The Keeving Experiment

Producing a naturally sweet cider by controlling the fermentation.

One of the problems I found with making cider from a variety of culinary apples was that the juice tended to ferment out to almost complete dryness.

Many drinkers appreciated the resulting distinctive taste of Royal County, but the rather excessive dryness did mean that it wasn't as widely popular as it might have been - and this tended to affect sales.

I therefore looked into various ways of achieving a slightly sweeter result. A farm cider maker that I visited in Cornwall added a quantity of a special powder obtained from a well-known firm of chemists, which apparently checked the fermentation. Not only did this product contain substances that I felt were better absent from my cider, but it was only available in a minimum quantity costing £300!

Some improvement was made by adding small quantities of fructose (fruit sugar) to each 5 gallon polybarrel just before despatch, but this meant there was a possibility that the cider would recommence fermenting if kept for some time in warm conditions. The other options, of using non-fermentable saccharin, I rejected on the grounds that its taste is apparent in the drink. Then in the autumn of 1997 I went to interview an amateur cider maker, Andrew Lea, for a magazine article.

I use the word 'amateur' because Andrew does not earn his living making cider, but he is a scientist who worked for the former Cider Research Station at Long Ashton, and he makes superb cider in his spare time from the 33 cider apple trees he has planted at his home near Wallingford in Oxfordshire. During my visit, I sampled some of his naturally-conditioned bottled cider, and to say that I was impressed is an understatement! It was rich and full-flavoured, with a sweetness that was obviously not the result of adding saccharin or any other artificial sweetener. Andrew explained that it was made by adapting a technique known as 'keeving', which is very popular among French cider-makers, especially in Normandy.

'Keeving' probably comes from the French word 'cuvage', and is a method of making cider that produces a natural sweetness. The minced pulp is held for 24 hours before pressing, allowing certain enzymatic changes to take place, and this should cause the juice to separate out into layers in the fermentation vessel - a floating layer of pectin gel on top; clear juice in the middle; and crud on the bottom. The traditional way of helping this to happen is to add a mixture of chalk (CaCO3) and salt (NaCl), but the French commercial producers now add calcium chloride (CaCl2). This reacts with the pectin in the juice to form calcium pectate, which is insoluble and floats to the top.

As the pectin traps much of the nitrogenous vitamins and amino acids that was present in the juice, the clear juice which is siphoned off is low in nutrients favourable to the yeast, and this inhibits its working. Slow fermentation on the natural yeast at low temperatures during the winter, allied with racking at a gravity of about 1025 degrees to arrest fermentation, produces a cider with a high proportion of residual sweetness but, of course, lower alcohol content.

I was determined to try this technique with my 1998 pressing, and Andrew was interested in seeing the results of it being used on a commercial basis and larger scale, and with a different type of fruit. He arranged to get me some calcium chloride, and also kindly let me have a small quantity of an enzyme which he had sent over from a manufacturer in Caen, in the Calvados region of Normandy. This was designed to assist the calcium chloride in producing the cap of pectate gel.

One aspect of the technique which I was not able to copy was leaving the pulp for 24 hours before pressing, as this process was done elsewhere, but the remainder was straightforward. The calcium chloride had to be completely dissolved in the juice, and then the enzyme added. The maximum amount of chloride allowed was 800 parts per million, but Andrew recommended only 400, which was 40 grams per 100 litres. This equated to about 90 grams or 3½ ounces per 50 gallons (235 litres), which is the size of my fermentation vessels. The enzyme, which was a brown liquid, had to be added at the rate of 7-10 millilitres per 100 litres of juice, so that each of my containers received 20 millilitres. This was done the day after pressing, when the juice was transferred to its fermentation vessels. No sulphur dioxide was added to any of the juice, other than that from the very minute amounts of sodium metabisulphite steriliser remaining in the vessels.

In all, I did four pressings in 1998; on 15 October (1), 29 October (2), 9 November (3) and 26 November (4) - each of about 120 gallons (550 litres). Pressing no.(1) received no treatment for three reasons:

  1. It was necessary to have a control with which to compare the keeved juice.
  2. If the experiment was too successful and produced cider that was too sweet, it would be useful to have some really dry stuff with which to blend it.
  3. I still hadn't got any calcium chloride!
The 'chapeau brun'

The three subsequent pressings did get the treatment, however, and I was amazed at the result. My apple juice always produces a layer of brown stuff on the top, as the fermenting yeast brings up small pieces of solids which clump together, but it was nothing like the cap produced in each vessel by the calcium chloride treatment. A layer of brown dross (which the French call the 'chapeau brun') nearly an inch (2.5 cms) thick formed on the top (top left), and this later produced a white ring of dry fluffy material like meringue (bottom left). I wondered if this was some of the calcium chloride coming out of solution, but Andrew informs me that it is still just calcium pectate.

The 'chapeau brun'

I skimmed off this dross as directed, but a second, lighter cap formed on (3) and (4), and this was also skimmed. It is significant that these two batches of fruit contained a higher proportion of cider apples in the mix, whereas the first two were almost entirely culinary fruit with a few crab apples.

It is also significant that, while (2) received the full recommended treatment of calcium chloride and enzyme, there was not enough enzyme available to give the two subsequent batches the full dose. (3) got half the amount and (4) nothing at all. Strangely enough, this did not seem to affect the production of the 'chapeau brun', but it certainly affected the fermentation, as this graph demonstrates.

Fermentation profile graph

This was produced by the gravity readings taken from each vessel at intervals from the day after pressing, and I think it is self-explanatory. Batch (1) (blue) behaved as in previous years, fermenting down very quickly. (2) (yellow), which received the full recommended dose of enzyme, was the slowest of all, and Andrew believes that this is because it formed the best gel and therefore the most complete removal of nutrients. (3) (green) and (4) (red) behaved in almost identical fashion. Again, this could be the result of the high proportion of cider apples in the mix.

There are a number of points which should be noted:

  1. Although the recommended procedure is to rack the juice after the cap has been skimmed, in order to remove the sludge on the bottom of the vessel, this was not done in most cases until some time afterwards.
  2. Because of the poor summer in 1998, with a lack of sunshine, much of the juice was low in sugar. It was therefore deemed necessary to add a small amount of cane sugar to each batch, in order to achieve a reasonable alcohol level in the finished cider - especially as the fermentation would produce less alcohol anyway. This was not done at the beginning of fermentation in each case, but, for convenience, the gravities on the graph assume that it was.
  3. The gravities were not measured in each vessel at regular intervals, although the graph suggests they were. The more recent gravities were taken more often, and these are the more important ones.
  4. No correction was made for temperature when taking the gravities. The 50 gallon containers are stored in a wooden shed and possess a considerable amount of hysteresis. Furthermore, the weather in this part of the UK over the past few months has been reasonably consistent in temperature, without any long cold spells.

This experiment was not conducted in a laboratory under strictly controlled scientific conditions, but I don't believe that any of the points made above affected the result unduly. The difference in the fermentation profiles of the three keeved batches compared with that of the control is so substantial as to nullify any slight errors. I believe that it proves the effectiveness of the process in checking fermentation.

The graph was produced at the beginning of January 1999, and all the batches were still fermenting. I shall continue to take readings at intervals until fermentation ceases, so an updated graph will be posted eventually. It is likely that all four batches will eventually be blended together to produce the 1998 Royal County, but, in view of the way they have behaved, I may produce two separate ciders - one from (1) and (2) and the other from (3) and (4).

Watch this space!

Since conducting this experiment in 1998/99, I have continued to keeve the juice every year, and the result is certainly a less assertive product. The only difference is that I have discontinued using the enzyme because (a) it is expensive and (b) it doesn't seem to make any difference.

Proprietor, The Lambourn Valley Cider Company.

E-mail: Click here