Content
Citation(s) & Footnote(s)
Book(s)Bibliography and full coverage
Figueiredo, Luisa P.; Borém, Flávio M.; Cirillo, Marcelo Â.; Ribeiro, Fabiana C.; Giomo, Gerson S.; and Salva, Terezinha de J. (2013). The Potential for High Quality Bourbon Coffees From Different Environments. Journal of Agricultural Science 5(10), 87–98. DOI 10.5539/jas.v5n10p87. Published online September 15, 2013. Supplied PDF is 12 pages, 328,368 bytes; all narrative, methods, conclusions, 23 references and final CC BY 3.0 notice read. All five tables and four figures (11 panels) visually inspected on printed pages 88–95. No appendix or separate supplement identified.
Bourbon here names Coffea arabica genotypes, not bourbon whiskey. Preserve the article and review as a documented search false positive / bounded comparative-method reference. Exclude from direct whiskey evidence.
Design — printed pages 87–90
Fourteen coffee genotypes (11 Bourbon and three commercial comparators) at three Brazilian sites: Lavras, São Sebastião da Grama, Santo Antônio do Amparo. Randomized blocks, three field replicates and ten plants per plot; 42 samples per site, 126 across sites. Duplicate chemical analyses are technical repeats, not extra independent field replicates.
Ripe cherries were selected, pulped by a semi-washed process, dried to 11% wet-basis moisture, held 60 days, husked and size-selected. Three accredited cuppers calibrated with two samples, then recoded those samples with the rest. Five cups per evaluation and three field repetitions; sites evaluated separately. Green-bean trigonelline, caffeine and 5-CQA measured by HPLC with standards. Three compounds cannot comprehensively measure beverage aroma or liking.
ANOVA/Scott-Knott comparisons plus separate within-site PCA of five sensory attributes and final score. Because final score includes sensory components, correlation with those components is partly built into the measurement; it is not independent predictive validation. PCA used genotype means, so plots do not show replicate uncertainty.
Findings — printed pages 90–96
Genotype-by-environment interaction was reported for trigonelline and caffeine, not 5-CQA. Mean trigonelline differed across sites (0.92%, 1.06%, 1.03% dry matter). This is an in-sample difference, not a validated classifier of unknown coffee origin.
Table 4 mean final scores 80.34, 80.54, 79.79 all carry the same significance letter. Crossing the paper's 80-point commercial cutoff is not evidence of statistically significant mean-quality differences. Table 5 also shows all genotype means share one significance letter in Lavras and Santo Antônio; only São Sebastião's top group is statistically separated.
Promising genotype groups differ by site: Lavras 7/10/12/5; São Sebastião 13/14/9/5; Santo Antônio 11/9/6. Genotype 6 moves from a low group at São Sebastião to a high group at Santo Antônio, suggesting contextual performance. The authors do not identify one genotype superior everywhere. Site/altitude/soil/climate co-vary, so altitude alone is not experimentally isolated.
Figure 1's three dendrograms separate the measured chemistry from most sensory attributes; body is comparatively distinct. Figure 2 compares standard and bean-extract chromatograms, with three identified peaks plus other sample peaks. Figures 3–4 give three site-specific score and loading plots; first two components explain 74.5%, 90.9% and 87.8% of variance. Enclosing ellipses are grouping illustrations, not demonstrated confidence regions. The changing second-component direction distinguishes body/balance from other attributes; it is not a causal chemical mechanism.
Critical audit
- Printed p.89 Table 2 gives Santo Antônio coordinates with minutes/seconds beyond conventional DMS ranges (20°91′66″S; 44°95′51″W). Do not map those as valid coordinates or silently invent replacements.
- Table 3's trigonelline mean for Santo Antônio is 1.03; averaging the 14 displayed rounded values gives about 1.041. Treat as a displayed-value discrepancy needing underlying data, not a license to change the source.
- The p.92 caffeine discussion names significant genotypes 3,4,5,8,9,10,14 while Table 3's capital letters also mark site differences for 1 and 2. Flag incomplete narrative enumeration.
- The p.92 assertion that Santo Antônio coffees are not specialty coffee generalizes an overall mean below 80 to all samples; Table 5 contains several scores at or above 80 there.
- The broad conclusion that Bourbon genotypes have the greatest potential everywhere exceeds the within-site significance evidence and the restricted sample. Nine lineages are unidentified; results should not be generalized to all Bourbon coffee.
- Harvest season/year, detailed roast and brewing settings, panel repeatability, full ANOVA outputs and external validation are not provided in the supplied article. Calibration discussion and separate site sessions are not proof that every source of sensory bias is removed.
No supplied supplement was referenced as required; missing methodological detail is a reproducibility limitation, not an unread-page gap.
Academy contribution
No direct claim about bourbon whiskey flavor, caffeine, terroir or quality follows. A bounded research-literacy comparison could illustrate genotype-by-environment designs, composite-score limitations, and statistical versus commercial thresholds. Compare conceptually with Academy grain-terroir studies while keeping species, processing and outcomes separate; coffee evidence cannot validate whiskey terroir.
Retain existing whiskey-specific studies as the evidence base. No new whiskey Zettel is warranted from this semantic match. The user requested every file be fully processed, so exclusion follows full reading and visual examination, not title-only screening. Cloud hash and browser rendering remain pending; original preserved.
