Playing Russian Roulette with the First Growths

At a wine dinner with Bordeaux first growths from 1985 to 1996, the big surprise was not the quality of the wines, but the huge variation between different bottles of the same wine. Although in each case the wines had been acquired from the same source and stored together, there was not a single instance in which two bottles of the same wine tasted the same.

The 1985 Haut Brion was the greatest puzzle. The first bottle showed a funky, quasi-medicinal nose, which seemed to suggest the possibility of Brett (unlikely though that might seem for this château), although the palate cleared a bit in the glass. It was actually subtle enough that I quite liked it. The second bottle went completely in the opposite direction, showing elegant fruits, but a squeaky-clean character with  that came close to eviscerating the character of Haut Brion.

Next came Angelus 2003. (Yes, I know this was not a first growth at the time, but the organizers evidently took a broad view of the term. Anyway, you wouldn’t balk at including Mouton Rothschild pre-1973 in a first growth tasting.) First bottle was fairly restrained, with rather flat aromatics, and the character of Cabernet Franc pushed a bit into the background. It never came to life. A second bottle showed more aromatic lift with a greater sense of structure at the end. A third bottle showed a more exotic impression, more sense of the precision of Cabernet Franc, with heightened sense of elegance; the very antithesis of any thought that the heat of 2003 might have given a jammy wine, it was one of the more elegant wines of the evening, while the first bottle was one of the most disappointing.On to Mission Haut Brion 1990, where the first bottle was absolutely true to the typicity of the chateau and appellation, with elegant fruits and faint sense of cigar box in the background. The next bottle showed flattened aromatics to the point at which all the life seemed to go out of the wine. While the first bottle was fabulous, the second was merely ordinary.

We went into high gear with Ausone 1996, where the aromatics of the first bottle seemed to point more to the elegance of the left bank than the richness of the right bank. Beautifully integrated, with a sense of seamless layers of flavor, the wine showed something of the ethereal quality of a great vintage of Lafite. A second bottle had a slightly sweaty nose, a faint sense of gunflint, and gave an overall impression of reduction. A third bottle was between the first two, with a flattened profile but not obviously reduced, and a fourth was almost as good as the first.

The first bottle of Lafite 1986 was a bit flat aromatically; although showing the precision and elegance of Lafite, a sense of austerity on the finish made it seem almost stern. I took the sense of a somewhat hard edge to the wine to be the character of the vintage and was uncertain whether it would dissipate with further aging. But a second example showed that this was the character of the bottle rather than the vintage: it really sung, with that ethereal quality of Lafite showing as a seamless impression of precise, elegant fruits, all lightness of being.

With Mouton Rothschild 1989 there was another sort of surprise. The first pour (from a decanter) showed the plush power of Pauillac, very much Cabernet-driven, with black, plumy fruits. A second pour (from another decanter) showed just a little more aromatic lift. The difference between these two was much slighter than between any of the preceding pairs. Here’s the rub: the Mouton came from a single Imperial. The fact that there was any difference at all is surprising, although I have had this experience before, when some pours from an Imperial seemed to be corked while others were pure (I Want My Glass From the Bottom of the Imperial). Interestingly this was also from a Mouton 1989.

The notion there can be differences within a single (large) bottle is disturbing. I think this warrants a proper investigation. I will undertake a thorough experiment if given a supply of Imperials of first growth claret (Mouton from 1989 would be preferred). We will extract the cork and take samples from the top and bottom using a very long pipette, without stirring up the wine at all. Then we will know if proximity to the cork and oxygen on the one hand, or to the sediment on the other, makes any difference within the bottle.

It is not so surprising there should be differences between bottles. After all, if you buy a case of wine and store it for ten or twenty years, you can see at a glance that every bottle has a different level. Differences in ullage imply differences in exposure to oxygen that might well affect the flavor spectrum. But the comparisons in this tasting went well beyond minor differences, to the point at which in each flight there was one bottle that was unquestionably first growth, and one bottle that was disappointing enough to cast doubt on that status.

One moral is that if you are at a tasting where there are second pours from a different bottle, always get a fresh glass for the second pour. Another is to ask whether there is really any point at all in tasting notes, projections of aging, or recommendations, if every single bottle is going to be different. Certainly this is not what the punter expects when he buys a bottle. The culprit must be the cork (inter alia, the sommelier reported that he had never rejected so many corked bottles in preparing for a tasting, so the worst cases had already been removed).

Is there any alternative? Experience with New World wines suggests that using screwcaps might cause the wines to age more slowly and a little differently, but with greater consistency. I’m sure the argument in Bordeaux would be that it’s a bad idea to risk damaging the product of one of the most successful wine regions in the world, but is it so successful if there is no predictability after twenty years?

Advertisements

A Visit to Amorim: Progress with Cork

“It’s very upsetting when people come and try to kill you. We’ve had two funerals already. But when you have 70% of wines under cork, I think you can see that after fifteen years the debate has been resolved,” Carlos de Jesus started our meeting at Amorim.

I had taken advantage of my trip to Douro and the Dão to visit Amorim’s plant just to the south of Porto. Half of the world’s corks come from Portugal, Amorim has more than a third of the market, and roughly two thirds of business is in wine stoppers. There’s a plant farther south, near the forests, when the first processing is done, and then the sheets of cork are sent to the plant in the north to be turned into corks.

Amorim1Sheets of corks come from the south

We visited the plant where both agglomerated corks (made of small particles) and solid corks (punched out from the sheet) are made. It’s actually technically easier to remove TCA from agglomerated corks, because the broken up particles can be heated to extract it before they’re stuck together into the cork. There’s a whole hall with a vast series of machines taking in raw cork and spewing out finished corks.

Amorim3Raw cork goes in one end and agglomerated corks, free of TCA, come out the other end

To make solid cork, the sheets are cut into bars of the right width, and then the corks are punched out. There are two separate parts of the plant devoted to this, one automated, one manual. Watching the operators of the manual line seems like a time warp. They hold the bar in a machine, operate a treadle to punch out the cork, and then move the bar along to punch out the next cork. This seemed old fashioned compared with the automated equipment, but Carlos explained that the operator gets better results, just like manual grape harvest is better than mechanized!

Amorim5Corks being punched out of bars on a automated line

After visiting the plant, we walked over to Amorim’s research facility, in a small building close by. The R&D department was created in 2000. A main objective is to find solutions to the TCA problem, but they’ve also been investigating the route for oxygen to enter bottles.

I was quite surprised when researcher Paulo Lopez explained that the routes for ingress are different for different closures. For synthetics and screwcaps, it’s permability from the exterior, but for cork it is the oxygen contained within all those little cellules in the cork itself. (Presumably this applies unless and until the cork loosens, but experiments show that this doesn’t happen in the first few years.)

Paulo says that the TCA problem is being significantly reduced. In 2003 the average TCA was 2.68 ng/l, now it is 0.62 ng/l. “We consider 0.5 ng/l to be zero as that is the limit of detection by available equipment. 1 ng/l is the sensitivity of the most sensitive human tasters.”

The lab does Q & A, testing 800 samples per day for TCA. This requires 24 hours to soak the cork and then 24 hours to get results. So it’s a check, but they have to wait 48 hours to release each batch.

Amorim7Serried ranks of gas chromatographs in the lab

The latest system is NDtech, which analyzes individual corks. (If not secret, it’s certainly more closely guarded as no photos were allowed here.) Eight years in development, NDtech basically takes the gas chromatography system to automated speeds. A single production line can handle about 15,000 corks per day, and it costs 15¢ per cork.

Each line spews corks out into one of three bins: guaranteed below detection; at the edge of detection; and not acceptable. It didn’t look casually as though the second bin was that much less full than the first (the third was much smaller), but Paolo says the proportions are 65%, 30%, and 5%. “There are still too many false positives,” he says, “the next stage is to reduce them.”

It was a major feat to automate the equipment to this point, but capacity is still limited. There will be about 15 million NDtech corks this year and 60 million next year, which is probably way below demand. “Looking at individual corks is the holy grail. NDtech analyzes individual corks, but I doubt that we’re going to apply it to billions of corks. We’ll start with 50 million corks and then we’ll scale up. We’re not here to defeat TCA and go out of business. We need to defeat TCA and stay in business,” says Carlos.

Experiments at Chateau Margaux: fining, filtration, and closures

You might think that once a wine has gone through alcoholic and malolactic fermentations, and been matured for months in oak barriques, that the die was set. Not a bit of it: the continuing experiments at Chateau Margaux show that there is an effect from every stage of what I suppose you might call finishing procedures. “Once again it is impossible to deny the differences, which is somewhat frightening,” was Paul Pontallier’s comment at the end of the tasting.

Once simply part of the routine of producing red wine, fining has become somewhat of a controversial issue, and is now one of the most obvious differences between Old and New World. When I was discussing Cabernet Sauvignon with producers in the course of researching my book Claret and Cabs,  virtually every producer in Bordeaux told me they fine, and almost all producers in Napa Valley said they avoid fining just like filtration.

Fining was introduced as a procedure to lighten the wine, the traditional process being to add egg whites to the barrel. Albumin protein in the egg whites is positively charged, and so reacts with negatively charged tannins to precipitate them. The conventional argument is that this softens the wine by removing harsh tannins and also polishes it by taking out other components. Critics ask why the egg whites should act only on harsh tannins and take the view that desirable tannins and other components might equally well be affected. And, of course, over the past ten or twenty years, tannins have become much riper so you might well ask whether they still need to be removed. (Most producers who continue to find do use fewer egg whites now than they used to.)

Well, the answer at least in the context of Bordeaux, is absolutely clear: the 2004 Chateau Margaux fined with 6 eggs tasted like a completely different wine from the unfined example. Some people thought they could see a difference on the noses of the two wines, but personally I thought they were indistinguishable (and I am a bit hard put to see why volatile compounds might be removed by fining). But the unfined example had more evident tannic grip, less finesse, and came up just a little shorter on the finish. The fined sample simply gave a distinctly more polished impression, not just because of less tannin, but with a sense of being altogether better rounded. Chateau Margaux as you will find bottles in the shops, by the way, has been fined with 5 egg whites per barrel since 1996.

Filtration seemed to have less effect, as tested by comparing unfiltered Chateau Margaux 1995 with sterile-filtered wine. There was no detectable difference on the nose, and the balance on the palate seemed very similar. The main effect to my mind was that the sterile-filtered example seemed like a slightly older, more developed wine, with a touch of sous bois that was not evident on the unfiltered wine. Most participants preferred the second wine, but that depends somewhat on whether you prefer your wines younger or older. Paul Pontallier felt that the filtered wine had actually deteriorated a little due to a touch of oxidation. I can’t say that I would describe the filtered wine as eviscerated or having lost character as a result of filtration, but I suppose it might be the case that the filtration removed components that protect against oxidation.

The closure trial compared Pavillon Rouge 2002 sealed with natural corks with the same wine sealed under screwcaps. There had also been a trial with synthetic corks, but apparently the results were disastrous, and in relatively short order the wine was spoiled. “It’s a good decision to use screwcaps for white wines that will be drunk in the first six months,” says Paul Pontallier, “and with what I know now I would do the same, but our dilemma is that we want to make wine that will age.” The two wines were quite different: open, round, and fruity under cork, but reserved, backward, and showing more austerity under screwcap. Interestingly, the participants split more or less equally as to which style they preferred.

Chateau Margaux is just about to undertake the construction of a new experimental cellar that will allow them to undertake even more experiments. Among future projects are looking into the properties of individual clones of grape varieties and investigating the effects of different types of pressing. “To my astonishment, many people take the view that, if it is new, it must be better,” says Paul Pontallier, “I admire their optimism, but I feel the need to experiment first.”

An Experiment with Corks and Screwcaps

It’s years since I did a scientific experiment but there is one I would like to see done with corks and screwcaps. It’s quite amazing that even with a more than a decade’s experience of bottling wines under screwcaps, the long term effects of the type of closure remain controversial. One issue that I think should be finally resolved is just what effect exposure to oxygen has on long term maturation.

When the same wine is bottled both under cork and under screwcap, it’s evident within a few months that they develop differently. Most of the comparative tastings that I have done have been with white wines, where the usual difference is that wine under screwcap retains brighter fruits with more evident freshness. Preferences are usually split at such tastings between the bottlings: some people prefer the fresh, young style of screwcaps, while others find more complexity in the greater development of the wine under cork.

When I was out in New Zealand and Australia earlier this year, I had several opportunities to compare older red wines that had been bottled under both types of closure. The results were completely consistent.  The wine under screwcap always seemed younger – in a blind tasting you might have said by a couple of years – with more primary fruits, whereas the wine under cork showed some development towards more savory, sometimes even tertiary, aromas and flavors. All of these wines were Pinot Noir (mostly from the first few years of this millennium) but I assume the results would be generally true for all red wines. (The tastings are described in more detail in my recent book, In Search of Pinot Noir). This may not be a completely fair comparison, because the reason for the switch to screwcaps was often the terrible condition of the corks available down under. In fact, when there was the opportunity to taste multiple bottles under cork, they often tasted as different from one another as they did from screwcap, an immediate validation of the decision to switch to screwcaps.

So wines under screwcap clearly develop more slowly: the question in my mind is whether they develop in the same way at a slower pace or whether the overall pattern of  development is different, reflecting a different relative timing of the loss of primary fruits and the appearance of tertiary flavors. Among the wines I tasted, when the wines under cork were in perfect condition, I generally preferred them: but that may be because my taste generally runs to older wines. I would be really interested to repeat the comparison in a few years when the screwcap wines have developed further to see which I prefer then.

Anyway, back to the experiment. The difference between screwcaps and corks is the rate with which oxygen gets into the bottle. It can be close to zero for screwcaps: in fact, there have been problems involving reduction for wines bottled under screwcap, just as damaging in their way as problems with oxidation for wines with faulty corks. Sulfur levels need to be reduced when bottling under screwcap and it may take a while to establish the most appropriate levels for wines intended for any aging. The very best corks (defined operationally as the tightest) have very low oxygenation levels close to those of screwcaps. But the unanswered question is whether you do actually need some level of oxygen exposure; for corks this comes from the supposed “breathing” of the cork; for screwcaps it could come in the future from liners with defined rates for passage of oxygen.

So the experiment I want to do is to determine definitively whether the difference between corks and screwcaps is solely due to oxygen exposure. It is very simple in principle. Bottle a (red) wine under both cork and screwcap. Take bottles with each type of closure and keep one set in normal cellar conditions (cool, dark, humid). Put the other set under identical conditions but in an atmosphere of nitrogen. Then see whether the wines with cork and screwcap closures develop in exactly the same way under nitrogen (which is what you would expect if oxygen is the sole relevant factor). And of course see what differences emerge with and between the wines in the normal cellar. All that’s required is a cellar filled with nitrogen (and I suppose a means of retrieving bottles for periodic testing). Then finally we would know the answer instead of speculating and arguing about it.