Peter Paul Rubens – Holy Family
Fabio Noli 30 November 2018


“Carrying this letter is Pietro Paolo the Flemish, my Painter, that I am sending there [to Rome] to copy and paint certain drawings, as more diffusely You will wish from him.”

This letter of recommendation written on July 18, 1601 by Vincenzo Gonzaga, Duke of Mantua and addressed to Cardinal Alessandro Peretti Montaldo in Rome, metaphorically and physically opened the doors of the eternal city for Peter Paul Rubens’s first and short Roman stay.

The young painter’s art had already reached a very notable level: after formally becoming a “master” in Antwerp, he did not come to Rome disoriented and looking for protection; he was sent there for work and for a specific purpose by a nobleman, Gonzaga, who would continue to pay his salary and wouldn’t request him to go to Spain until the following year.

It was, in all likelihood, in Rome, during this first stay, that Rubens began to shape his extremely original iconographic type that is the “Holy Family with Saint Elizabeth, the Infant Saint John and a dove”.


AII79423 Holy Family of the Basket (panel) by Rubens, Peter Paul (1577-1640); Palazzo Pitti, Florence, Italy; Flemish, out of copyright.

From the mid 18th century on this composition, set against a romantic backdrop, was known in print. Along its lower edge is the phrase: “Cum essem parvulus sapiebam ut parvulus” (from the letter of Saint Paul to the Corinthians 13,11). These printed versions were doubtfully attributed to Rubens and were put in relation with other paintings from the painter’s Italian period (especially Rubens’s “Holy Family”, now in Palazzo Pitti).

It wasn’t until 1893 that H. Hymans reported (publishing only in 1927) a painting with the same subject. In 1925 the painting was in London with A.L. Nicholson where it then passed in the same year to William H. Moore’s collection in New York. In 1955, the painting was donated by Mrs. Ada Small Moore to the Metropolitan Museum of New York, where it still is (oil on panel, 66 x 51.4 cm).

In 1927-1928 the painting was first displayed to the public in an exhibit held at the P. Jackson Higgs Gallery in New York: on that occasion, Valentiner suggested its dating to 1602-1605, based on the structure of the composition; this dating, however, was not agreed upon by a number of critics: both E. Simple and F.E. Washburn Freund, in their reviews of the exhibition, strongly claimed the painting should be dated to a later period. Washburn Freund, in particular, pointed out the contradiction between the composition, that appears inspired by Italian models (and therefore is such as to logically justify Valentiner’s suggested dating), and the painting style and technique, that appear to belong to a more mature time.
In 1930, L. Burchard made the second painting public, which is now displayed at the Los Angeles County Museum as a Rubens original dating to 1609 ca.


3588514 The Holy Family with St Elizabeth, St John and a Dove, c.1609 (oil on wood) by Rubens, Peter Paul (1577-1640); 167.6×149.9 cm; Los Angeles County Museum of Art, CA, USA.


At the beginning of the century, the painting (oil on panel, 138.4 x 120.7 cm) was in the Count of Kospoth’s collection in Briese bei Ols, Silesia; in 1930 it was purchased by Berliner Gustav Nebehay and it later became a part of Gustav Oberlander’s collection in Reading (PA), where it remained until 1947.

In 1936, the two paintings were displayed together for the first time, in Detroit; based on the direct comparison on that occasion, the Oberlander panel was deemed to be the final work, while the Moore painting was its model; both works were dated to 1609-1610. As for the print, E. Siple, in her review of the exhibition, stated that it was taken from the larger painting, although certain details of the drapery seemed to follow the smaller one. In 1946, the Oberlander version was sold on New York’s antique market; from there it entered New Yorker Frederick Mont’s collection and was displayed in an exhibit organized by Valentinier in Los Angeles, where it was shown alongside the painting we are presenting in this auction.


3928815 The Holy Family with Saint Elizabeth, Saint John, and a Dove, c.1608-9 (oil on wood) by Rubens, Peter Paul (1577-1640); 66 x 51.4 cm; Metropolitan Museum of Art, New York, USA;

Later on, in 1952, the work became a part of Ch.F. Sandborn’s collection in Los Angeles; following his death, it was acquired by the County Museum thanks to the Colonel and Mrs. George J.Denis Fund.

As for the autography of the painting in the County Museum, critics (starting from Valentinier, who defined it “one of the most important compositions after his return from Italy”) had often expressed an unanimous positive opinion; after a new cleaning, shortly before 1948, the critics had once again reaffirmed its autography.

However, recently, the piece has been demoted to a copy and this seems to be a final conclusion.

The Metropolitan Museum painting is a whole other story: up to recent times, it was mostly considered to be a model, although an excessively refined one at that, for the County Museum painting, and was generally dated to after Rubens’s return to Antwerp between 1608 and 1610

But others, and especially M. Jaffé among others, believed it to be a simple copy of the larger one: notably, the Metropolitan curators themselves classified it as “copy after Rubens” in 1980.



In 1984, Liedtke, observing that “although the quality of the picture is inconsistent, there is no indication of more than one hand having produced it”, therefore identifying it as a “discontinuous” model in its parts, and noticing that the panel was particularly large for an unfinished “modello”, classified it as “attributed to Rubens”. Scribner seconded the attribution, hypothesizing that one if not both panels should, however, be attributed at least in part to assistants; Scribner thus believes that the assistants, starting from the first draft (the Metropolitan panel), then enlarged its proportions in the Los Angeles version, to which Rubens then added some final touches.

Liedtke spoke again on the subject in 1992, reaffirming the opinion expressed in 1984, and underlining the stylistic comparisons with Rubens’s “Virgin adorned with flowers”, now found in Vaduz in the collection of the princes of Liechtenstein. He believes the New York version to be autograph, and despite its executive inconsistencies, he is inclined to consider it a finished work (to keep or sell) rather than a model. At the same time, Liedtke believes the Los Angeles painting to be a later autograph redaction, perhaps made for a church.

Since the Los Angeles version is now considered to be a copy, since 1995 the Metropolitan curators have classified their painting as original. This has also entailed reconsidering the painting’s status: no longer believed to be a model for the Los Angeles piece, it has become a standalone work, probably, due to its small size, destined to the Antwerp bourgeoisie.

The painting we are presenting in this auction (oil on panel 102 x 78 cm) was purchased in Rome around 1910, and traditionally comes from a branch of the noble Roman family of the Colonna princes. Since then, it has always remained in private collections: first in the US, where it had arrived in the period between the two wars, and then, thanks to an Italian collector, it returned to Italy and reached the hands of its current owners by inheritance, after almost 80 years in which it was never put up for sale on the market.



The painting was first displayed to the public in the exhibit curated by Valentinier in Detroit in 1946, where it was presented as a Rubens original, datable to 1607 ca. and thus to his second Roman stay.

In October 1946, the painting was examined and studied by Rodolfo Pallucchini and Antonio Morassi, both acting independently: both confirm the attribution to Rubens’s hand, with interesting stylistic observations. Despite being displayed in the 1946 exhibit, the painting, undoubtedly due to the rarity of the exhibition’s catalog, remained unknown to scholars, who never mentioned it.

In fact, it was never published until 1985, where it was presented as an original and dated to 1605 ca. by D. Bodart, who pointed out its “stylistic affinity with the world of Caravaggio-esque painters, especially with Orazio Borgianni” and declared its provenance from the Colonna family, underlining how this confirmed the hypotheses of contacts between Rubens and Cardinal Ascanio Colonna, protector of Flanders, who at that time had employed Rubens’s brother Philip as a librarian. 

In 2001, the painting underwent several thorough examinations and multispectral analyses that helped understand its conservative history. 



The painting is in good conditions: it underwent a restoration, probably in the period immediately following World War II, during which the panel was thinned and reduced to a thickness of 0.25 cm and then thickly boarded. For this reason, the rear side of the panel does not hold any real information about its history.

From a technical point of view, it can be seen that the base coat is swiftly spread, and that in several spots the preparation is left in view so as to give the idea of a clear transparence and a dark depth, which would have been lost with a uniformly-spread dark color. Using preparation instead for background color is a practice that is well-known to Baroque and 18th century painters. The earth in the preparation refract the light in a subdued manner and make the dark backdrop into an almost hollow space, ready to welcome its characters. This technical expedient, which testifies to a swift and expert painting technique, is lacking in the other two versions, in which the color black is uniformly and compactly spread, a sign of the fact that, quite likely, the author was following a model and not an original idea to bring to life.



In 1997, the panel we are now presenting underwent a quick cleaning that removed the old coat and repainted parts on the background. Just this year (2018), however, the painting underwent complete restoration and cleaning, that brought other details to light that had been missed in 1997. For this reason, the painting has never been seen nor published in its current state and appearance.

The UV fluorescence has highlighted the presence of coating in certain points where integrations were made during the old restoration. The most significant are in Joseph’s beard, on his left forearm and on part of the Infant Saint John’s head.

The examination has brought to light the most significant aspect in the painting’s technique, that also draws precise lines around its history and its likely derivation from the master’s hand.

Beneath the paint, the panel reveals oblique radial lines that make up a sort of perspective scheme within which the characters and their parts were to be placed. Also visible beneath the paint are parts of the underdrawing, especially in the faces of Mary, John and Saint Elizabeth.

There are countless areas that show clear changes of heart and alterations to the figures:  Jesus’s right cheek wasn’t initially so pronounced and neither was the lock of hair at the nape of his neck.



The Infant Saint John’s right leg was poised differently. Joseph’s figure may have been sketched with a different height, and Saint Elizabeth’s back as well. 

The underdrawing is clearly visible on the Virgin Mary’s face, while her left sleeve used to be longer.

The hands of all characters were changed throughout; the most significant changes are the Infant Saint John’s left hand, inscribed in a circle as to make its drawing easier, the Virgin Mary’s right index finger and the dove’s left wing painted “a corpo” over Jesus’s left arm and Elizabeth’s cloak. This is another sign of the authenticity of the composition as, if it had been based on a model, it would not appear so reworked.

Another significant reconsideration, a sign of how undecided the painter was on the characters’ static, is the visible preparation of fingers on Baby Jesus’s left side.  This early version of Mary’s gesture would have fully reflected the tradition, dating back to the Gothic period and carried on throughout the Renaissance, according to which the mother seems to envelop the child when He is in her womb. 

Baby Jesus’s left leg was reworked, and the left hand is also probably a second draft of a hand that seemed to be holding up the veil on his side (in the Metropolitan painting this veil is golden, embellished with ornamental motives, clearly thought for its position since the beginning).

Jesus’s hand, instead of holding the veil that has been “passed” to Mary, serves a different purpose: it is indeed clear that the dove isn’t being passed from the Infant Saint John to Jesus, rather he is tearing it away from his cousin’s grasp.

This gesture significantly changes the overall meaning of the painting. In fact, the dove might represent a sign of the future sacrifice that Jesus Christ was preparing to make. The action going on between the two figures is therefore not just a playful one (as some scholars have pointed out), but an example of sacrifice and responsibility that Christ is offering the viewer, bearing the burden of his own sacrifice for the sake of humanity.

The depiction of the dove in Jesus’s hand or otherwise depicted with Baby Jesus associates this idea with Orazio Borgianni’s painting “Holy family with Saint Elizabeth, the Infant Saint John, the dove and an angel playing the violin”, today kept in Palazzo Barberini. In this composition, however, the Infant John is offering the dove from his own hands to Jesus, explicitly anticipating his role as the precursor of Christ, as the person who offers him grace and recognizes the superiority of the Messiah’s role and mission over his own.



In Rubens’s composition, however, the Baby Jesus is snatching the dove from his cousin’s hands, with near arrogance, as the cousin tries to hold on to it, barely managing to rip off a few feathers.

These dove plumes that remain in John’s hands may be interpreted as a symbol of the eloquence unequivocally bestowed upon the Baptist by the Holy Ghost.

In the two American versions, some feathers can be seen falling around the two infants, which is something that would likely happen if such a scene were real. Although this is an extra detail for the painting, it is also true that it partially steals focus from the dove’s symbolic meaning, as the dove is not just a bird whose feathers are being ripped off, but holds a precise religious meaning in its every part.

This would also reconnect to the phrase from Paul’s letter to the Corinthians that Rubens had chosen to accompany the print: “When I was a child, I talked like a child, I thought like a child, I reasoned like a child. When I became a man, I put the ways of childhood behind me.” The two infants, thus, compete for the dove since their young age, but Jesus, being aware that he is the one that shall bear that burden, makes a gesture that is not a childish one: sacrifice himself for others, fully conscient that only like this he is fulfilling the Father’s will.



Borgianni’s panel seems to “normalize” the invention of the type of gesture that Rubens’s infants make, hinting at the awareness of future events that both characters share.

As an example of the iconographic success of the scene painted by Rubens, it is worth mentioning a painting by Carlo Saraceni, depicting “Saint Anne passing the dove to the Baby Jesus sitting in Mary’s lap”, which is also kept in Palazzo Barberini in Rome.

Lastly, the reflectoscopic analysis has brought out, in the bottom right corner approx. 15 mm below the foot, towards the right, an alphanumeric print in extremely small characters, interpretable as an upper-case “P” followed by four digits that are readable as “1602”. These five elements are all well aligned and evenly spaced: the reciprocal sizes of the components are consistent, it is therefore absolutely certain that it is a form of print and not casual stains in the pigment or flaws in the wood. The print is completely invisible to the bare eye, as it is covered by the brown backdrop.  Some PR-printed microframes have revealed a scarce chromatic difference between the print and the original backdrop, the earth-based preparation: this leads to think that the print was traced using the same pigment as for the backdrop, tone on tone. A microphotograph taken with UV fluorescence shows that the letter readable as “P” has clearly been painted over, while the digits have not: they seem to be contextual to the original backdrop and don’t seem to have been rewritten or manipulated: this goes to prove their authenticity and makes them a documentation of paramount importance for the painting’s dating. At the time, it is not possible to establish whether the letter “P” was, in an undefined time, added to the pre-existing characters (thus limited to the series of numbers) or if someone just repainted an existing trace (of a “P”). What’s certain is that, unlike today, the whole print had to be clearly visible in the past, enough so to make it possible to add a letter or paint over it. 




This leads to the conclusion that, at least in that spot, the pigment is made up by at least four layers: the earth-based preparation, the print, the possible addition of the “P”, the covering that is now visible.

It is also possible that the print was traced not with a brush but with graphite: in this case, it would still be over the preparation layer and not under it. From the above, it can be presumed that at least the digits are authentically contextual to the realization of the painting. The numbers “1602” are, logically speaking, interpretable as a date, which may also refer to the period when the painting was made, therefore placing it not during the second Roman stay but during the first.

It is therefore a small-sized date that the author placed, tone on tone, on the brown preparation layer, with the intention of not making it excessively visible, as if it were a note, a memo meant more for himself than for the public: it is logical to think that only a remake of the backdrop at a later date would have covered it.


The comparison with other paintings by Rubens can show how some elements return with more mature features in the Master’s later production: in the Genovese altarpiece for the Chiesa del Gesù e dei Santi Ambrogio e Andrea, depicting the “Circumcision” (1608) or in the “Madonna Adored by Angels (Madonna della Vallicella)” (1608) in Rome, for example, the face from the Virgin Mary in our panel.

It is possible to make various comparisons as to the poses, the drawing and the execution between the hands of other characters in the painting and other paintings from Rubens’s first Roman stay: for example, we can compare Saint Sebastian’s left hand in the “Martyrdom of Saint Sebastian”, now at the Galleria Corsini, with the hand of the Infant Saint John grasping the dove’s feathers; Saint Sebastian’s right hand with Saint Anne’s right; the kneeling angel’s left hand as he unties the Saint’s feet with Joseph’s right.

Another peculiar element that may link our painting to Rubens’s paternity is the fact that in one point on the panel it is possible to observe one of those tiny allegoric figures, almost playfully hiding in the composition, that some have chosen to see in the Flemish painter’s work: specifically, below Mary’s left eye someone has suggested seeing a rooster, like in “Susanna and the elders” from Galleria Borghese a fish-like figure hides between the character’s legs, or in the famous “Portrait of Suzanne Fourment” (the older sister of Rubens’s second wife) an owl is painted by the left arm, or in the “Nativity” in Fermo a lion skin is clearly visible on a shepherd’s shoulder, never reproduced in the following versions of the famous work. 

The last element to consider, when assessing not just this painting’s authenticity but especially its temporal priority compared to the American panels and especially to the New York panel which is considered to be the oldest of the two, is the detail of the frame. This important frame, in fact, certainly contemporary to the panel and richly carved, frames the figures exactly where they are depicted in the ex-Moore painting.

Joseph’s cloak indeed is a few centimeters longer in the Roman version that we are presenting, ideally closing the circle created by the cloak in the lower part of the figure. Thus, in the same way, the upper edge of the painting, that somewhat “smothers” the scene in the American painting, in this version gives the composition more air and verticality. To the right, Mary’s dress doesn’t exactly finish at the panel’s edge, and also the crib and the fabrics in the lower part are not, in fact, cut off but end in a soft curve. It hence seems that the smaller proportions in the New York panel have been somewhat received by a painting that didn’t completely leave room to the composition but rather closed it within a frame.

Fabio Noli