Special selections, unrepresentative samples, etc 22 June 2006

From Dave Ingram, a group of related questions about 'some things that have worried me for a number of years'. With primary responses from Chris Williams (winemaker at Meerlust and The Foundry) and Tim James chipping in:

1. Sifting out the best – and worst of a vintage

At a CWG auction a few years ago, the winemaker spoke about taking their best barrels and making them auction wines: eg from 100 barrels they select one to 5 and call them special / reserve / etc, and sell them at inflated prices. The bottom few barrels of the 100, what happens to them? Surely the ‘left over’ 95 barrels should be blended together to balance the whole ‘vintage.  

CW: Even the same wine put into the same barrels develops differently. Some winemakers like to put together a ‘tête de cuvée’, their idea of the best blend of selected barrels, often sold at auction. Problem is, the remaining blend is poorer for having the ‘crème de la crème’ removed. These producers then sell this remaining blend as their regular vintage.

TJ: The ‘bottom few barrels’ are also, in fact, often removed from a quality-conscious property’s standard best wine, and put into a ‘second label’ wine. The introduction of these ‘second labels’ at many wineries must surely be one of the best ways to improve the quality of a winery’s flagship. Or else the lesser quality wine, not thought good enough for the flagship, might be sold off for anonymous use elsewhere

2. Uniformity with large production

How can a winery produce 100 000 cases of a wine and ensure the quality of all the bottles is the same ? One cannot put all the barrels into a single container as even this would result in more solids being at the bottom of the container, due to gravity. If we agitate to prevent this we cause premature ageing or oxidising.

CW: Incorrect. Some really big blends do originate from the same tank. However, really large blends such as Chateau Libertas or Kumala dual variatals are blended proportionally, fractionally over time, and there will be a difference, both in the proportions and the age of wine. The trick is to blend to a consistent taste (something which the best Champagne houses and Cognac producers have refined to an art).

TJ: Just look at the incredible size of some of the tanks at the biggest wineries ... and if the blending of all the smaller tanks takes place near the end of the process, just before bottling even, there should be no problem.

3. Unrepresentative samples

If Boland, for example, who have sometimes been accused of this, use the principle of making a strict selection (as in question 1) and submit their best barrel to competitions, put the second best in the tasting room, and then sell the rest through the retail chain at less wine quality-conscious outlets (I wonder for example, how the supermarkets can have such good pricing – I accept loss leaders as marketing but ...) when the wine does not match up to expectation we the uneducated winemaker blames the food we just ate, or the fact that we left the shopping in the car for a few hours.

CW: Indeed!

TJ: This is a real problem, and very difficult to deal with (see the comments related to competition entries following the article about Du Toitskloof's dubious practices with awards, also the point made in a comment about supermarket bullying over prices).

4. An example from Meerlust

1985 Meerlust released their first ‘Meerlust Red’. I bought a few cases, and they varied significantly in quality. Was this because all bottled wine was called ‘Red’ or was it because of bottling, where some barrels were better?

CW: Our first Red was produced in 1985, as far as I know it was released in about 1989. It was our ‘de-classified’ Rubicon 1985 – there was no Rubicon 1985 released. After barrel aging, Meerlust did not believe that it was up to the standards set for Rubicon, although it did have the qualities of a Meerlust wine, it perhaps lacked the complexity, intensity and depth of Rubicon, hence the de-classification, mainly to uphold the integrity of Rubicon but also to give our fans something to drink while they waited for their other vintages to mature! The 1985 Red was bottled at the Bergkelder, in one run, from the same blend. So why any differences? Corks, perhaps. Because cork is a natural product and the quality varies between  each individual cork, even within the same batch, cork is variable as a bottle closure. Apart from the obvious problems of around 5% cork taint, corks also vary in their ability to properly seal a bottle. In less than perfectly sealed bottles, oxygen permeates through and alongside the cork, oxidising the wine to varying degrees. The differences becomes accentuated over time. In the wine industry we have a saying, ‘there are no great wines, only great bottles’, and this is true of all bottles sealed with cork. That is one of the main reasons in favour of screwcaps. Bottom line, wines sealed with cork have huge variability, especially over time.