A FEW YEARS ago I published an essay in Art and Antiques on 'Winslow Homer's Mystery Woman', in which I speculated about Homer's relationship with a red-haired woman who appears in a number of his paintings and water-colours of the 1870s. Since the article appeared, information has come to light which demonstrates that most of my con- jectures were incorrect. It is worth recapitulating the 'mystery woman' theory before indicating just where I went astray.

A TINY sketch of 1879 (Fig.18)' is the first witness to the spatial and compositional preoccupations that were to dominate Degas's classroom pictures for almost two dec- ades. Degas produced some forty pictures in an elongated, frieze-like format: some of them are race-course scenes, but most show ballet dancers in a classroom setting. The ballet friezes may themselves be divided into two groups. Some, which we shall call the rightward friezes, start from a shallow foreground on the left, backed by a wall receding steeply towards the right; further to the right, this wall has a setback that widens the room (Figs. 17, 21, 22, 23, 24 and 33).2 The leftward friezes have a similar but reversed compo- sition: they depict an L-shaped space that opens out from the shallow lower right to a widened background on the left - the gain in width offsetting the narrowing perspective (Figs.25, 32 and 34).

I. 'Olympia'

WE now know that Manet's Olympia did not appear at the 1865 Salon with the first stanza ofZacharie Astruc's poem La Fille des Iles affixed to the frame.I A particularly stark juxtaposition of image and text has been metamorphosed from fact to mythology. It has also been shown that Astruc's verse did not even appear in all editions of the Salon catalogue.2 The link gets weaker as patient research progresses. Weakened as it is, however, the connexion remains, augmenting the rage of critics in 1865 and attracting the attention of scholars ever since. Astruc's words appear in the catalogue, above his name, as follows:

Quand, lasse de songer, Olympia s'eveille, Leprintemps entre au bras du doux messager noir; C'est l'esclave a la nuit amoureusepareille, Qui vient fleurir lejour ddlicieux a voir: L'augustejeunefille en qui laflamme veille

MENTION has been made of 'L'Acadimie Vitti' in some biographies of Paul Gauguin, most recently in the catalogue of the great Gauguin exhibition of 1989.1 Gauguin's instruction at the academy has been dated to the beginning of 1890, to a period when he had returned to Paris from Britanny and was staying with Emile Schuffenecker. The artist mentioned this episode briefly in a letter to Daniel de Monfreid of May 1899, referring to 'le peu de temps queje corrigeais a l'atelier Montparnasse . . .'

SEVERAL of Paul Signac's letters which have surfaced recently - in the catalogue of a British dealer' and in the family archive of the French lithographer, Auguste Clot -- suggest there is a need to revise the information in the catalogue raisonne covering his colour lithographs.