See my top wines at the bottom of this article.

A wine tasting horde descended on Bordeaux last week. Around 3,000 wine merchants and wine critics, almost 30 per cent more than last year and more than ever before, commuted frantically from tasting table to tasting table to see what sort of wines Europe's hottest summer ever produced in the world's most important fine wine region.

American wine importers, apparently undeterred by the unaccommodating dollar-euro exchange rate, were particularly prevalent. Perhaps because they had bought the 2001s sparingly and the 2002s hardly at all, not just because the growing season was difficult but also because anti-French sentiment was at its height a year ago, they felt confident that their customers would have money to spend on this vintage, of which much has been expected ever since last August's heatwave headlines.

The wines are qualitatively more varied than in any previous vintage I have known but by common consent the most successful spot on Bordeaux's wine map by far was what we professionals tend to call rather sloppily the northern Médoc, wine estates around the town of Pauillac and, especially, the village of St Estèphe. Châteaux Montrose, Lafite and Cos d'Estournel, clustered on a usefully damp patch of clay a good hour's drive north of the city of Bordeaux, all excelled themselves this year.

Each of these famous Châteaux produced wines that manage to be both classically long-lived Médoc yet with an extra energy and brilliance that made them taste more like health drinks than anything I have ever found in a health food shop. Lafite is likely to cost, even those lucky wine lovers who manage to buy it at opening prices, well over 100 or $200 a bottle.

Yet just a few kilometres further north, in the real northern Médoc, called just Médoc on wine labels, wine producers are in a financial situation described as "a catastrophe" by Jean-Guillaume Prats of Château Cos d'Estournel.

In the last five or six years prices of basic red Médoc, for long seen as one of Bordeaux's smarter appellations, have fallen by more 50 per cent, from what the local growers still call 20,000 francs a tonneau to 8-10,000 francs, a price for bulk wine that, once blended into a merchant's brand or supermarket's own label, would work out at the equivalent of well under a pound or two dollars per bottle.

This sharp fall in price is a direct result of twin forces in France and abroad. On export markets virtually all French wine has been suffering because of competition from newer wine producing countries, especially those of the New World. A ho- hum wine labelled just Bordeaux or Médoc is now a much harder sell than one carrying the name of an actively marketed Shiraz or Chardonnay.

At the same time wine consumption within France, where two- thirds of all bordeaux has habitually been drunk, was already plummeting before a much tougher campaign against drinking and driving, including highly visible roadside checks, was instigated. And then there is the controversial Loi Evin, the recent, health-inspired law which limits advertising of alcohol within France to such an extent that the Bordeaux wine growers [the Conseil Interprofessionnel du Vin de Bordeaux] has just been fined because its new, surely rather sensible, slogan 'Drink less but better' (Buvons moins, buvons meilleur) was judged illegal because it "incited people to purchase wine". Small wonder that the Bordeaux merchants are buying less and less from the region's producers.

Life is tough for all of France's less famous vignerons but those of Médoc are feeling a particular pinch. Partly because there was much more land available on this, the left rather than the right bank (of the Gironde estuary), the total area of vineyard doubled to 15,000 hectares (37,500 acres) in the last quarter of the 20th century, fuelled by what seemed like unquenchable international demand for fine red bordeaux. While their counterparts in the large Entre-Deux-Mers region between the left and right banks have benefited from lower production costs and some particularly dynamic co-ops, Médoc wine producers have been caught by complacency and heavy borrowings based on an outmoded financial model.

In the late 1990s when Asia was alight with enthusiasm for red bordeaux, many Médoc producers invested in new equipment, at a time when yields were generally higher than they can afford to be today. As the market has shrunk, quality has had to improve, which means that not only have prices slumped but average yields have fallen from around 62 hectolitres per hectare to closer to 55, and considerably less than in this in the drought-shrunk year of 2003.

According to the outgoing president of the organisation representing the Médoc's delightfully named crus bourgeois, Dominique Hessel of Château Moulin a Vent in Moulis and Château Tour Blanche in St Christoly- Médoc, there are not yet begging bowls in the streets of Pauillac but dramatic economies are being made. Since new barrels represent once of the most sizeable outlays for any wine producer, he says the most obvious losers are the coopers. Smaller, less glamorous Châteaux can afford to buy only 20 to 30 per cent new barrels whereas in the glory days of the late 1990s they might have splashed out on 80 per cent.

This contrast in economic fortunes between the world-famous apex of the Bordeaux pyramid, the first growths like Lafite and a few dozen of the best-known Châteaux, and the broad base of also-rans which make up the bulk of wine produced in Bordeaux is a fair reflection of the character of the 2003 vintage.

There are perhaps 40 really successful wines made in Bordeaux after the sweltering summer of 2003, for which demand will be extremely high. But the great majority of 2003s are either of average or much lower than average quality and there will be no queue of eager buyers for them. Even some of the classed growths, the Médoc's supposed nobility located just below first growths like Lafite and above the crus bourgeois, found it difficult to sell their 2001s and almost impossible to sell the unpopular 2002s, despite cutting prices. Proprietors at all levels have become considerable and unwilling stockholders.

So how will the 2003s be priced? Producers of the best wines can probably afford to raise prices a little while such an increase would be disastrous for the rest. But which of the proud, status-conscious Château owners is going to so publicly admit that theirs is not one of the finest 2003s?

They may like to use the argument that the 2003 crop was small, even smaller than 2002's reduced level, but the majority must know in their hearts and palates, if not their pockets, that they have not produced great wine this year. This is particularly true of the lighter soils of Pomerol where yellow-leaved vines were visibly expiring even quite early in August. Jacques and Fiona Thienpont made no friends among their neighbours there when they announced there would be no Le Pin this year. And I have never seen Jacques' cousin Alexandre less upbeat than when showing me his 2003 at Vieux Château Certan just round the corner. On the other hand, Châteaux Lafleur and Eglise Clinet, hardly a spit away, somehow managed to make quite thrilling Pomerol, with Châteaux Petrus and Trotanoy thoroughly respectable too.

The plateau of St-Emilion, which is effectively an extension of Pomerol, suffered equally and perhaps more disastrously in view of its much greater extent. But again, this is by no means an appellation to write off. Château Ausone on the steep limestone slopes below the town on St-Emilion itself has made probably its best wine ever - certainly the best wine of the current regime led by Alain Vauthier. And pioneer 'garage wine' Château de Valandraud, a wine I have sometimes found too rich and uncomfortably concentrated in the past, seemed to me refreshing and successful in 2003.

In sultry 2003 the refreshment factor is vital. What both of these great St-Emilions have in common are an unusually high, dominant proportion of lively, aromatic Cabernet Franc grapes rather than the much more heat-affected Merlot and yields that are almost as high as 30 hl/ha. Excessively low yields or, heaven forfend, deliberate concentration of the must after picking, which modish winemaking trick some properties reportedly used even in 2003, only succeeded in making already dangerously unbalanced, low-acid musts even more unbalanced.

If one generalisation can be made in one of Bordeaux's most heterogeneous vintages it is that 2003 was a Cabernet and not a Merlot year. The earlier-ripening Merlot vines, especially young ones without deep roots, really struggled to find sufficient water and it was Merlot grapes, which had already turned colour before the worst of the heatwave in early August, that were most likely to be burnt and shrivelled by the merciless rays of the sun - especially in properties which followed another fashionable practice designed to maximise ripeness and, usually therefore, quality: thinning the vines' leaves to allow more sunshine on to the berries.

Cabernet Sauvignon and Cabernet Franc vines were at a much less fragile stage of their development when France sweltered most seriously. In fact several experienced winemakers still seem almost mystified by how well their older Cabernet vines withstood temperatures over 40 degrees Centigrade - virtually unknown in this part of the world. Further proof perhaps that Bordeaux really is the spiritual home of this well-travelled grape variety.

They know that 2003 was not strictly not a drought year and such problems as there were were caused by excess heat and evaporation rather than by a shortage of underground moisture. The water table had been usefully topped up by quite a wet winter and rainfall in June, July and August was about average at Bordeaux's central weather station. Useful, if highly localised, showers towards the end of August restarted the ripening process to develop flavour compounds and ripen the tannins. Older Cabernet vines with deep roots, especially those in damper, cooler soils such as those around St Estèphe and Pauillac, continued against all the odds to look remarkably green-leaved and healthy throughout the season.

The most common fault in the least successful wines is sheer lack of fruity middle and, sometimes, unappetisingly low acidity. As acids fell, to record low levels, the sugars reached unknown highs so that the precocious Merlots just had to be picked - but there had not been time for the development of either interesting flavours or gentle, ripe tannins. A green streak on the finish spoils many a St Emilion and Pomerol whereas, ironically enough, the troubled area in the north of the Médoc, thanks to its cool soils, has produced some of its best wine for years.

The Bordeaux market is notoriously cyclical and the last time there was a similar crisis was as recently as the early 1990s. But that "difficult" period, as the locals tend to call it, was caused by global economic malaise rather than by today's fundamental, and apparently long-term, shifts in buying patterns. Today's Bordeaux wine map, and Château ownership, may well change considerably over the next few years - but this time Médoc estates are hardly likely to be snapped up by insurance companies and banks. A more likely lifeline may even be official financial help with pulling up less promising vines - unthinkable as recently as the late 1990s but now under discussion.

Subscribers to purple pages can find tasting notes and ratings on hundreds of 2003s.

My top 2003 wines
LEFT BANK
Ch Montrose
Ch Lafite
Ch Latour
Ch Cos d'Estournel
Ch Léoville Las Cases
Ch Léoville Poyferré
Ch Margaux
Ch Pichon Longueville (Comtesse)
Ch Pichon Longueville (Baron)


RIGHT BANK
Ch Ausone
Ch Lafleur
Ch Pétrus
Ch L'Eglise Clinet
Ch de Valandraud