What cider science does not yet know

The mechanism is well established: tannins precipitate salivary proteins, and chain length governs the split between bitterness and astringency. Model-solution work and fractionation studies support this robustly, and it is one of the better-understood pieces of cider chemistry.

What does not follow is prediction. Measure total phenolics, or even the degree of polymerisation, in a finished cider, and the correlation with what a trained panel reports is moderate at best. The reasons are known in outline — ethanol interferes with tannin–protein binding, residual sugar suppresses the perception, polysaccharides compete for binding sites, pH shifts the precipitation behaviour, and salivary flow rate differs several-fold between individuals — but no model combining them predicts panel scores reliably across ciders.

The practical consequence is that a phenolic figure is a statement about composition and not a forecast of the palate. It also means that classification thresholds built on total tannin, including the Long Ashton ones, carry a known limitation their own authors acknowledged. This site records phenolic measurements and sensory bands as separate things because collapsing them would assert a relationship that is not there.

That they are present is not in doubt. Spontaneous cider fermentation begins with a substantial population of apiculate and other non-Saccharomyces yeasts — Hanseniaspora, Metschnikowia, Candida, Pichia, Torulaspora — and their succession has been characterised repeatedly by culture and by sequencing.

What they contribute to the finished cider is much less clear. They are active mainly in the first days, at low alcohol, and are then displaced; some of what they produce is subsequently metabolised or stripped. Studies pairing spontaneous and inoculated ferments generally find differences in the volatile profile, but attributing those differences to particular organisms is difficult, because the two treatments differ in more than the organism list — in kinetics, in temperature history, in nitrogen consumption and in oxygen exposure.

Controlled work using defined co-inoculations is beginning to isolate specific contributions, and some organisms have reasonably well characterised effects. But a general account of what the wild population does to a finished cider, and how much of it survives maturation and blending, does not yet exist. Claims that a particular wild yeast is responsible for a particular characteristic in a finished commercial cider should be read as hypotheses.

The claim appears constantly: that the microbial population of a particular orchard, cellar or region imparts a character that could not be reproduced elsewhere. It is an attractive claim, and it is not established.

What is established is that microbial populations differ between sites, and that a cellar develops a resident population over years of use. That much is well documented, in cider as in wine. What has not been shown is the next step: that those population differences produce differences in the finished, matured, blended product that a panel can attribute to region above chance, across producers and across seasons.

The methodological difficulty is severe. Producers in a region share fruit, equipment, cellar temperatures, water, timing and technique, so any regional difference has many candidate causes besides the yeast. Separating them requires either fermenting identical juice with populations from different regions under identical conditions, or a large multi-producer study with the confounders controlled, and neither has been done at a scale that would settle it for cider.

This site therefore records that regional populations exist and differ, and does not assert that they produce a regional signature in the glass. Where a producer or a tradition makes that claim, it is reported as a claim.

The analytical foundation of English-language cider science is the Long Ashton work, carried out on fruit grown at one Somerset site across the twentieth century. It is genuinely valuable and it is the reason so much can be said at all. It is also historic, site-specific, and reaching the present largely through secondary compilations rather than through the original tables, which are not consolidated in any open archive since the station closed.

Modern analysis exists for a minority of cultivars. North American programmes have published juice chemistry for a set of European varieties grown in their climates, which is the single most useful counterweight to treating English figures as universal; French and Spanish institutes publish for their own varieties. Between them these cover perhaps a few dozen cultivars well, against several thousand named cider apples and several hundred named perry pears.

For most named cultivars, what exists is a description — how the fruit looks, when it ripens, what a maker thinks of it — and a classification, and nothing measured within living memory. Many perry pears are known from a handful of surviving trees and have never been analysed at all. This is why so many records on this site carry a classification and a described blend profile rather than measurements, and why the profiles are marked with the basis they rest on.

Cider writing has an unusually high ratio of confident assertion to published evidence, for understandable reasons: the field is small, much of the practical knowledge is craft knowledge held by producers, and the research base is scattered across a handful of institutes in several languages.

The risk that creates is that a plausible mechanism, repeated often enough, becomes indistinguishable from a finding. Several of the claims on this page are repeated routinely in cider writing as established, and they are not.

Stating that plainly is not scepticism for its own sake. It is what allows the rest of the site to be trusted: a reader who can see where the boundary of the evidence is drawn can rely on what sits inside it.