An equivalent legend could apparently be made out on the left hand of the two full sheets of the same work that form a coherent unit (which also includes the British Isles). In this case, he could read just the end of the figure, cccviii islands, i.e. 308. Kupčik mistook both these incomplete figures to represent dates - in this case, for an unexplained reason, 1508. The cccviii instead of the expected ccclxviii figure is perhaps due to careless copying, though the ending (suggesting 368 rather than Benincasa's invariable 367) might indicate a Beccari influence. See further below under Works formerly attributed to Benincasa or copies of his productions (Rovigo).
This study was a response to an article by Paul Harvey (2009) on the use of colour on medieval maps. Towards the end he mentions portolan charts and, in particular, the way "they coloured small islands overall in a variety of colours" (p.51). He concludes by saying that "the convention deserves detailed investigation. Its origin may be functional or decorative - or neither" (p.52). He mentioned Benincasa specifically.
The 'Colour & Shape Analysis' (C&SA) involved noting the colour, convention-type and sometimes shape of selected medium-sized islands and the major estuaries. In all, 105 features were involved. [For the C&SA project, its data and conclusions, and links to the illustrations page see The colours and shapes used to denote some of the smaller islands and the major estuaries on portolan charts up to 1500.] All the charts and atlases available to me in high quality scans were examined. The analysis was carried out during the first half of 2010, after the rest of this 'Benincasa' essay had been written. It both confirmed the earlier findings and strengthened them.
The 2007 Pujades DVD was of course essential for this exercise. Unfortunately his cut-off date of 1469, roughly in the middle of Grazioso's production and excluding all of Andrea's, meant that the analysis is incomplete for the Benincasas. However, sampling of the period after 1469 confirmed the earlier pattern, which was one of remarkable consistency. The findings as a whole highlighted a pattern in which the chartmakers before him, while often true to certain repeated Catalan, Genoese or Venetian conventions respectively, or to their own family 'house style', had not settled on a general standardisation. The colour and shape of selected islands and the larger estuaries, or the other unusual conventions that this analysis identified, were sometimes the same across the work of a single individual, but often not. Only Benincasa's near-contemporary Roselli came near his level of consistency, suggesting that the concept of almost total uniformity in respect of these features was not much valued before the middle of the 15th century.
When trying to distinguish work produced by the master (or at least under his direction) from that of a copyist, the most significant complement to the discovery of Benincasa's own consistency in relation to the C&SA features is that the sequences of colours he gave to the islands and estuaries are usually different from that of any other chartmaker and never, when taken as a whole, the same. This development confirms, beyond any reasonable doubt, the judgements about authenticity and imitation set out below in the section on Works formerly attributed to Benincasa or copies of his productions.
The C&SA findings constitute a large sub-section of these portolan chart pages. Access is via the section's Menu. The following individual pages have direct relevance for Benincasa studies:
Colour and shape on the work of Benincasa [a Microsoft Word table]
Colour and shape on the productions of Benincasa's successors and imitators [a Microsoft Word table]
Colour Feature List [descriptions of the varied treatment of the islands and estuaries considered in the analysis, with several references to Benincasa]
Summary table of the features analysed [a Microsoft Word table]
Explanatory notes and wider implications - see particularly the sections on Benincasa's consistency
Anonymous works and the question of their attribution to individual chartmakers or to their supposed workshops [see for a comment contradicting my early statement about the 'inexperienced hand of an apprentice' here]
For a demonstration that Benincasa's chromatic pattern was different from that of other 15th-century Italian chartmakers and from Roselli, see the appropriate 'Colour & Shapes' pages.
Before the C&SA findings became available, identified elements of a Benincasan 'visual signature' had been restricted almost entirely to the British Isles. Now, they could be related to Mediterranean and Black Sea features as well. These are often very minor, occasionally invisible under insufficient magnification or inadequate scanning resolution. Several of them can be considered trivial. Yet they gain a special importance because of their role as pointers towards a Benincasan authorship. No earlier surviving chart, for example, set a white name label into the fictitious islands of till and scurce (circular and oval respectively, either side of Scotland), nor, it seems, did any unrelated subsequent charts, which marks out the work of Benincasa's successors from all other productions. Benincasa was also the first to edge the three elements of the British Isles in different colours or to colour Minorca green (apart from a single internally inconsistent, unsigned Italian atlas (Bodleian, Douce MS 390)).
It was usual to leave a white device at the centre of Tenerife (inferno), to represent the snow-capped volcano, and this was given a variety of shapes. Benincasa already showed on his earliest works what can be interpreted as either a cross or a four- leaf clover . It appears thus on one of his two 1461 charts (Florence A.S., C.N.5) and habitually thereafter, but as a five-leaf version on the other (whose missing date is assumed to be 1461 - Florence A.S., C.N.6). Like the two stripes over Majorca, slanting from bottom left to top right, that particular clover/cross device was seen earlier only on the work of the Catalan chartmaker Rafel Soler (rescued from oblivion by Pujades), who was probably active some time in the period 1425-50. Because those two features are not otherwise seen on Italian charts, their adoption by Benincasa indicates a different source from the Beccarian model he usually followed.
From the Colour and shape on the work of Benincasa [a Microsoft Word document], it can be readily seen that he usually applied the same colour to each feature, if no great distinction is made between certain secondary colours, for example, brown, pink, mauve, purple. These would presumably have been mixed up in the workshop from the purchased primary colours. The most obvious discrepancy relates to the use of red on his charts for the islands in the Rhône Delta and for Corfu, where the atlases used gold instead (with the 1473 atlas being the only noted exception). Occasionally he left a feature uncoloured. Otherwise, most of his colour choices were invariable and, in sufficient cases for identification purposes, different from those of other chartmakers. [See further on this under Benincasa's consistency.]
The C&SA also considered the shape of selected islands. The crisp professionalism of Benincasa's atlases, and his past shipboard experiences as he gathered information for his portolano of the Adriatic, Aegean & Black Sea, might have prompted expectations of hydrographic refinement on his charts. [What does seem to have been a major contribution was his regular citing of soundings, e.g. 'passo XX di aqua' or 'V pie de aqua', presumably where this was of significance for the pilot.] {This sentence added 4 December 2019}. Instead, examination of the Benincasan outlines on selected islands in the Ionian group and the Aegean Sea finds that he was either unusually conservative - sometimes being less realistic than Vesconte a century and a half before - or, when chartmakers had generally abandoned a realistic treatment of these islands in favour of easily recognisable, imaginary outlines, Benincasa tended to be the most inventive. In the Ionian Islands his Corfu, for example, develops an arrow-head projection [on which see the note about the later copies in the Cornaro Atlas], while Zakynthos looks like a piece from a modern jigsaw puzzle. He either did not see or chose to ignore the more realistic outline for this island that emerges in some Venetian charts from the first half of the 15th century. Where earlier charts had given the Aegean island of Skyros two symmetrical rectangular bays, Benincasa increased that to three, so that it looks like castellation. [See Island shapes as a mnemonic device, for the suggestion that there was a serious purpose behind these fanciful creations, which Benincasa copied and sometimes elaborated.]
The single most useful visual indicator of Benincasa's work, and the one that on its own allows unsigned works, or those with obliterated signatures, to be attributed with confidence to him or his son, is his treatment of the headings for England, Scotland and Ireland. See an illustration (for whose quality I apologise) here. Cortesão (1971, 2:190) had already drawn attention to this. Several of the conventions Benincasa uses can be seen already on the 1403 and 1435 Beccari charts, but it is the combination of precisely handled elements that makes Benincasa's work distinctive. See Table 1: British Isles and the Baltic.
('Invariable' in the following, does not take account of those atlases by Grazioso whose location is known but for which no scans were available: 1471 (Vatican), 1473 Bologna and 1480 (Vienna), nor the 1476 Geneva atlas by Andrea. See below about the 1466 Paris and 1474 Budapest atlases, which omit the country headings)
INgliterra:
A. IN [with a capital N] - invariable
B. IN [two bars across the N, which is written as an upside-down U] - usual, but noted as missing on three atlases of 1468-9
C. INgl [squiggle going up and to the right of the 'l'] - invariable, except for Florence, Archivio di Stato, CN.5, which may therefore be earlier in the same year than CN.6 (the other signed chart made in Genoa, but from which the year has been erased). CN.6 is usually given the same date as CN.5, his earliest dated surviving work, i.e.1461.
D. Scocia [diagonal line going up from the top end of the S] - occasional
E. IRlanda [capital R] - invariable
F. Large dot before and after each of the British Isles country headings - invariable after Florence, Archivio di Stato, CN.5 & CN.6 (the two earliest charts). [This device was noticed on the 1436 Bianco atlas, but not on the 1448 chart - was it used by others before Benincasa?]
As far as related works are concerned only the damaged chart in the Florence Archives (CN.9), which seems clearly to be in Grazioso's hand, reproduces any of the features of Grazioso's characteristic handling of those country names. This strongly suggests that Grazioso's idiosyncratic flourishes here serve both to identify the autograph work of one of the Benincasas and also to distinguish that from the productions of copyists. This is hardly surprising, since an imitator presumably did not wish to broadcast his debt to another chartmaker.
As is so often the case, generalisations have to be modified. At least two of the Benincasa works that include the British Isles [of those that could be examined] omit the characteristic headings altogether: the 1466 atlas in Paris and the 1474 in Budapest. In most respects both those works conform to the Benincasa style but the Budapest atlas [seen only in small images] has an Ireland whose shape does not reproduce the usual Benincasa outline, perhaps due to careless copying.
It is unfortunate that the two charts that Pujades assigns to a possible Benincasa atelier (see below under Successors) do not, in their current trimmed form, extend as far as the British Isles. Andrea (in his atlas of 1476 and charts of 1490 and 1508) reproduced Grazioso's capitalisation of the second letter of INgliterra (A), that of Ireland (B - though the name was omitted entirely in 1508) and the framing dots (F). However, Andrea did not imitate Grazioso's three calligraphic flourishes (B-D). {Minor alterations to the preceding three paragraphs, 29 March 2012}
For an illustration of a work by a successor, Angelo Freducci, which perpetuates elements of Benincasa's style, see the treatment of the British Isles on the relevant chart of the 1555 atlas in the National Maritime Museum.
Just as the prominent placement of the names of the three main elements of the British Isles was rarely repeated for other countries or provinces - whether in the work of Benincasa or of other chartmakers - so that region was singled out visually in a further way. A distinctive feature of Grazioso Benincasa's work is the way he added colour to the outline of the British Isles. Instead of the discreet yellow often used to mark the coastlines, some earlier chartmakers had added a stronger wash, but none applied it to all three countries, 'England' [i.e. England & Wales], Scotland and Ireland, or, if they did, not with three different colours. From the time of his earliest work in 1461 Benincasa gave each entity its own colour. By 1463, the date of his first atlases, he went one stage further and added a softer version of each colour as a wash, inside the coastline where the place-names are written. The colours varied over time, although (interestingly, in the context of other comments about the development of Benincasa's visual style) the four atlases of 1466 and 1467 share broadly the same tones. Occasionally, England was left uncoloured, but the supposed physical separateness of Scotland was always emphasised.
There are obvious antecedents for this treatment of the British islands, the largest such entities on the portolan charts. The 14th-century Catalans, from Dulceti onwards, had covered the whole of Great Britain and Ireland with an overall wash. However, the combination of strengthened British coastal outlines and an internal wash can be found on the 1435 Battista Beccari chart, though not on the work of his father Francesco, nor on Battista's own earlier chart of 1426. Because of the interchange between the Beccaris and the Catalan masters, it is not surprising to find comparable treatment on the work of, e.g., Roselli and Bertran & Ripoll, even if Benincasa's British Isles is marked more strongly than theirs.
On his two earliest charts, Ireland was edged in blue but - on all those works I have been able to see in colour (including Andrea's chart of 1508) - from 1463 onwards Grazioso invariably used green instead. Whether intentionally, or because the green has faded less than the other colours, his Ireland is thus given special prominence. While it might be tempting to see here a reference to the 6th-century saint, St Patrick, and his green shamrock, this association with Ireland's famour colour has not been traced earlier than the 18th century, along with the supposed symbolic reference to the Christian trinity.
That the 1537 Conte di Ottomanno Freducci chart uses blue for Ireland instead, is just another possible pointer to the authorship of the Turin chart, which does the same.
None of the elements used by Benincasa to emphasise the British islands appeared first on his charts. Indeed, it can be assumed that he took over the idea from Battista Beccari. But the strong yet controlled way he presented the British coasts seems to form part of his distinctive 'signature'.
Two other features may be mentioned. First the diagonal scale across each corner. Corners had been cordoned off by earlier practitioners, for example Vesconte and Ziroldi, to provide space for religious or other illustrations. But in those cases, no scale was placed within the diagonal lines. However, an anonymous Venetian atlas, assigned by Pujades to the second quarter of the 15th century (Vatican, Vat.Lat. 9015 - Pujades A 31, 'Census' 158), adds a scale bar to the diagonal lines enclosing religious pictures. Nevertheless, the way the scale is presented, as alternative groups of six verticals, is quite different from Benincasa's version (see Table 3: Corner scales, illuminated initial 'G', and borders).
In any event, close examination of Benincasa's diagonal scales, and of others on works that are clearly copies of his, allows two assertions to be made. First, that it was Benincasa who introduced the diagonal scales (of a type now to be described) and, second, that, while diagonal scales have been found on some 'pseudo-Benincasa' works (see below under Works formerly attributed to Benincasa or copies of his productions), none of those yet seen in reproduction is in his recognisable style.
Some patterns can be observed in the precise lay-out of the diagonal atlas scales:
It may be reading too much into what might have been casual decisions by the draughtsman (presumably Benincasa himself or an assistant) but this element was handled in an internally consistent way on each individual atlas sheet, suggesting that the variations were both intentional and a reflection of a developing style.
Diagonal scales are the most recognisable diagnostic feature on those pseudo-Benincasa atlases that do not extend beyond the Mediterranean, and they provide powerful corroboration of the link with Benincasa, or of direct copying from his work. Examples are Milan (Bib. Triv., 2295 - see the Medea Chart Database), Paris (MS. Ital. 1710 - Medea Chart Database), Parma (II, 29, 1621 - Medea Chart Database), Rovigo (Concord. 486 - Medea Chart Database) and the recently identified Rotterdam atlas (see the Medea Chart Database). Another of the Rovigo atlases (Silv. 68 - see the Medea Chart Database), reproduced in Kupčik (2004) has a single, much larger scale bar across four of its five sheets. Others, not yet seen in reproduction, may need to be added to that list, on which see below under Works formerly attributed to Benincasa or copies of his productions. What unites the Paris, Parma and Rovigo (Concord. 486) atlases (at least), and distinguishes them from the examined Benincasa works, is the fact that the six vertical dividers separate alternate compartments with mile dots from those sections left blank. This arrangement is also found on the 1476 atlas signed by Andrea Benincasa. Is that significant, or is the same found on Grazioso's latest surviving atlas, of 1480 (for which no illustration is available)?
Those working later in the Benincasa style repeated the diagonal scales, with minor differences, such as Conte di Ottomanno Freducci (up to at least 1539) and, from the same period, Georgio Sideri (=Callapoda) [see Tolias, 1999, pl. 6-9]. Battista Agnese used a simplified device, just a line of dots across one corner with an explanation, which is visually very different. The diagonal scale device was also adopted generally by the compiler of the collection of archive copies of the work of assorted chartmakers, the Cornaro Atlas (c.1489) in the British Library (Egerton MS.73). This includes five copies of Benincasa's charts, making him the most reproduced chartmaker in that work.
Finally, the form of Benincasa's signature. While on the charts it was usually placed as a simple, single-line inscription, within ruled borders across the neck, the atlases offered possibilities for elaboration of the enlarged initial G of Grazioso, with which the inscription starts. This authorship inscription is almost always found in spare space to the east of the Adriatic on the chart of the central Mediterranean, extending down across two (or occasionally three) lines. On the oldest surviving atlas (1463) the 'G' is framed by parallel lines, but for the next surviving atlas (1465) this had mutated into the form that would continue certainly until 1469, with the 'G' inside a ruled square, supplied with a long foliate tail in red ink. Not all of the later works could be examined but the signatures on the 1471, 1473 and two 1474 atlases are different from the earlier forms. It seems that by 1471 the flamboyance of his earlier initial was starting to be curtailed, first by a reduction in the length of the ornate tail and then, in 1473, by that's reduction to a faint outline. {Alterations and additions to the previous two sentences 24 May 2015}. See Table 3: Corner scales, illuminated initial 'G', and borders.
The only comparable instance I could identify was on the 1455 chart by Bartolomeo Pareto, a priest in Genoa. The P introducing Presbiter at the beginning of his authorship statement could have given Benincasa the idea. Pareto has already been mentioned as a possible intermediary between the Genoese Beccari family and Benincasa, whose career seems to have started in that city.
The flamboyant (boastful?) appearance of Benincasa's signature seems not to have been imitated by any of his contemporaries [though I would be happy to hear of other instances]. Significantly, perhaps, Andrea supplies a simple signature for his sole surviving atlas. However, his devoted follower Georgio Sideri (Callapoda) took advantage of the shared initial letter for his first name to imitate Grazioso's ornate style on his earliest surviving work, an atlas of 1537 (Tolias, 1999, pl. 8). {This sentence added 20 December 2011}
It is sometimes the case that the meticulous gathering of visual information which, at the time, had no obvious significance, can later prove valuable. This seems to apply here to what most would consider a trivial feature, namely small variations in the treatment of the ruled borders.
Richard Pflederer, in the CD-ROM accompanying his survey of the portolan charts in several libraries, made a point of noting for the Benincasa works in the British Library (which holds the best collection of those) the number of ruled lines that made up such borders. This information has been extracted from the CD and is codified towards the right of Table 3: Corner scales, illuminated initial 'G', and borders. Two columns note the number of ruled lines in, respectively, the outer border and the pair framing each diagonal corner scale. The data offered by Pflederer have been supplemented from the Pujades DVD for the pre-1470 works.
Whether those edgings comprised three or four lines would have no interest were it not for the fact that they seem to point to a clear chronological development. As always, the analysis is incomplete because of the lack of reproductions of Benincasa's later works, but it looks as if, around 1466, Grazioso increased from three to four the number of lines he placed at either side of his corner scales, and the next year did the same for the outer border. This minute feature might even be sufficient to place the British Library's 1467 atlas (Add. MS 11547) a little earlier in that year than Paris's DD 6269. This suggestion is supported by the inclusion of Pegli (just to the west of Genoa) on all early Benincasa works (up to and including both Add. 11547 and the third 1467 atlas, BnF DD 1988) and its omission from all later works (including DD 6269) {sentence added 20 April 2011}
The border detail might also suggest an early date for the source of the Parma copy, while the additional ruled lines in the Rotterdam atlas add to the evidence that it is not an entirely slavish copy of a Benincasa work. [As an aside: in seeking to establish a chronological sequence among Benincasa works of the same date it is unfortunate that the only two to include additionally the month and day (the charts of 1461 and 1470) are the sole surviving works from those particular years, whereas for other years we have multiple examples.]
Pflederer also noted the distance in millimetres between the outer border and the edge of the atlas sheets, even if the sample is even smaller - just the five British Library atlases. For the atlases of 1463 and 1467 it was 3 mm, for 1468 and 1469 it was 4 mm, and for 1473 it was 5 mm.
Further analysis would be needed to determine if those chronological patterns were more than coincidence. But is it so unlikely?
Each Benincasa atlas comprised four or more sheets. Three atlases survive for 1467 alone and he presumably produced a large number during his career. Even if the drafting sequence of the invariable background elements has not yet been conclusively established, every sheet required such tedious housekeeping tasks. Had Benincasa used helpers or apprentices we can assume he would have delegated such work to them. If so, he would surely have taught them a precise procedure to follow. On the other hand, if he had to fill in those mathematical elements himself would he not have followed a similar, unthinking routine? If, at a certain point, the addition of an extra line seemed to be an improvement, that might naturally have become the new norm. Which is also why a copyist would find it easiest to make an unthinking imitation of what was in front of him.
There are insufficient separate charts among Benincasa's surviving work, and their legibility tends to be poor. But the treatment of the borders to the scales found in the upper and lower margins of his charts might possibly reveal a parallel development.
The value of the scales Benincasa used is also relevant in the context of drafting. Combining once again data from Pflederer and Pujades, this time in Table 4: Scales and distances, different patterns become clear. The Pflederer CD-ROM records information about the scale and the distance between six paired points, measured on the five atlases and a single chart in the British Library. Pujades calculated the scale and five paired distances on the larger group of pre-1470 surviving Benincasa works (2007, pp. 206, 208-9 (original table in Catalan), 477-8 (explanation in English)).
The distance information from the two sources is not directly comparable and the sample is concentrated largely in the early period but it might inspire a future researcher to complete the data-gathering. What is very evident, though, from the two scale columns, is that for most of the time Benincasa used just two standard scales. There is, though, the tantalysing suggestion of a third, larger version, as found on the 1468 Servera chart in Majorca, which focuses on just the western part of the usual portolan chart region.
For convenience, Table 4: Scales and distances has been presented with a default sort order (which can be changed) arranged by ascending scale. This shows the groupings around the main scales but also demonstrates - for a range of reasons which Pujades cites (p.477) - why the figures show slightly erratic patterns. These may reflect the difficulty of calculating distance via reproduction, distortion of the vellum or, more interestingly, real differences in the coastal outlines.
The use of two main scales led to atlases which, as Pujades (pp.208-9) shows, tended to fall into one of two sizes, roughly 28 x 36 cm or 34 x 44 cm. No doubt those were sold for different prices. They also raise questions about the drafting process and the models used.
Table 2: West Africa and the Cape Verde Islands shows how the sheet lines used by Benincasa to divide up the two (later three) charts covering the coast of west Africa were not fully consistent. However, had the scale been factored in there as well it would have been apparent that the two instances where the north-west Africa coverage started on a chart that extended well above Lisbon were the two drawn at the smaller scale. Nevertheless, even if a group of four charts start at petronella three others at that same larger scale do not. There is even more variation in the terminal names on the second west African sheet. Why? If he had been working from an atlas exemplar for each of the two scales his sheet lines would have been constant.
It is probable that we are seeing here a reflection of different large models that he was using for those two main scales, and not ones already broken down into separate atlas sheets. Is it not more likely that he had entire charts drawn at Pujades's 1 cm and 1.3 cm scales [see Table 4: Scales and distances], from which he traced off what he needed for each sheet? If he had such models the task of scaling up (or down) would have had to be done once only. Some time after the 'Chapter' appeared (1987) I studied an unsigned Dieppe-School atlas in the British Library ('Egerton MS 1513: a remarkable display of cartographical invention', Imago Mundi 48 (1996), pp.93-102, see particularly p. 95). That dealt with a late-16th century world atlas, and three different scales were identified as models for the overlapping sheets. Coincidentally, the larger scale of the Servera chart might perhaps indicate that Benincasa also had master charts drawn at three different scales.
The scans on the Pujades DVD are a magnificent research tool but they cannot be downloaded. It is not, therefore, possible to print off enlarged details to compare the same area from different works. One day, if that became possible, visual or cartometric comparison could test the hypothesis that, like Egerton MS 1513's creator, Benincasa was drafting sheets, to his convenience, by extracting copies from one of two (or three) models. In the Egerton case, variations in outline for areas covered at different scales showed that the models involved were not hydrographically identical. Might the same apply in Benincasa's case? Pujades (2009) p.352 refers to the three scales used by Gabriel de Vallseca. Perhaps this was a usual 15th-century chartmaking practice
It was pointed out that Egerton MS 1513's drafting exemplars had evidently been inherited from a previous master, in that case Guillaume Le Testu. Any cartometric examination of Benincasa's work should also consider that of his known successors. If they had inherited his working models this would help to explain their long perseverance with outdated outlines.
This region was included intermittently on portolan charts from the early 14th century onwards. Blocked from access to the Baltic by the Hanseatic League after 1323, subsequent Italian chartmakers usually omitted that region or repeated earlier outlines.
An alternative explanation for the neglect of the northern reaches of the Baltic by southern chartmakers would be that the Swedish king Magnus IV had declared a trade embargo barring all foreign vessels from sailing farther north into the Gulf of Bothnia than the authorized foreign trade centres of Stockholm and Turku. Successive commercial codes ("Bottniska Handelstvånget") were issued in the period 1349-57. This information came from Ulla Ehrensvärd, Pellervo Kokkonen & Juha Nurminen, Mare Balticum. The Baltic - Two Thousand Years (Helsinki: John Nurminen Foundation, 1995) p.132, supplemented by a private communication from Ulla Ehrensvärd. {This paragraph added 26 June 2014}
In this context it is worth noting that the Baltic is included on the following works by Grazioso, his successors and imitators:
The outlines on the two Grazioso charts apparently derive from the 1436 Bianco atlas or more likely via the 1455 Pareto chart. Like the other works above there is nothing original in the Baltic outlines, which essentially repeat what can be seen on the Catalan Atlas. Indeed, the Dulceti chart of 1330 demonstrates a better understanding of the Gulf of Bothnia than any of those, though not of the Gulf of Finland.
Surprisingly, apart from the first two, the works above do not generally seem to have been directly copied from one another or from a common model. Most distinct is the Rotterdam form, which, like the Bianco atlas, gives no hint of the Gulf of Bothnia. It is not, however, directly copied from Bianco, the only previously known atlas sheet devoted to the Baltic - see 'Chapter', fig.19.16, p.409, reproducing the western half (with south at the top); the whole folded sheet can be seen in the DVD accompanying Pujades, 2007 (under A 18; select '7' and rotate it through 180 degrees).
So what was the source for the Rotterdam atlas, and why did its compiler trouble to include a separate chart (sheet 6) for the Baltic? No surviving work by Benincasa contains an extra chart of that type, covering the Atlantic coast from Bruges up to Norway, and then into an unrecognisable Baltic.
The partial coverage of the Baltic on Vatican, Borgiano V gives a distinctive shape to Denmark but this work has no obvious connection with Benincasa.
Since, apart from the hint (no more) of the Gulf of Bothnia, there was no evident improvement in outlines spanning at least the period 1330-1508, it is tempting to suggest that the visible differences in the coastlines were due to the careless copying of outlines that were largely unknown to the chartmaker and of little interest to him.
The four islands of the mythical Antilia group, first shown on the 1424 chart, and repeated by Battista Beccari, Bianco and Pareto, appear on Benincasa's first atlas (1463), apparently modelled on Pareto (see Cortesão, The nautical chart of 1424 p. 84 - but note that the large northerly island, saluaga, has been scraped off that 1455 chart).
There was no place for the Red Sea in the atlases that form the greater part of Grazioso Benincasa's output. On the other hand, charts that traditionally extended as far south-west as Cape Bojador (in present-day Western Sahara) could, according to the tilting of the Mediterranean axis that was accepted at that time, have included the upper half of the Red Sea. Benincasa's surviving general charts of 1461 (one dated, the other assigned to that year) and 1470 were in the restrained Italian style and omitted inland detail, so the Red Sea is not shown. For that reason, the Red Sea appears only on his final work, the chart of 1482. Since this is in the Catalan style, with much illustrative and inland detail, the Red Sea is included. I can find no reproduction for the eastern half of the undated chart reliably attributed to Grazioso (Florence, C.N.9) but its inclusion of illustrative elements such as the Atlas Mountains makes it likely that it too showed the Red Sea. Both of the cut-down charts suggested by Pujades as deriving from Benincasa's workshop are in the Italian style and would presumably not have included the Red Sea (Paris D 21815 and Vicenza).
This means there is only one depiction of the Red Sea by Grazioso that can be considered here. Two features are worth noting. First, the Nile is shown as being formed of two equal streams (converging from the south-west and south-east respectively) which meet at a point level with the north end of the Red Sea. Second, the river shown flowing down from the north and then curving westwards into a lake above the Red Sea (Gulf of Suez) does not continue on to the Nile. These points are made because other charts gave a different interpretation of the region's water features. Freducci, on his chart of 1497, displays a single Nile stream flowing due north, though he follows 14th-century models in having a confluence of twin Nile feeder branches much further to the south, i.e., at a point below the bottom border of the 1482 Benincasa chart. Freducci's schema also connects the lake at the head of the Red Sea to the Nile. In those two respects he matches the arrangement found, for example, on Roselli's chart of 1466 in Minneapolis (Pujades DVD, C67). Grazioso's son Andrea, in his 1508 chart (Vaticana, Borgiano VIII), follows his father with the broken connection between the Nile and the Red Sea but imitates Freducci with a single Nile stream flowing due north.
So, from Benincasa's time onwards, there were various interpretations of the course of the Nile and of other real or supposed rivers in the region. This discussion is not concerned with who most closely approached reality, nor with the chartmaker who first adopted any particular outline, but rather how these variations can assist with the creation of the context against which to set the 'pseudo-Benincasa' works. For assorted reasons, the Turin chart [E.4] is, at present, the only candidate in this case. With the simple geometry of a south-north Nile and a river (with its subsidiary link to the Red Sea) flowing into the Nile at right-angles from the east, this conforms closely to the pattern found on the 1497 Freducci chart, thus providing further support for an attribution to that chartmaker. For comments on Catalan depictions of the Nile see Relaño (2001a, pp.104-5).
(on this see Cortesão, 1971, 2:193, and 'Chapter', p.412)
Benincasa is most celebrated for being the first to document the west African discoveries of Alvaro Fernandes, Pedro de Sintra and others (1446-61). As will be shown, this reputation is not fully deserved.
First a comment on the logistics of chart-making. It should be remembered that these new coastlines were literally off the bottom of the standard portolan chart, whose coverage was determined by the shape of the animal skin used. Extending to the Baltic in the north, to the Antilia group in the west or to Sierra Leone in the south presented logistical problems. This of course applied to portolan charts, but not to atlases, and Benincasa was primarily a creator of atlases - at least as measured (perhaps incorrectly, because of the greater vulnerability of separate charts) by those currently locatable or that survived into recent times (17 atlases against nine charts, if the three confident attributions are included).
Benincasa's depiction of Africa's north-west coast went through several major changes. The toponymic sequences of these, and those of selected predecessors, are set out in Table 5: Comparing west African place names from samotamat to cauo de sancta maria. This can be summarised as follows.
STAGE 1
In the first period, 1461-3 (including his earliest surviving atlas) Benincasa imitated the usual 14th-century toponymy down as far as cauo de buçedor (Cabo Bojador at 26o North, in present-day Western Sahara). This coastal stretch had been included on the 1330 Dalorto and 1339 Dulceti charts (probably the work of one individual) as far as the vellum would allow, that is just short of Bojador. The outlines and names were then extended down to that cape on the Catalan Atlas of 1375 (or they possibly first appeared on the G. Soler chart that Pujades dates to the period 1368-85 - his C 14 - although that is more likely to belong to the end of that period). This same suite of names appears on the undated, and jointly-authored, Corbitis and Pinelli-Walckenaer atlases (Pujades, 2007, A 11 and A 12). Their dating has been uncertain in the past but they can probably be assigned with some confidence to the early 15th century. {These two sentences added 11 November 2013}
There is no difference between the Benincasa toponymy and that of the 1375 Catalan Atlas. He did not, at this stage, repeat the few additional names listed by Andrea Bianco in 1448. Since Bojador lies just south of the Canaries, its inclusion along with that island group provided a logical and neat cut-off point. For that reason most of Benincasa's own charts end there, including his final work of 1482.
STAGE 2
The next reflection of growing Portuguese knowledge of that coastline comes with the Bianco chart of 1448. This added a further 33 names below buçedor, reflecting Portuguese discoveries up to 1446. However, its unconvincing outlines and the casual placement of the two final and important names, cauo verde and cauo rosso, suggests verbal information rather than copying from a Portuguese chart. For that reason, and given the repetition of cauo verde in particular, it is probably not justifiable to describe the Bianco chart as providing a definite cartographic record of the position and context of Alvaro Fernandes's discovery in 1446 of Cabo Roxo, at the southern border of Senegal (12.20o N). For a theory that the Bianco chart was drawn neither in 1448 nor in London, see Andrea Bianco's "London" chart of 1448.
It is not until 1465 (not 1463, as I stated in the 'Chapter', p.412, following Cortesão), with his second surviving atlas, that Benincasa made the necessary adjustment to his sheet arrangement so as to provide space for the Bianco names. The five Benincasa atlases that survive for the period 1465-7 saw coverage extended to cauo rosso, on a single sheet that started above Lisbon. Produced after a gap of 17 years, this is the second dated chart to extend that far down the coast, though, if its dating of c.1462-4 is accepted, the Catalan Estense world map (reliably attributed to Roselli) slightly preceded Benincasa. [On that see under Catalan Estense world map in 'Anonymous works and the question of their attribution to individual chartmakers or to their supposed workshops'.]
Benincasa followed Bianco's outline and toponymy. However, his omission of seven of Bianco's names and the inclusion of three additional ones must point to an intermediate source. It seems likely that Benincasa only became aware of the west African information once he had moved to Bianco's home city, Venice, at some point between 1461 and 1463.
STAGE 3
This involved substituting a new coastal outline and replacing the sequence of names down to Cape Rosso, as well as adding a large number beyond that. The significance of the next development has not been sufficiently appreciated before. Cortesão (1971, 2:152) described a Roselli chart of 1464 (in Nuremberg) as being the first after Bianco's to extend beyond Bojador, 'with eighteen or twenty place names on it'. He mentioned a few of those names. With the aid of the welcome zooming facility on the Pujades DVD, it is possible to arrive at approximations of this new sequence of names (apparently 23 in all, of which only three had definitely appeared on Bianco's chart). These names prove to be (with variations) the first part of the sequence that Benincasa was to include four years later, and on which much of his reputation rests. Although most of the names first seen in 1468 are new, confusingly a few are repeated. These include the names of genuine discoveries, but now spaced out correctly and aligned to a surveyed coastline. Some of the names repeated from the 1448 Bianco sequence are notable as apparently representing the end point of earlier Portuguese voyages, such as rio d'oro (Baldaia, 1436), c.bianco (Tristão, 1441) and c.rosso (Fernandes, 1446). The repetition of a few salient names should not distract attention from the majority of new names inserted between them.
Roselli has space on his 1464 chart to extend approximately seven names below the navigationally important headland of cauo bianco, in other words, perhaps two names beyond the island of falcon, though, even at magnification, the writing is very unclear. This seems to represent the discoveries made by Nuno Tristão in 1444 but not the major extension down to Cape Verde, achieved the next year by Dinis Dias. On his chart of 1465 Roselli repeats those names (apparently adding one extra) but stops off before cauo bianco at what seems to be restinga. Neither of these cut-off points seems to have had any significance and must merely reflect no more than the space on the vellum available to Roselli. That Bianco had been able to extend much further on his 1448 chart was because that was restricted to the Atlantic coasts and arranged vertically not horizontally.
This point is important because, if the longer name list that Benincasa was able to include in his atlases was the result of a single transmission from a Portuguese source, it is fair to surmise that Roselli (who did not produce atlases) might have had access to information about the entire coastline down to present-day Liberia. Is it likely that, twenty years after the event, Roselli would have received information about the Tristão discoveries of 1444 but not those immediately afterwards (1445-7) that had extended knowledge as far as present-day Conakry? It seems highly implausible that Roselli, in 1464, would not have known as much as Bianco did in 1448. Even if the names are not readily recognisable, Bianco's sequence does go considerably further than Roselli's.
Benincasa has always been credited as the intermediary, but perhaps this was simply because Roselli did not find a way to use that information in the format of a single vellum sheet. Of course, Roselli might not have known about Pedro de Sintra's voyages of 1460 and 1461, which extended the point reached in 1447 by a mere two and a half degrees, and provided Benincasa with his final twenty or so names.
Cortesão (1971, 2:192) considered that the mixture of Portuguese and italianised names pointed to Benincasa's sources as being a combination of lost Portuguese charts and first-hand information from Alvise da Cadamosto, whom he met in Venice in 1463. The chronology described above does not sit easily with that hypothesis, since Benincasa was still not using in 1467 whatever information Cadamosto might have supplied. Nor did Cadamosto's voyages in the 1450s produce further coastal discoveries.
I am grateful to Richard Pflederer for drawing my attention to a review (Geographical Journal 91, 6 (1938) pp.556-7), by G.H.T.K. [i.e. George Herbert Tinley Kimble] of G.R. Crone's edition of The Voyages of Cadamosto (Hakluyt Society, 1937). Kimble took issue with Crone's contention that Benincasa was indebted to Cadamosto, listing a number of reasons why Benincasa was not reproducing Cadamosto's place-names. I should also acknowledge that Kimble had also anticipated my comments (below) about the lack of significance attaching to Benincasa's various west African terminal names, ascribing those instead, as I do, to the 'exigencies of the shape and size of the charts'. The Wikipedia entry on Cadamosto also contains interesting material. {This paragraph added April 2012}
Comparing the Roselli names with the equivalent stretch on Benincasa's atlases - and making allowances for the limited legibility of Roselli's - it looks as if about six of the Roselli names are not repeated by Benincasa, and eleven of Benincasa's are missing from Roselli's list. Therefore, while evidently recording the same discoveries, neither used precisely the same source, unless each made his own selection from a common list. The change of toponymy was accompanied by a redrawn, and improved coastline, for which Roselli and Benincasa evidently used the same model.
For the atlas-maker Benincasa there was no problem about a major extension down the west African coast. He simply readjusted his page arrangement to give a second sheet to that coastline. It must be just a coincidence that the first of those two sheets normally ends beyond cauo bianco, roughly where the 1464 Roselli chart finishes.
Benincasa first displayed the new information on one of his two surviving atlases of 1468 (BL Add 31315). The names now extend to Bassa Point in Liberia, named as cauo de sancta maria (6o N). [The ex-Kraus atlas of 1468 was incorrectly described earlier as having the same extent, but a scan provided by Christie's for their sale of 19 November 2014 reveals that it actually stops at cauo de sca anna, in other words one name earlier than any others in this late group.] {This correction added 4 November 2014}.
This means that Benincasa listed about 90 names below buçedor. Only 13 of those had appeared in the Bianco sequence. The new coastal outline is notably different and, although it does not yet approach the accuracy of the earliest surviving Portuguese charts from the 1480s, it does contain sufficient features and correctly-placed names to confirm that Benincasa was copying a chart and not a written account. No serious attempt has been made in Table 5: Comparing west African place names from samotamat to cauo de sancta maria to match the Bianco and Roselli/Benincasa sequences, nor to relate them to present-day names or to explorers' accounts. The matching of names is particularly difficult in relation to the Catalan (Estense) world map [Modena, Estense, C.G.A.1], confidently attributed to Roselli and apparently produced in the period 1462-4 [On that see under Catalan Estense world map in 'Anonymous works and the question of their attribution to individual chartmakers or to their supposed workshops'.] Not surprisingly, given Portuguese reluctance to share coastal knowledge, Benincasa was never to pick up the discoveries of the 1470s, which appear for the first time on undated Portuguese charts assigned to the 1480s, and on a dated chart by Jorge de Aguiar of 1492.
The additional names required an extra sheet [or, on the larger-scale atlas of 1468 offered for sale by Christie's on 19 November 2014 (and again in New York on 14 June 2018 Lot 15), four sheets altogether - even if the name selection appears to be unchanged] but, as was common with Benincasa, he did not always end his sheets at the same place [on which see Table 2. West Africa and the Cape Verde Islands.] This provides the most obvious explanation for the fact that most of his atlases stop just before his furthest extent, cauo de s. maria. From 1468, there were three sheets covering the Atlantic coasts: (1) comprising the British Isles, France and northern Spain, (2) Portugal to approximately cauo bianco, and (3) from there down to Liberia. How each was divided determined how much space was left at the bottom of the final sheet [see the list below]. It was clearly not a priority for Benincasa to include the full complement of names - they would have been meaningless to most of his users and only the Portuguese navigated those coasts - and so he continued until he ran out of space, sometimes half obscuring the final name in the margin, or, in the case of the 1474 atlas rediscovered in Kiev in May 2015 (E.27), running riu de up against the outer border without finding space for the operative final word, palmeri. In other words, what was shown depended on drafting decisions, not hydrographic ones. Andrea continued in the same way. The north-west Africa sheet in his 1476 Geneva atlas stopped at buçedor, despite there being considerable empty space below. He did not, like his father, enlarge the five-sheet atlas format to make room for the Portuguese discoveries.
When Grazioso added the coastal names in 1468 he simultaneously inserted the main body of the Cape Verde Islands and the Bijargos Islands, following that up the next year with further islands in the Cape Verde group (see the next section).
One other Benincasa peculiarity should be noted. Apparently on a single occasion (the 1468 British Library atlas that displays the extended list of west African names) he altered slightly the otherwise invariable sequence above buçedor (see Column 7 of Table 5: Comparing west African place names from samotamat to cauo de sancta maria). Some names were omitted and a few others included, possibly owing their origin to Ziroldi (see his 1446 atlas in Florence, Pujades A20).
On his earliest works (1461-3), including his oldest known atlas, Benincasa covered the coast only as far as Cape Bojador, repeating what can be seen in the Catalan Atlas of 1375. Even though he was working in Venice by 1463, he did not demonstrate any knowledge of the work of the Venetian, Andrea Bianco (1448), or of the source the latter had used, until 1465. He then picked up a version of the Bianco names, down to a randomly placed cauo rosso, and used those for two years. During that time he was clearly unaware that already, in Palma, Majorca, by 1464 if not earlier, Petrus Roselli had introduced a new coastal outline with realistic names, at least beyond cauo bianco, which wholly superseded the information from Bianco. Finally, in 1468, Benincasa's flexible atlas format permitted him to introduce the configuration and toponymy of the coasts discovered by the Portuguese up to 1461 as far south as modern-day Liberia. He was, as far as we know, the first to display that information, but it seems the Majorcan Roselli may have possessed it before him. The differences in the name lists used by Roselli and Benincasa allow copies by others to be assigned to one or other of those two models. So far, it has been possible to test only two of the 'pseudo-Benincasa' atlases - Paris, BnF Ital. 1710 and the work in Rotterdam. Each of those followed Benincasa, not Roselli.
The following list indicates the final west African name on each Benincasa work for which there is information. This feature is also noted on the copies of his atlases by others, on the assumption that the draftsman concerned would have copied exactly what he saw, including the same terminal name. For example, the unsigned atlas in Paris (BnF Ital.1710) ends at riu de palmeri, even though there was space below for further names, and the usual corner scale had been omitted to permit those. The obvious differences in the name lists confirm which copies should be assigned to pre- or post-1467/8 models:
[with the extended coastline - the final six names]
NB this section was altered on 4 November 2014 after a scan became
available of the ex-Kraus 1468 atlas, for sale at Christie's on
19 November 2014 (and again in New York on 14 June 2018 Lot 15). This
showed that the terminal name in that case was cauo de sca anna and not, as previously thought, cauo
de sca maria]
Benincasa's other hydrographic innovation related to the Cape Verde Islands, though for the same reason (space) that his charts did not extend down that coast, the Cape Verde Islands appear only in the atlases. The group was discovered in 1455-6 but subsequent confusion about the names, shapes and placement of the separate islands makes interpretation of their representation on portolan charts difficult. Cortesão (1971, 2:184-8) credits Benincasa as being the first, in 1468, to show the eight main islands of the group (discovered by Diogo Gomes and Antonio da Noli in 1460). In the following year Benincasa increased that total to eleven (see Cortesão's comparative illustrations on pp. 186 & 187, or the scans on the Pujades DVD). Cortesão does not make clear if the further three islands represented discoveries from a later voyage. In his later atlases Benincasa continued to include the group, although, as they press up against the western edge of the west African sheet, he sometimes restricted coverage to the four larger, eastern islands. Since it would have been easy to move the African coast a little to the right on the sheet, it seems inescapable that, as for the final coastal names, he felt little urgency about including all he knew (for example, in the 1471 atlas in the Vatican and the 1480 atlas in Vienna). The degree of sophistication and completeness in the depiction of that island group cannot therefore be used as a reliable dating aid.
For our purposes, interest focuses on 'pseudo-Benincasa' atlases that include those islands: Rotterdam and Paris, MS. Ital. 1710. Scans of the first were made available to me for private study and Vagnon (2005) conveniently reproduces the relevant sheet from the Paris atlas (fig.4). As described below under Works formerly attributed to Benincasa, neither of these atlases is in Benincasa's hand, nor are they the work of the same individual.
Comparison between the depictions on the unsigned Paris atlas and the signed Milan atlas of 1469 (Cortesão, fig.84) shows such close similarities in shape and positioning as to leave no doubt the Paris atlas was directly copied from one of Benincasa's. However, Cortesão (1971, 2:191-2) noted differences in the island names. Each has the same arrangement, with four main islands, in a north-south arrangement, and a parallel group, of mostly smaller islands, to the west, extending further to the north. In reality, the Cape Verde Islands are formed of three separate groups arranged respectively northwest-southeast, north-south and east-west. However, most of Benincasa's names match the present ones and, with a little adjustment for position, the group's arrangement can be understood.
By contrast, the Rotterdam atlas's configuration, while roughly equivalent to the four main islands of Benincasa, adds a second group to the east of the first, not the west, and distributes the names in an arbitrary fashion. Its four main islands are labelled thus from the north (with the usual Benincasa forms in brackets: san nicollo (sal), s.luzia (bonauista), s.vizenzo (mais), s.antonio (san jacomo - though unhelpfully named 'bonauista' across the island itself, just as Benincasa himself did). The smaller group on the Rotterdam sheet are named, from the top: desal, boauista, dormaio, santta maria grande, illcos, fogo. Several of those names are normal ones misplaced; some forms are not recorded by Cortesão (1971, 2:185-92). What was the model for this? There are slight similarities in shape and relative position between the Rotterdam and Benincasa treatment of the lesser islands. But what would explain their being moved to the other side of the main group? How could the Rotterdam depiction have been based on a chart of Benincasa's, since it was inferior even to his first attempt to depict them? And yet the Rotterdam atlas is otherwise clearly based on Benincasa. There is clearly scope for further research here.
"There is no doubt that he was a specialised craftsman, for not even the most ardent sceptics deny that such regular and systematic characteristics needed the support of a well organised atelier employing permanent collaborators" (Pujades, 2007, p.497a). How much does the evidence from the new 'Colour & Shape Analysis' bear that out? [For a correction to my 1987 statement (p.431a) about signs of careless apprentice work on one of the Benincasa atlases see Attributions: Benincasa.]
A number of works have been provisionally attributed to Benincasa himself, have been assigned to a possible Benincasa workshop, seen as the work of (presumably) unauthorised copyists, or produced by his successors: his son Andrea or members of the Freducci family, or Callopoda/Sideri. How can unsigned works in his style be reliably assigned to the correct category?
The obvious points to focus on are handwriting and content. Letter forms are clearly vital here but only limited remarks can be made without access to high resolution scans of more than just a few of the relevant works. It is of no great significance that the distinctive diagonal scales appear only on atlases not charts, since most of the works for which Benincasa's authorship has been suggested are indeed atlases. A detailed toponymic study for the second half of the 15th century would certainly be valuable. In the meantime, the large array of features included in the 'Colour & Shape Analysis' means that we now have a new litmus paper for determining, first, if a Benincasan model was used and, second, if the work in question was produced by Grazioso or Andrea.
Much work remains to be done in this area. Unfortunately, the condition of most surviving charts makes it difficult, if not impossible, to study their letter forms. Nor is it possible to print out enlarged sections of the scans on the Pujades DVD, for comparative purposes.
Comments made by Pujades (2009 p.351) in the context of the 1439 Vallseca chart may explain some of the features described below: "On the 1439 chart he often resorts to techniques typical of notarial or chancellery deeds, such as the unnecessary elongation of the descending strokes on letters such as y or h, or the superfluous reinforcements on others, such as capital C or E". It is accepted that determining who actually drew a particular chart, even assuming it was a single individual, is not straightforward. In his detailed study on Gabriel de Vallseca, Pujades admits uncertainty as to whether or not the three signed charts were in the same hand (2009, pp.350-1).
The questions that need answering are these:
what are the unique elements of Grazioso's hand?
Given the limited access to reproductions of adequate quality for analysis of the hands involved, and my own lack of palaeographical expertise, only a few tentative comments can be made.
First, Grazioso's own handwriting 'signature'. Perhaps more so than his contemporaries, Grazioso liked to use different letter forms, e.g. a 'g' with either one or two loops [both versions have been seen in the same word], a capital 'A' in two different styles, a squashed or occasionally pulled-out 's', alternating short and 'j' forms for the letter 'i' (sometimes in successive versions of the same name or even within a single name), and a terminal 'm' with or without a curling descender at the end. What does seem to be peculiar to him is the way that, in the dipthongs de and ve, often only the loop of the e is preserved.
does Andrea replicate those or have his own peculiarities?
Three works of Andrea's survive: an atlas of 1476 in the Bib.Pub.et Univ., Geneva, MS Lat.81 (a single sheet of which is reproduced in Cortesão (1954), Pl. XV); a chart of 1490 (the burnt remnant of which is now in the Soprintendenza per i Beni archeologici delle Marche, Ancona - apparently unreproduced); and the 1508 chart in the Vatican, Borgia. VIII. While the following features all appear occasionally in Grazioso's work, they are frequent or habitual in Andrea's. The terminal 'a' tends to have an upcurl at the end, the terminal 'm' a backward-curling extension at the end and, at least in 1508, the 'h' extends in an s- type curve.
are any of the unsigned charts - or those truncated at the point where any signature would have been - in the hand of either Benincasa?
It is almost certain that the following was drawn by Grazioso or Andrea:
Pujades considers that the following two incomplete charts are the products of a hypothetical Benincasa workshop:
As can be seen on the Pujades DVD, each chart has been cut down on all four sides. Any Benincasa signature would have been across the neck at the left hand side and thus removed, i.e. it need not be assumed that these charts were originally unsigned. The handwriting of each is very similar to that of the Benincasas, and the toponymy of both is closely in line with Benincasa's work of the 1460s (as shown in Pujades's analysis, pp.356-73 & 392-3).
The 'Colour & Shape Analysis' shows almost complete conformity between these two productions and those single charts signed by Benincasa [see the Microsoft Word table: Colour and shape on the work of Benincasa]. On the Paris chart Chios (sio) is coloured pink, whereas on the three other Benincasa charts for which this could be noted, the colour was green or yellow. On the Vicenza chart, the colouring of the two 'islands' in the Dnipro estuary (to the west of Crimea) is not typical. However, these features were not ones on which Benincasa's colouring was wholly consistent, and those minor differences are certainly not sufficient grounds, alone, for considering either of these charts to be the work of another practitioner.
The notion that the peripatetic chartmaker could have had a workshop staffed with apprentices and/or assistants presents difficulties, except perhaps for the period between 1468 and 1480 when Grazioso may have been firmly based in Venice (though there is no clear evidence for this). We should instead assign these two charts formally to the Benincasas. If so, the Paris chart is more likely to be the work of Grazioso. Conversely, since the Vicenza chart illustrates those characteristics assigned above to Andrea, that work can perhaps be tentatively attributed to him. The names on the Vicenza chart are placed so regularly at right-angles to the coast as to give a feeling of rigidity. This might indicate the work of a trainee, perhaps Andrea, who is likely to have been an adult 'apprentice'. Grazioso, who was thought to have been born around 1410, married in 1435, the year he started compiling his portolano of the Adriatic, Aegean & Black Sea. Thus he would have been about 50 when he embarked on chartmaking, after the loss of his ship to pirates. That implies that Andrea could have been mature himself by then.
Since we can safely rule out the possibility that any chart copyist was attempting to produce Benincasa forgeries, those works that mimic his decorative and hydrographic conventions, his colour choices and island shapes, as well as his toponymy, must have been drawn by one of the two Benincasas or by someone they had trained.
are two or more of those unsigned works that are thought to be copies of Benincasa's work drawn by the same hand or can any be assigned to his known successors?
Some of the following are definitely not in the hand of either Benincasa; others are considered, by those who have examined them, not to be his work. As yet no handwriting seems to be repeated, but some works might prove to be in the same hand, once adequate handwriting samples are available for those not yet seen. For only three of the works below, the atlases in Parma and Rotterdam and the Turin chart, were adequate scans available so that they could be subjected to the 'Colour & Shape Analysis'. Where the information is available, the following are included, however, in the analysis on the various Benincasa Tables.
For the Colour & Shape Analysis findings for the chartmakers mentioned in this section please consult Colour and shape on the productions of Benincasa's successors and imitators [a Microsoft Word document]
No direct evidence has yet been found to corroborate what seems likely from the close imitation of Benincasa's work by subsequent chartmakers, namely that he left behind a chain of chartmakers, each of whom taught the next. Grazioso may himself have been responsible for training the first of these, Conte di Ottomanno Freducci, known from works dated between 1497-1539 [though for the claim that this was more than one individual see the Jesi fragment]. Grazioso's son Andrea, for whom works are known up to 1508, was occupied with public office and could not been a full-time chartmaker. Ottomanno Freducci presumably inducted his own son Angelo (fl. 1547-56). Both Freduccis signed their works from Ancona, Grazioso's home city, to which he had returned by 1480.
Astengo (2007, p.221, and p. 202, note 165), cites another Anconitan chartmaker, Rocco Dalolmo, whose sole surviving production, a chart of 1542, 'is very close to the work of the Benincasa family and the elder Freducci'. Dalolmo might perhaps have been taught by Ottomanno and (from what can be discerned in an illustration) he does seem to have reproduced the Benincasan forms for the British Isles headings and the Lacus fortunatus legend. However, from the C&SA findings it appears that Dalolmo introduced sufficient variations to the Benincasan canon to suggest that he was an imitator rather than a pupil. The full toponymy of his chart is given by Baldacci (1990).
Finally, there is the Cretan, Giorgio Sideri (known as Callapoda) who was active in the period 1537-70, working in Venice for most of that time. Having noted that his work was generally imitative of various others, Astengo (p.217) states that his 'nautical charts are clearly derived from the work of the Ancona cartographers Benincasa and Freducci', and goes on to itemise the borrowings. The listing cited as a link at the head of this section includes works by the two Freduccis and Callapoda. That analysis demonstrates how faithfully they kept alive, for 110 years (up to 1570), most of the conventions and colour choices that Benincasa had introduced in the 1460s. They also mark out these Benincasa-style charts from the work of all others active in the first half of the 16th century - at least as far as the available scans are concerned. {This paragraph altered 19 March 2014 in the light of the rediscovery of a 1570 Giorgio Sideri (Callapoda) chart in Central St Martins Library, London, which extended his period of known activity by five years}
The treatment of the larger mythical islands around the British Isles demonstrates both the stylistic borrowing and the clinging to outmoded geographical concepts. See the readings for till, scurçe, montorio/brazil and man [Nos 1, 2, 10 & 11 in the Colour & Shape Analysis]. The serrated outline given to the circular montorio/brazil is the single feature most distinctive of a Benincasan source. Otherwise it is the imitation of Benincasa's colour sequences and island shapes that betrays the debt.
The insulation of the Benincasa followers from what was going on around them is so marked that it seemed sensible to provide two 'latest date seen' figures in the Summary table of the features analysed [a Microsoft Word Table]. Dates in square brackets are the last noted occurrence of a feature on work by non-Benincasan chartmakers, while those in wiggly brackets refer to the charts of Benincasa's successors. The gap in years indicates the length of time that only the latter persevered with those obsolete conventions.
Given that these successors continued using Benincasan models for decades, might any of the unsigned productions turn out to be earlier than thought - even possibly the work of Grazioso or Andrea? Further analysis would need to be done by those with palaeographical expertise. However, there do seem to be some stylistic peculiarities in the handwriting of the Freduccis that could be investigated. For example, the upward looping extension to the 'e' (found on the 1497 and 1537 works) and, most distinctive of all, a long, straight extension to the double-looped 'g', going down and to the left (found on the 1537 and 1555 works).
Works used in the preparation of this note and in the Benincasa Tables
(for full details of the following post-1986 works see the Bibliography)
Astengo (1990c) [on the Turin chart]
Falchetta (1995), pp.54-60
Kupčik (2004)
Pflederer (2001)
Pujades (2007)
Sider (1992)
Tolias (1999)
Vagnon (2005)
Pre-1986 references
Almagià, Roberto. Monumenta Cartographica Vaticana. Volume 1. (Vatican City: Biblioteca Vaticana,
1944).
Caraci, Giuseppe. 'A proposito di alcune carte nautiche di Grazioso Benincasa', Memorie Geografiche dall'Istituto di Scienze, Geografiche e Cartografiche, 1 (1954): 283-90.
'Census'. Campbell, Tony. 'Census of pre-sixteenth-century portolan charts', Imago Mundi 38 (1986): 67-94. [Available online via JSTOR subscription (in the 'Arts & Sciences Complement')].
'Chapter'. Campbell, Tony. 'Portolan charts from the late thirteenth century to 1500' in: J.B. Harley & David Woodward (eds) The History of Cartography . Vol.1 (University of Chicago Press, 1987), pp. 371-463 [available online in pdf format since June 2011.]
Codazzi, Angela. 'Benincasa, Grazioso' in the Dizionario Biografico degli Italiani Volume 8 [1966 - accessed 22 July 2011].
Cortesão, Armando. The nautical chart of 1424 (1954).
Cortesão, Armando. History of Portuguese Cartography, 2 vols (1969-71). [Citing volume 2, 1971.]
Destombes, Marcel. 'A Venetian nautical atlas of the late 15th century', Imago Mundi 12 (1955 ), p.30.
DNA. Cumming, W.P., Skelton, R.A. & D.B. Quinn, The discovery of North America (New York, 1972).
Emiliani, Marina. 'Le carte nautiche dei Benincasa, cartografi anconetani', Bollettino della Reale Società Geograficana Italiana 73 (1936): 485-510.
Frabetti, Pietro. Carte nautiche italiane dal XIV al XVII secolo conservate in Emilia-Romagna - Archivi e Biblioteche Pubbliche (Florence: Olschki, 1978).
Kamal, Youssouf. Monumenta cartographica Africae et Aegypti, 5 vols in 16 fascicules (Cairo, 1926-51).
Kretschmer, Konrad. Die italienischen Portolane des Mittelalters: Ein Beitrag zur Geschichte der Kartographie und Nautik. Veröffentlichungen des Instituts für Meereskunde und des Geographischen Instituts an der Universität Berlin, vol. 13 (Berlin, 1909; reprinted Hildesheim: Georg Olms, 1962). [Harvard University has mounted a scanned version. You can search for any word and enlarge the retrieved page.]
La Roncière, Monique de la, Michel Mollat du Jourdin, et al. Les portulans: cartes marines du XIIIe au XVIIe siècle (Fribourg: Office du Livre, 1984) [English edition: Sea charts of the early explorers: 13th to 17th century (New York: Thames & Hudson, 1984).
Manoscritti cartografici nella Biblioteca Vaticana (Vatican City: Biblioteca Vaticana, 1981). [Notes on items 1-54, i.e. the cartographic works, by Romain H. Rainero; from a catalogue to an exhibition, June-December 1981, organized for the 9th International Conference on the History of Cartography.]
Nordenskiöld, A.E. Periplus: an essay on the early history of charts and sailing-directions (Stockholm, 1897).
