Brilliant vision: the making of Van Dyck’s dazzling portrait of Rachel de Ruvigny

Universally acknowledged as one of the greatest exponents of the art of portraiture, Anthony van Dyck’s portrait of Rachel de Ruvigny, the Countess of Southampton (1603-1640) is a superlative example of the skill and panache of the Flemish painter.

During his short but brilliant life Van Dyck (1599-1641) frequently criss-crossed Europe, however his career can be divided into four main phases: his artistic apprenticeship under Rubens in Antwerp (up to 1620), six crucial years in Italy, where he absorbed the art of his great hero Titian (1621-27), a successful return to Antwerp (1627-32) and finally the triumphant culmination of his career in England (1632-41), to which the NGV portrait belongs. From his earliest years as a child prodigy he was recognised for his outstanding talent and flamboyant personality, qualities that are abundantly on show in this bold portrayal of Rachel de Ruvigny. The painterly brilliance of this portrait was an inspiration to later British artists including J.M.W. Turner, who was captivated by the crystal sphere by Rachel’s side. This particular work was so famous that there are now at least ten copies and variants of it in existence. All of the copies of the painting are based on the NGV’s portrait, however one important variant with the same dimensions is now in the Fitzwilliam Museum in Cambridge. For a time, there was a debate as to which of the two portraits was the primary (original) version, which was only resolved after the NGV painting was restored in 1998-99. Seen together, they provide an illuminating example of van Dyck and his practice during his last years and illustrate the some of the differences between primary and later versions of the same portrait.

The version of the portrait in the Fitzwilliam Museum, Cambridge, United Kingdom

The materials and working methods of van Dyck have been well known for centuries, due to the great fame he enjoyed during his lifetime as well as his contemporary chroniclers leaving us with a good record of his working practice. The French banker and collector Everhard Jabach (1618-95), who had his portrait painted by van Dyck around 1636 – about the same time that Rachel’s portrait was painted – vividly described the painter’s busy working day in which his patrons would visit his studio for a series of prompt one hour-long appointments. During these sittings the artist would concentrate on painting the head, and would then make a black and white compositional sketch on grey paper outlining the sitter’s pose and clothing. He would arrange for the sitter to leave a set of clothing to be included in the portrait which he could work on in his own time. Parts of the preparatory work for the rest of the image were delegated to his studio assistants. Van Dyck refined and finished off the painting, leaving the hands to the last, asking his young assistants to pose as hand models – which is why the hands in van Dyck’s portraits sometimes appear to flatter the sitter. Jabach’s account inevitably invites discussion as to the involvement of assistants in van Dyck’s work, however for his most important clients van Dyck was expected to be wholly responsible for the execution of the work.

Detail of the head from the NGV version of the portrait during conservation treatment in 1998

Jabach’s description about how van Dyck’s paintings were constructed played a role in resolving the debate about which of the two portraits was the primary version. A number of important details in the execution of the head were revealed in 1998 when old restorer’s retouchings were removed. They showed a distinct line of separation between the head and the rest of her upper body that had been masked by the pearl necklace. Moreover, van Dyck had daubed small white sketch marks to locate the positioning of Rachel’s necklace, teardrop earrings and hair-restraint. All of these preparatory details are signs that the portrait was painted directly from life. Marks like these are rarely if ever found on replicas because the whole image can be transcribed seamlessly without having to be first marked out in this relatively crude way. These preparatory markings were not present in the Fitzwilliam version. In 1999, the two paintings were brought together in London for the first time in over 350 years for a private viewing before a large group of British art scholars. The matter was resolved with the consensus that the Melbourne portrait was the primary version.

The two paintings are viewed together in London in August 1999

The Cambridge painting remains an important variant because of the unique modifications made by van Dyck– namely the appearance of a foot resting on the skull in the lower part of the painting, and a gold sceptre topped with a phoenix, which was originally planned for the NGV version but was removed by van Dyck. A likely explanation for the changes is that a few years may have passed between the first and the second, with the critical intervening fact that Rachel died during childbirth in 1640. The added details on the Cambridge version appear to refer to her triumph over death, making it a posthumous adaptation of the earlier Melbourne work which was undoubtedly painted from life.

Another important aspect to learn about the NGV portrait from another copy of the painting is that the majestic satin dress worn by Rachel was almost certainly once a darker shade of blue. A small enamel replica of the portrait was made by the French artist Jean Petitot in 1642, a few years after the original was painted. Every aspect of the original portrait is copied in exquisite detail, yet the tonal difference in the blue drapery is striking. A mezzotint engraving of the portrait by James MacArdell from 1758 (also in the NGV Collection) suggests that the dress had a more muted tonality than we see now. If this is the case, then some part of the upper layer of indigo and azurite blue has been lost. The likely reason for this is that van Dyck was known to mix his blue pigments not with a standard oil medium, but with water-based media such as egg or animal glue, which could have made the blue areas of his paintings especially fragile.

Enamel copy of the NGV portrait by Jean Petitot in the Devonshire Collection, Chatsworth House, United Kingdom

An important contemporary historical record account of the techniques of famous artists known as the de Mayerne manuscript was compiled by a prominent expatriate Swiss physician, Theodore de Mayerne (1573-1655), based on his conversations with van Dyck and other prominent artists working in London. De Mayerne provided information about van Dyck’s use of paint media, specifically for his blues, for which he used water-based mediums such as gum water, fish glue, animal glue or egg white. Scientific analysis has shown that he also used egg yolk and, on occasion, linseed oil.

The practice of avoiding linseed oil for blue pigments was already well established in Italy by the sixteenth century. It was done in an effort to preserve the cool tonality of the blue pigment which could develop a greenish tinge if the paint medium or covering varnish was prone to yellowing. An unwanted consequence of this practice may be that these passages of van Dyck’s paintings were more vulnerable than the rest of the painting and might have suffered from overcleaning in the past. If this is the case then some of van Dyck’s original colour and tonal harmonies may be lost, yet paradoxically the ravishing dress and the painting still dazzles the eye in a way that remains faithful to the artist’s intent to inspire awe in the viewer with his unique talent and flair.

Notes