Dataset

The USDA-PGRU germplasm characterisation

The largest single-season cider characterisation on the site: forty-seven accessions from the United States national apple collection, twelve properties each, all picked in one autumn.

What does a national germplasm collection know about cider apples?

For one harvest, a great deal and all of it at once. A Cornell project sampled accessions from the USDA-ARS collection at Geneva, New York and measured the same twelve things for every one: fruit weight and diameter, starch pattern index and flesh firmness, soluble solids, titratable acidity, pH, total phenolics, and the four sugars separately — glucose, fructose, sucrose and sorbitol.

The four-sugar breakdown is what makes it unusual. Almost every other dataset reports a gravity or a Brix and stops, and sorbitol in particular matters because yeast cannot ferment it: it survives into the finished drink as sweetness that no amount of fermentation removes.

The programme

Transcribed from the source as published. Nothing in this panel is CiderHQ’s wording.

Site: USDA-ARS Plant Genetic Resources Unit, Geneva, New York, in the Finger Lakes region of western New York State
Period: 2017
Cultivars transcribed: 47
Measurements: 560
Values withheld: 14 — see limitations
Classifies under: long-ashton

Climate. Humid continental, with cold winters and warm summers moderated by the Finger Lakes — a sharper seasonal swing than any of the European cider regions whose fruit the collection holds, and warm enough in late summer to bring acid down faster than a maritime autumn does.

Method. One harvest season. Fruit assessed for weight, diameter, starch pattern index on the Cornell 1–8 chart and flesh firmness by penetrometer; juice pressed and measured for soluble solids by refractometer, titratable acidity by titration to pH 8.1 as malic acid, pH as the autotitrator’s starting reading on juice that had been frozen, total phenolics by Folin–Ciocalteu assay as gallic acid equivalent, and glucose, fructose, sucrose and sorbitol by enzymatic assay.

What the source says about itself

The programme sorted its accessions under the Long Ashton scheme using thresholds stated in modern units — bitter above 2.0 g/L tannin, sharp above 4.5 g/L malic acid — and found 29% of the 158 accessions bittersweet, 13% bittersharp, 28% sweet and 30% sharp. CiderHQ records the numbers rather than the derived classes, because the class follows from the numbers and the numbers are what a reader can check. Every entry names the accession the tree was held under: a collection label is a claim about identity, not a proof of one.

One cultivar, four measurement programmes, four sets of numbers — and the fourth is not a juice analysis at all. This is why CiderHQ never averages across studies.

Described in full

Shape: Four stacked panels, one per programme, each naming the place, the period and the substrate that was analysed. Within each panel the figures that programme published for Kingston Black are listed with their units. The fourth panel is set apart by a dashed border and a heading that reads "finished cider, not juice".
WSU Mount Vernon, Skagit Valley, Washington, 2002–2017, juice: pH 3.51, total acid 0.61% w/v, total tannin 0.15% w/v, specific gravity 1.055, 13.3 °Bx. A maritime North American site chosen as the closest analogue to the English West Country, averaged across the seasons the cultivar was in trial.
USDA-PGRU Geneva, New York, 2017, juice: pH 4.23, titratable acidity 2.1 g/L, total phenolics 1,080 mg/L, 12.5 °Bx. One season, one harvest, in a continental climate with hotter summers and colder winters than the Skagit Valley. The pH is 0.72 units above the Mount Vernon figure for the same cultivar.
Long Ashton, Somerset, 1905–1975, juice: Total acid 0.58% w/v, total tannin 0.20% w/v. A seventy-year composite from the station whose figures underlie the English classification. Its tannin figure is a third higher than Mount Vernon’s for the same fruit.
Geneva, New York, late 1970s, finished cider: Finished pH 3.56, finished total acid 0.58% w/v, finished total tannin 0.062% w/v. Set apart because it is a different quantity with the same name: fermentation, fining and time remove phenolic material, so this tannin figure is roughly a third of the Long Ashton juice figure and cannot be read as disagreeing with it.
What the figure argues: The pH spread across the three juice analyses is 3.51 to 4.23, which is large enough to change what a maker would do about it. Season, site and method account for differences of that size, and none of the three is known to be wrong — though the Geneva programme’s pH readings run above every other programme’s for the same kinds of fruit, which its study page records. A measurement is a statement about a fruit and a place and a season and a method. Averaging the four would produce a number that no orchard ever grew and no laboratory ever measured.

What it adds

The sugar profile, cultivar by cultivar, which nothing else on the site carries for apples at this scale.
Starch pattern index and flesh firmness beside the juice chemistry, which lets a reader connect maturity at picking to what the juice measured.
It records the accession number and the name the collection holds each tree under. A germplasm accession is a tree with a label, and the label is a claim about identity rather than a proof of one — so a reader comparing this with a Somerset analysis of the same cultivar can see that the two are statements about two trees.
It publishes a standard deviation beside every mean, which is carried as the measurement’s range wherever the deviation is usable.

What it cannot be asked

Stated as limits rather than as apology. A dataset with known bounds is more useful than one presented as complete.

It is one season. Everything was picked in the autumn of 2017 from trees at one site, and a cider apple’s acid and phenolic load move substantially between seasons. A single-year figure stays a single-year figure however many accessions stand beside it.
Fourteen of the forty-seven accessions carry at least one published deviation that cannot be right — Foxwhelp’s juice pH at 3.49 give or take 3.40, Liberty’s phenolics at 560 mg/L give or take 2,200. In each case the mean is published and the interval withheld, with the published figures and the reason recorded.
The phenolics are a Folin–Ciocalteu assay in gallic acid equivalents, which counts everything that reduces the reagent and is not a tannin titration. The thesis converts them to Löwenthal tannic acid equivalents by a laboratory factor of 1.6 only to assign classes; the figures here are the unconverted ones. Until October 2026 this site labelled them as Löwenthal figures, which was wrong.
The four sugars and the refractometer reading do not always describe the same juice. Bramley’s Seedling is recorded at 139 g/L of glucose, fructose and sucrose against 8.9 °Brix, which allows about 92 g/L of dissolved solids of every kind, and several other varieties are similarly out of line in one direction or the other. The figures were checked against the thesis’s supplemental tables in October 2026 and are transcribed exactly, so the disagreement is in the published source: its own means are 125 g/L of the three sugars at 12.3 °Brix. The thesis does not say which reading is wrong. The site therefore uses the three sugars only as proportions of one another, in summary, and does not set them against the Brix.
The pH column runs high. Twelve of the forty-seven juices read above 4.5 and six above 4.7, against a highest of 4.7 for apple juice in every other programme on the site. The column is the starting reading of the autotitrator on juice that had been stored frozen and thawed, transcribed exactly. The programme’s acidity agrees with its pH — the low-acid fruit is the high-pH fruit — so the column may be elevated by the handling or the measurement, rather than wrong. Readings above 4.7 are shown on their cultivar pages and left out of every comparison; the rest are used, with this note.

Why these figures are not universal

One orchard, one season, one set of trees whose identities are collection labels. The Finger Lakes are continental where the English cider counties are maritime, with hotter summers and colder winters, and the fruit reflects it — Kingston Black reads pH 4.23 here against 3.51 in the Skagit Valley. Both are correct measurements of what grew where it grew.

What was measured

Every property in the transcription, with how many cultivars carry it. A property measured on some cultivars and not others is a gap in the source, not in the transcription.

Properties recorded across 47 cultivars in The USDA-PGRU germplasm characterisation
Property	Cultivars	Reported in
Fruit weight	47	Grams
Fruit diameter	47	Millimetres
Brix	47	Degrees Brix
Titratable acidity	47	Grams per litre
pH	47	pH
Total phenolics	47	Milligrams per litre
Glucose	47	Grams per litre
Fructose	47	Grams per litre
Sucrose	47	Grams per litre
Sorbitol	47	Grams per litre
Starch pattern index	45	Starch pattern index
Fruit firmness	45	Newtons

Total phenolics across every cultivar in this dataset that carries it — 47 of them, from 4290 down to 450 mg/L. Every figure here comes from this one programme, measured by one method in one place, which is the only arrangement under which a comparison like this is sound. Nothing on this chart may be read against another dataset’s column.
Cultivar	Total phenolics (mg/L)	Where it sits in this study’s range
Médaille d’Or	4290
Bramtot	3530
Coat Jersey	3190
Doux Normandie	2720
Stembridge Jersey	2520
Domaines	2050
Browns Apple	2040
Muscadet de Dieppe	1900
Bulmer’s Norman	1840
Black Twig	1810
Fillbarrel	1810
Cap of Liberty	1740
Nehou	1670
Michelin	1660
Binet Rouge	1650
Skyrme’s Kernel	1570
Ellis Bitter	1440
Foxwhelp	1440
Tale Sweet	1420
Northern Spy	1190
Court Royal	1120
Kingston Black	1080
Golden Russet	1070
Rhode Island Greening	980
Newtown Pippin	970
Harrison	960
Stembridge Cluster	930
Arkansas Black	920
Gewürzluiken	920
Grenadier	920
Le Bret	920
Ashmead’s Kernel	870
Bramley’s Seedling	850
Esopus Spitzenburg	850
Tardive Forestier	840
Reine des Pommes	810
Jonathan	780
Roxbury Russet	760
Winesap	760
Taylor’s	720
Ben Davis	680
Stayman Winesap	650
Sturmer Pippin	650
Golden Delicious	630
Glockenapfel	600
Liberty	560
Wickson Crab	450

Ordering only. Where a figure was published as a range the midpoint is used to place the row, and the range itself — not the midpoint — is what appears on the cultivar’s own page.

The whole distribution

Everything Characterization of Malus Genotypes within the USDA-PGRU Malus Germplasm Collection for their Potential Use within the Hard Cider Industry measured on apple juice, as published on this site: one method per row, repeat samples of a variety kept separate. Drawn live from the cultivar records.
Property	Lowest	Lower quartile	Median	Upper quartile	Highest	Samples (varieties)
Titratable acidity as malic (g/L)	0.9	2.1	4.1	5.3	9.2	47 (47)
pH (pH)	3.09	3.48	3.64	4.33	4.63	41 (41)
Total phenolics (mg/L)	450	810	980	1740	4290	47 (47)
Brix (°Bx)	8.1	11.3	12.5	13.6	18.7	47 (47)
Fruit diameter (mm)	47.0	59.0	65.0	74.0	94.0	47 (47)
Fruit weight (g)	53.0	95.0	124.0	174.0	294.0	47 (47)
Fruit firmness (N)	26.0	50.9	58.9	71.1	114.1	45 (45)
Starch pattern index (SPI)	5.5	7.4	8.0	8.0	8.0	45 (45)

What it can be read beside

The commonest way to misuse this corpus is not to misread one figure but to put two side by side that were never measuring the same thing. This is the judgement, stated.

Whether The USDA-PGRU germplasm characterisation may be read in one column with the site’s other datasets
Dataset	Verdict	Why
The Virginia polyphenol characterisation	Partly comparable	Both American, both reporting total phenolics by the Folin–Ciocalteu assay in gallic acid equivalents. The assay is shared; the laboratory, the season, the site and the handling of the juice are not, so the two sets are read side by side with care and not pooled.
The Geneva finished-cider analyses	Not comparable	The same institution and nearly the same ground, forty years apart — and the earlier work measured finished cider rather than juice, so nothing but fruit weight can cross between them.

Where a dataset is not listed, no judgement has been recorded and the safe reading is that the two have not been checked against each other. CiderHQ never averages across studies: how evidence is handled sets out why.

Cultivars in this dataset

Each carries this programme’s figures on its own page, alongside whatever else has measured it.

Where the site uses these figures

Sources

What this page rests on. Where a source is marked as registered rather than read, CiderHQ is recording that the body is authoritative on the subject without claiming to have worked through the document itself. See our evidence policy for what each state means.

Characterization of Malus Genotypes within the USDA-PGRU Malus Germplasm Collection for their Potential Use within the Hard Cider Industry
Cornell University / USDA-ARS Plant Genetic Resources Unit · university · passage verified 2026-08-25 · covers 2017 harvest
A Cornell masters thesis, open access, read in full on 2026-08-25 including the supplemental tables. The largest single-season cider characterisation CiderHQ holds: 158 accessions from the United States national apple collection, each measured for fruit weight and diameter, starch pattern index, firmness, soluble solids, titratable acidity, pH, total phenolics and the four sugars separately. Every mean is published with its standard deviation, which is why CiderHQ can store dispersion here and cannot for most historic tables. Forty-seven of the accessions correspond to cultivars the site already holds; the figures are transcribed for those and the accession identifier is recorded against each, because a germplasm label is a claim about identity rather than a proof of one.