Art historians are by their very nature people with a highly developed eye for detail, with access to sound research facilities and normally featuring a commendable desire to achieve accuracy in their written work. But as far as descriptive accuracy on dogs portrayed in art is concerned, this eye for detail has a very definite blind spot. Time and time again, in art gallery catalogues, historic house guide books, fine art auctioneers' brochures and museum captions, all sorts of elementary mistakes are made in identifying dogs, especially specific breeds, in paintings, on pieces of sculpture and other works of art.

 The two most common errors are easily identified. First there is superficial uncorroborated research, mainly relying on dog books by Victorian writers who copied each other's mistakes and knew little of foreign breeds. Prolific writers like "Stonehenge", "Idstone", Youatt, Stables, Rawdon Lee and Drury, whose books contain quite elementary errors, seem to be used as authorities again and again. Reliable writers like Ash, Leighton and Watson are sadly overlooked. This results in the words of quite distinguished art historians misleading the public rather too often over many canine studies in works of art.

 Secondly there is the tendency to identify dogs portrayed in ancient art as contemporary breeds of dog, breeds which were only created or stabilised in the last hundred years or so. This tendency is assisted by overzealous breed historians who out of a desire to bring ancient lineage to their breed, can see look-alikes in every art gallery in the world. Of course dogs like to-day's greyhounds, deerhounds, springer spaniels, bloodhounds, basset hounds and mastiffs have existed for many centuries but more as breed types than precisely like contemporary pedigree breeds. Lesser known breeds from overseas portrayed in art are quite often given the name of the nearest British breed of dog in appearance.

 In this way Dutch paintings depicting the koikkerhondje, (the decoy dog of Holland and a most valuable animal before the advent of firearms), almost always end up in British descriptions with their being referred to as toy spaniels or even Cavalier King Charles spaniels. This tendency is repeated in descriptions of dogs painted by Stubbs. Dutch sporting dogs like the Drensche Patrijshond portrayed in Dutch paintings are nearly always described as English Setters by British art 'experts'. This is not just misleading but insulting.

 Every small spitz-type dog Stubbs painted is called a pomeranian and each medium-sized grey-haired spitz is called a keeshond. But in 'The Dog Turk', 'Turk and Crab', 'The Prince of Wales's Phaeton' and 'John Christian Santhague', German not Dutch bargedogs are portrayed, with 'Turk' being a wolfspitz, the grey variety. The three sizes of the German spitz are delightfully depicted in 'The Prince of Wales's Phaeton': gross-spitz on the left, klein-spitz in the middle and mittel-spitz on the right. A wolfspitz is also portrayed in 'A Prancing Horse with Two Dogs'.

 The charming little white dog in Stubbs's 'Dog and Butterfly' is usually described as a pomeranian; but the painting displays most accurately the Italian toy breed, the volpino or loulou d'Italie, which has a longer back and less closely carried tail. The volpino is an ancient breed, long prized by the Italian nobility and often given by them to notable visitors as prestige gifts.

 Similarly, Stubbs's 'A Rough Dog', never properly identified, looks to me remarkably like two old related French herding breeds: the labrit (probably named after the town of that name in the south of France, although since it is like a briard in conformation, some believe the name comes from 'de la Brie') and the berger de Bresse, now almost certainly extinct. The leading dog in Stubbs's 'Gamekeeper with Three Dogs' has been called a mongrel and a bloodhound; but it more closely resembles the old southern hound type of harrier. Blue-mottled and linked with the hounds of Gascony, the Hailsham harriers displayed the longer leathers and shorter stature of this type, which Stubbs very accurately captured.

 But genetic impossibilities also feature in the art historian's canine identifications. At Dunham Massey, a National Trust property in Cheshire, there is a painting by Knyff of a compact short-faced brindle dog which is described in the official guide book as a pug. But a pug cannot be brindle. Chestnut brindle pugs have been recorded in China, where the breed originated, but not in England. Pug/bulldog crosses (used to shorten the face of the bulldog) have been recorded however and brindle does feature in the bulldog coat colours. The dog depicted at Dunham Massey is much more "of bulldog type" and should never be described as a pug.

 At Shugborough, the National Trust property in Staffordshire, the official guide book there describes a William Webb painting in the mansion house as depicting 'Two Dogs and a Dead Pheasant'. This painting however displays an outstanding portrayal of a black pointer and an excellent illustration of the mainly white setter of that time (1822), a breed now called the English setter. To describe such a valuable record of these gundogs as 'Two Dogs and a Dead Pheasant' is very inadequate.

 In the Louvre there is a Decamps painting entitled 'Bulldog and Scotch terrier'; but the terrier depicted is a bullterrier and has not the slightest resemblance to any Scottish breed of that time or this. Also in the Louvre is a fine Desportes painting entitled 'Pack of Hounds of Louis XIV'; I would say that the dogs in this painting are pointers and, as usual with Desportes, represent a most accurate portrayal of such dogs at that time in France, which in itself is of value.

 Most of the time, in identifying dogs in paintings it is the lack of knowledge of art historians or gallery caption writers which leads to such mis-namings. In this way the dog in Memling's 'Vanity' is described as a Maltese, whereas it is much more likely to have been a bichon, probably a Bolognese. The portrayal of the old Dutch breed the steenbrak in Jan Gossaert's 'The Adoration of the Kings' has led to all sorts of art expert identification, ranging from small setter and large spaniel to hound-marked mongrel. A leading London fine art auctioneer in a recently-issued catalogue described a dog in an 18th century foreign painting as a whippet, a breed developed in England by north country miners at the end of the last century.

 Art is a record of history and the interpretation of art has historical importance. As Guy Paget wrote in his 'Sporting Pictures of England' (1945): "Those who wish to understand England and its growth during the seventeenth and eighteenth centuries cannot do better than study the sporting pictures of that period". But if these pictures are inaccurately described by art experts then scholars and researchers are misled. The correct identification of a breed of dog in a work of art so often provides much wider information too: location, sporting interests, period of time and even social level.

 In his 'British Sporting Paintings (1650-1850)', Oliver Millar wrote: "If we are prepared to devote study to a collection of family portraits in the hall of a country house...we cannot logically ignore the record the same patron may have commissioned of the animals in his stable and their breeding." We cannot logically ignore either, the need for accuracy in describing animals in paintings, whether in the stable or the kennel. Art historians really do need to know the difference between a bullenbeisser, a Broholmer and a braque du Bourbonnais, and if they do not, they should at least have the humility to say so.