Arts: Fairy, fairy, quite contrary

Exhibitions: The Royal Academy has rediscovered the Victorians' love of little people. But the show suggests that sprites make better subjects for stories than paintings

It's Largely whimsy and nonsense, but there are three or four things that I like about the Royal Academy's "Victorian Fairy Painting" exhibition. First, it is good to be reminded that Richard Dadd was an exceptional artist. Secondly, Arthur Rackham's illustration prompted me to reread Puck of Pook's Hill, an extraordinary book and all the more queer because it is about being English and yet seems unEnglish: the result no doubt of Kipling's magical Indian background. Thirdly, the show has returned Estella Canziani's The Pipe of Dreams to the public domain. Its former reputation is intriguing, while the picture itself is worthy of respect.

Most of all I appreciate the catalogue. This is one of the times when an exhibition of little visual interest comes with a publication of permanent value. It's the work of a number of authors and has been edited by the RA's Jane Martineau. One of its essays, on "Victorian Fairy Book Illustration" is by Pamela White Trimpe, who comes from the University of Iowa Museum of Art and was the instigator of the exhibition. Iowa seems an odd place to find academic fairy studies, but Trimpe has done her work well and so have her fellow writers. Detailed, knowledgeable and illuminating, the catalogue is the best book on 19th-century British art to have appeared for some years.

If only the quality of the painting were higher. This show presents us with inescapably minor art. The fairy literature of the time - from Scott through Ruskin, Christina Rossetti, Charles Kingsley, George Macdonald, Kipling and Yeats - was far better. The word is more suggestive than the brush. Fairy paintings are nearly always banal. Although the Victorians produced so many of them, I think they went against the spirit of the age. The central virtue of Victorian painting in general, and of Pre-Raphaelitism in particular, was that it represented real people in believable contemporary situations and did so with honesty, clear observation and social concern. Fairy painting could do none of these things. It could not be sturdy and least of all could it be truthful.

Fairy painting might be called a sub-branch of the Victorian symbolist tradition (amply documented in the current exhibition at the Tate), but I think the RA show correctly associates them with theatrical art at the beginning of the century and with illustration at its end. The theatrical paintings are mainly concerned with A 8Midsummer Night's Dream. Oberon and Titania, Bottom and the rest of them are imagined by John Lamb, Thomas Grieve, Landseer, Maclise and others; and all these paintings are terrible. They are false, arch, ill- designed and badly lit, and they cannot find a convincing scale for the people or personages that they represent. An embarrassing truth about British painting is that Shakespeare seldom inspired artists. We know dozens of interesting, charming or informative pictures taken from his plays, but no major ones.

Lack of sensitivity toward Shakespeare approaches its nadir with interpretations of the The Tempest. Ariel (who should properly be regarded as a spirit rather than as a fairy) is maligned in a lithograph by Richard James Lane and in paintings by Henry Singleton and Maclise. It's quite possible that Ariel's beautiful nature is beyond pictorial art, but why should such a spirit have produced such ugly pictures? An unexpected result of this exhibition is that it makes one want to read books rather than look at art. It's as though painting were feebly aspiring to the condition of literature. And this is the reason why the works on paper in this show are more successful than the oils on canvas. They were frankly, indeed abjectly, illustrations to books.

They also got better as the century went on, in line with more sophisticated publishing and colour printing. Arthur Hughes's illustration of William Allingham's fairy poem, for The Music Master, is nice. However, it is of 1855, so the wood engraving is stiff. I also like Richard Doyle, so rich in fun, but suspect that the best medium for fairy art was colour lithography within books of the 1890s. By that time a more aesthetic elfinism was offered by publishers to children.

So another curious thing about "Victorian Fairy Painting" is that it's not a show for children, or anyway not for today's children. Christmas comes, does it not, with grave and fearsome step. There will be no fairies for children at my table. They will be given chips and a couple of roast coot and can eat with their fingers while watching TV in a different room. That's what they want. It's obvious that TV culture has finally put an end to any child's half-belief in fairies and Fairyland. But when did Fairyland first disappear? This is the question raised by Estella Canziani's The Piper of Dreams.

She was born in London on 1887 to an English mother and an Italian father. The Piper of Dreams (1914) feels like a painting but technically is a watercolour heightened by body colour. It did not go with a book. But the Medici Society realised that it was a kind of illustration. They bought reproduction rights and sold no fewer than 250,000 copies in one year. Amazingly, the image became a talisman for British soldiers in the trenches. Perhaps it spoke of some kind of England, not an accurate one, that is also evoked by Puck of Pook's Hill? Did Medici prints go to soldiers at Christmas more than at other times? So that this excellent goblin replaced the useless Christianity of Christmas cards? The picture has been in private hands for many years. Given its national importance (and also because I rather love it) I would like to see it in the Tate one day.

! Royal Academy, W1 (0171 300 8000), to 8 Feb 1998.