L.11. Madonna and Child with standing Male Saint holding a staff at the Left and St. Catherine of Alexandria at the right, and Other Saints?, possibly The Mystic Marriage of St. Catherine, with an Angel playing a Lute in the center below.

Late 1514 or 1515


On the Angel Playing a Lute in the Uffizi, a fragment of the lost altarpiece considered here, see P.4.

The painting from which the Angel Playing a Lute was cut could only have been an altarpiece of the kind painted by Fra Bartolommeo, Andrea del Sarto and others in the first two decades of the sixteenth century in Florence, as considered below. In a reflectograph of the small painting (Fig.Reflectograph85) part of a staff appears on a diagonal immediately above the center of the wing at the left (the angel’s right wing), as first recognized and published by Antonio Natali in 1985. He also recognized part of the wheel of St. Catherine of Alexandria at the tips of the feathers at the far right (the angel’s left wing) close to the edge of the panel. These and other details are clearer in a later reflectograph made with finer instruments in June of 1996 and published again by Natali the same year (Fig.Reflectograph96). Because of the way the two details in the reflectographs might be read as converging toward the center of the picture he suggested in 1985 a comparison with the figures of St. Onofrius leaning on his staff and St. Catherine of Alexandria with her wheel in Sarto’s Sarzana Altarpiece of 1528 (Fig.Sarto,Sarzana)1. Between two standing and/or kneeling saints there would have appeared the angel in the Uffizi panel. Above the angel the Virgin and Child would have been seen raised upon a pedestal, with the angel seated on a step below, as in Fra Bartolommeo’s altarpiece of 1509 in the Cathedral of Lucca (Fig.Frate,Lucca)2 and in his much larger Mystic Marriage of St. Catherine of Siena of 1512 (Fig.Frate,Pitti)3. The incised horizontal lines visible in the Uffizi panel that marked in the gesso ground of the panel the edges and moldings of steps and of the Virgin’s pedestal, as visible in the S. Maria Nuova Altarpiece (but not in reproductions of it). One such incised guideline appears in that altarpiece at the top of the heads of the two seated angels, exactly at the place where an incised small line appears in the Angel Playing a Lute. The two lines very close to each other that run through the angel’s right hand (Fig.Reflectograph96bis; with details marked in green lines [incisions], red lines [details from the reflectographs] and blue hatching [ornament]) could define a molding as do pairs of incised lines in the Dei Altarpiece that mark the top and bottom of the rounded front edges of the steps.4 Measuring 37.8 cm. across, the small Uffizi panel is smaller than the area occupied by the two angels in the altarpiece of 1518 (around 73 cm, wing tip to wing tip). Rosso’s single angel is slightly larger than each of the angels in the 1518 altarpiece, so other figures in the lost altarpiece would probably have been of comparable larger size. But the single angel still occupies less space than the other two, suggesting that the lost altarpiece was smaller than Rosso’s extant altarpiece. With the limited evidence available set within the usual format for an altarpiece with a single saint at either side of the Virgin and Child raised upon a pedestal, and with an angel seated in the center below, the lost altarpiece may have resembled Sarto’s lost altarpiece of around 1514-1515 with St. John the Baptist and St. Ambrose, but with the angel below holding the bishop’s mitre, known from one copy at Stokes Poges (Fig.Sarto,StokePoges). It measures 167 x 132 cm. as against Rosso’s Uffizi altarpiece at 172 x 141, a minimal difference.5

The later reflectograph shows several details not visible in the first reflectograph. Visible at the lower right is evidence of the edges of the step on which the angel is seated, and a perspectival diagonal edge at the far left suggesting something more of the architectural setting of the scene. Above the angel’s head appears what may be the draped knee and lower leg of the Virgin. To the left of the knee and just below the hand of the angel holding the arm of the lute may be two decorative heads, the upper one possibly an infant’s, the other a bearded old man’s (shaded in blue in Fig.Reflectograph96bis). If these are actually details of ornament they might suggest an aspect of Rosso’s style unlike anything appearing in the Frate’s altarpieces but seen occasionally in Sarto’s pictures and throughout the framing of the Scalzo narrative scenes.

No altarpiece is recorded by Vasari, or by any other source, from which the small Uffizi panel would have been cut. As the picture was already in the Uffizi in 1605 it is likely that it was with with Rosso until his departure for Rome, in Florence, and remained there.  If this was the case it is unlikely that the original painting, thought small as an altarpiece as indicated by the size of the surviving fragment, would have escaped Vasari’s notice. This is probably true even if the picture was elsewhere. Consequently it is possible that the destruction of the altarpiece and the saving of the fragment had taken place well before Vasari began to write the first edition of the Lives in the 1540s. However, it is also possible that the lost altarpiece was abandoned, likely unfinished and forgotten, and hence unknown to Vasari, only to be discovered around 1600. At that time the picture as a whole was not sufficiently finished or appreciated to warrant it being completed by another painter. From it the fragment in the Uffizi possibly of a finished part of the altarpiece, may then have been created, as suggested by the format that the image was finally given and by the manner in which its background was repainted, including a slightly lighter passage with a diagonal edge in the lower right corner.6 Perhaps the cast shadow on the arm of the lute and the shadowing on the angel’s left arm that emphasizes the highlight along its inner side to the elbow are additions of this moment(Fig.P.4c). A certain interest in fragments seems not to have been unusual around this time,7 although the format of the Angel Playing a Lute does suggest that it is was meant to be seen in its final state as a fragment.

There is some possibility that a hint of the lost altarpiece can be detected in Vasari’s “Life” of Rosso.  Against other chronological evidence the only mistake that he makes in the arrangement of Rosso’s Italian works is the placement of the S. Maria Nuova Altarpiece of 1518 between the Assumption of 1513-1514 and the lost coat-of-arms of Leo X at the Annunziata paid for in September 1515 (L.9).  It is therefore possible that Vasari confused two items of information, or that the information he received was unclear, causing him not to recognize the existence at one time of another altarpiece. This confusion could have resulted from the fact that there was also a controversy connected to the lost altarpiece, which because it was unfinished and abandoned was not very clearly remembered by anyone.  Nor, it would seem, were the circumstances surrounding it clearly remembered as distinct from those related to the controversy connected to the existing S. Maria Nuova Altarpiece.  The period indicated by Vasari’s mistake is precisely that in which the Angel Playing the Lute can, on stylistic grounds, be most comfortably placed.  It could be assumed that the remainder of the lost altarpiece was destroyed.

But a slightly different history is possible. Rosso’s Portrait of a Young Man in Washington (P.8), of around 1520, is, as revealed by an x-radiograph, painted over, as reported in 1994, entirely and not merely sketched, the head and shoulder of a young woman seen in profile facing left (Fig.P.8dFig.P.8e).  This female figure occupies the area where the sitter’s right hand appears, but upside down showing that the panel was originally used in the other direction (Fig.P.8f). Both this panel and that of the Uffizi Angel Playing a Lute are of poplar, which may not be usual but still has some significance in connecting these two fragments, both of which also have a vertical grain. The Angel, then, appears to have been cut from a tall plank and not shaped as an independent horizontal picture from a small piece of wood with a horizonthal grain. The profile head is about the same size as the Virgin’s in Rosso’s S. Maria Nuova Altarpiece. It is reasonable to believe that the this female head is by Rosso and that in painting the portrait he was using a panel on which he had already worked. The profile head of the x-radiograph may once have belonged to the same painting as the Angel Playing a Lute to which it seems stylistically compatible, as stated by Béguin and reported in P.8. It could be the head and shoulder of  the young and beautiful St. Catherine of Alexandria at whose feet was the wheel associated with her, part of which Natali recognized just to the right of the angel’s wing at the far right in the reflectographs of the small Uffizi painting. The female saint at the right would be looking toward the Virgin and Child, either standing in attendance or in a scene of the Mystic Marriage of St. Catherine (of Siena or of Alexandria) as painted around the same time by Fra Bartolommeo (Fig.Frate,Marriage,Paris)8and Andrea del Sarto(Fig.Sarto,Marriage,Dresden).9 On the other side of the Virgin would have been a saint with a staff, or a staff with a cross at the top as frequently held by St. John the Baptist. The width of the lost altarpiece could possibly have allowed for more saints as in the Frate’s picture in Paris, or with only two more saints as in Rosso’s own altarpiece of 1518. Although the visual evidence for the appearance of Rosso’s lost altarpiece is too scant to allow more that a glance at what it may have looked like, the  placement of these details on top of a shadow of the copy of Sarto’s painting in Stokes Poges (Fig.Sarto,StokePoges)10 may be marginally helpful, at least in showing how much of the composition has been lost (Fig.Composite). An alternative would be an image of the Mystic Marriage of St. Catherine of Alexandria, with a model like that of Ridolfo Ghirlandaio’s altarpiece of 1506-1507 at the Villa La Quiete near Florence (Fig.Ridolfo,Marriage; Fig.Composite2)11

Béguin compared the profile head of the Washington portrait to that of St. Apollonia’s in Rosso’s Marriage of the Virgin in San Lorenzo, dated 1523 (Fig.Apollonia-Catherine), suggesting that their similarities indicate the same date. However, in detail there are as many differences, like the actual line and shape of the noses, and as similarities, the flat-topped ears that are found throughout Rosso’s work. Here, however, St. Apollonia’s ear is emphatically flatter than St. Catherine’s, suggesting that this indiosyncratic detail has gone from an early spontaneous artistic gesture to an established pattern by 1523. The Washington profile should also be compared to the profile of the young and slender St. John the Baptist in the S. Maria Nuova Altarpiece, with the Washington profile of an equally young but fuller female face of what would be the beautiful St. Catherine of Alexandria (Fig.John-Catherine). Both, again, have flat-topped ears and mouths slightly open. They could be contemporary creations and earlier in date than the possibly older in appearance and vaguely mystic St. Apollonia. It should not be too quickly assumed that because these two profiles make the only comparison of two female heads to be found in Rosso’s early Florentine works that they are contemporary. Béguin herself said in regard to the bringing together of these two fragments that “Cette hypothèse n’est pas absolument infirmeé par le style.”

This second scenario would mean that the altarpiece of late 1514 or 1515 was first cut up by Rosso himself around 1520 for a section of the poplar panel to be used for the creation of the Washington portrait. He would have salvaged the Angel Playing a Lute himself, but not to reshape it into an independent picture, as the irregular edges of the original panel suggest. Then around 1600 the fragment was discovered, the panel’s thickness planed down, as suggested by Natali, the irregularly cut panel straightened by added strips, a new background painted over the scraps of original details of the altarpiece and a frame made for the evenly shaped panel. A new independent picture was created and by 1606 had become part of the scheme of the hanging of the pictures in the Tribuna of the Uffizi. This may suggest that the re-creation of the fragment into a picture of its own may have been done for this very purpose.

The discovery of the inscription of Rosso’s full name and of the date 1521 on the Angel Playing a Lute adds another chapter to the story of this much reproduced painting. Before finding these inscriptions the small painting had been dated by various art historians: around 1518 (Kusenberg, 1935), as early as 1514-1515 (Barocchi, 1950; Carroll, 1987), 1518 or even earlier (Natali, 1985), c. 1519-1520 or 1523-1524 (Franklin, 1994), and every year in between.

These newly seen inscriptions, discovered in the reflectograph of 1996, have now focused attention on a specific date and on the possibility of an origin for the work outside Florence. Falciani pointed out that the signature is faint and Franklin refers to the inscriptions as “apparently found.” Unless the small painting is further cleaned around the area of the inscriptions, possibly ruining the integrity of this much loved picture, the actual nature of these inscriptions may never be understood. Already in 1605 it was attributed to Rosso and while it was for a while given to Beccafumi and then to Francesco Vanni it was again attributed to Rosso, whose authorship of the picture has held until the present.

The two cursive inscriptions, that were mentioned in a note by Natali in 1996 and published by Falciani the same year, require considerations.12As Rosso’s name seems to read: R[u]beus florentin[?] f[?]c[?] it has been suggested, according to Ciardi’s interpretation in 1987 of the inclusion of the artist’s origins, that it was written by Rosso when he was away from Florence. The date has been read as: MDXXI. These inscriptions, not visible in the small illustration published by Natali, would likely place the making of the painting, or, rather, the addition of the inscriptions, in Piombino or Volterra. On the altarpiece of which the small painting had been a part Rosso’s name and the date would most likely have been painted on the front of a step at the center of the picture above the angel, or on the pedestal upon which the Virgin is seated. Falciani reported that the two inscriptions were discovered at the lower right near the edge of the lute. However, he pointed out that they are not aligned with the parallel incisions marking out the architectural lines found in the gesso of the panel, but instead on the border of a balustrade (balustra) on which the angel rests his elbow. In my viewing of an enlargement of Natali surprisingly good reproduction there is no balustrade and the angel’s elbow is simply situated above the top of the step that he is seated on. Pictures of lute players I have seen show the fingering arm of the player free of any support. If Falciani is correct in the inscriptions not being aligned with the lines of the steps, and hence were likely not foreseen for this placement in the original altarpiece, they would then only have been painted on the small picture when it had been cut out of the large painting. Why they became obscured is difficult to answer. Perhaps they were painted out when there were arguments over the correctness of Rosso’s authorship of the painting and of the authenticity of the inscriptions. Perhaps the inscriptions were simply a guess at the time of the painting’s recreation around 1600 when Rosso’s name was generally remarked as Rosso Fiorentino and with the date of 1521 known from the report of the inscriptions on the Volterra Deposition.

The date of the execution of the original altarpiece is the important issue now. I had dated the small painting cut from in late 1514 or in 1515. The Angel Playing a Lute would have been cut into its original shape from the central vertical poplar panel of the lost altarpiece, several other joined panels to the right would also have been taken for the support of the Portait of a Young Man in Washington around 1520 (P.8). The portrait done very shortly before the artist left for Piombino, Rosso perhaps then carried the small fragment of the angel, that he saved and kept for his own pleasure at the time he cut out the panels for his portrait, to Piombino and Volterra. He might have painted more on the panel, to finish details and to repair damage that had happened to it in his studio, signed and dated it in 1521, and brought it back to Florence to have the fragment around in his Florentine studio again. The reshaping of its irregular edges made by several thin additions to the panel, the painting out of the background and possible additions to the shading on the arm of the lute and the inner arm of the angel, and the framing of the work may have been done to prepare it for a place in the design of the hanging of the pictures chosen for the Tribuna of the Uffizi. So it was that long after his death the small painting bestowed upon the artist a significantly greater measure of fame than almost any other work by him that remains in the city of his birth.

A last possibility comes to mind related to court documents (DOCS. 3-6) about the artist’s refusal to pay a debt of eighty-six lire piccioli owed to Jacopo de Leonardo da Colle. The earliest known claim of December 18, 1517 was presented to the “Signori uffitiale et vostra Corte… nell’arte et università degli Spetiali della città di Firenze.” The authority of this court was over matters related to the guild of pharmacists (Speziali) and might be sought for disputes related to the pigments that pharmacists supplied to painters. Here the suit may have specifically been about the pigments Rosso bought for the execution of this lost altarpiece. If this work was begun in 1515 and worked on for an extended period of time, it may no longer have been wanted by whomever commissioned it. Jacopo di Leonardo, having waited long enough to be paid for the pigments he supplied, sought for a court order in December of 1517. It is possible that there has been an earlier claim to the “Corte nell’arte et università degli Spetiali.” Having been unsuccessful there, nine months later, on November 24, 1518, Jacopo, the merchant of these pigments, turned to the higher Tribunale de Mercanzia, the merchants’ court. On December 1, 1518, he tried again at the same court. By this time and already for over a month, Rosso was also involved with the refusal by the spedalingo of the Hospital of S. Maria Nuova of the altarpiece that he had commissioned of the artist on 30 January of that year. Although the spedalingo had in the end to accept the painting he did not pay Rosso the last nine florins of the twenty-five contracted for the painting. No record has yet been discovered showing that Jacopo di Leonardo was paid his eighty-six lire piccioli by the impecunious artist.13

1 Freedberg, 1963, I, 85-85, Fig. 205, II, 171-175, no. 76.

2 Freedberg, 1961, I, 196-197, II, Fig. 255.

3Freedberg, 1961, I, 204-205, II, Fig. 264.

4These incised lines are best seen in a raking light and are generally not recorded in a photograph. In the Dei Altarpiece they do not go through the figures although one can partly be seen in the drapery of the seated female saint at the lower left.

5Freedberg, 1963, I, Fig. 58, II, 47-49, no, 25, as c. 1515. Shearman, 1965, I, 46, II, 225-226, no. 37(i), dated it in 1514 or more probably early 1515.

6This detail, that may show the upper lighted surface of a ledge, is not always evident in reproductions of the painting and does not appear in the reflectographs.  It may have been added to suggest a spatial placement for the angel in an otherwise empty setting.

7Shearman, 1977, 363, makes this comment in relation to the fragments of Perino del Vaga’s altarpiece at Hampton Court.

8Freedberg, 1961, I, 201-202, II, Fig. 260.

9Freedberg, 1963, I, 19-21, Figs. 38-39, II, 34-35, no. 16.

10 See n. 5.

11For Ridolfo Ghirlandaio’s altarpiece, see DeBenedictis, Cristina, ed., Villa La Quiete, Il patrimonia artistic del Conservatorio delle Montalve, Florence, 1997, Color Plate, 184-186 with Fig.The composite images were made by Peter Charlap, with the superimposed drawing in red, green, blue and black by me.

12Natali, in Empoli e Volterra, 1994 (1996), 169, 170, ns. *, 1, 214, Fig. 93; Falciani, in L’officina della maniera, 1996, 350-351, no. 129, Fig. 129.

13I have been able to find one Jacopo di Leonardo da Colle who in the summer of 1390 was appointed a new trumpeter for Siena on July 6, 1390. But he seems never to have taken up his post. See Frank A. D’Accone, The Civic Muse, Chicago and London, 1997, 429, 767. That he was the same “cittadino fiorentino” who, twenty-seven years later, and now a pharmacist, brought Rosso to court in Florence, in unlikely, but possible enough to be worth reporting.