Content
Citation(s) & Footnote(s)
Book(s)Whiskey Knowledge Databases › Literature Notes
Complete analytical Literature Note produced from a full-source review. Claims below are bounded by the recorded evidence and limitations.
Scope and core summary
Complete synthesis of a 204-page dissertation integrating flavor chemistry, gas chromatography–olfactometry, reconstitution and omission, descriptive and rapid sensory methods, and multivariate statistics across aroma mixtures and culinary systems.
Author argument
Johnson argues that instrumental and sensory methods become most powerful when designed together: volatile composition alone cannot explain perceived flavor because masking, synergy, emergence, matrix effects, context, and assessor response shape the experienced whole.
Researcher synthesis
This provides the methodological backbone for the project’s sensory and blending work. Treat a blend as an interacting system, distinguish detection from causation, combine structured tasting with chemical evidence when available, and choose rapid or rigorous methods according to the decision being made.
Evidence assessment
Substantial primary doctoral research plus literature synthesis. Particularly strong for method integration and the logic of reconstitution/omission.
Limitations and open questions
Limitations: Most experiments concern non-whiskey systems such as lavender, bitters, foods, and vinegars. Transfer to whiskey must be labeled as inference, and laboratory aroma methods do not capture the entire drinking experience.
Open questions: Which whiskey-specific studies can test the dissertation’s mixture-emergence framework in commercial-whiskey blends?
Connected records
- Contributors: 1
- Verified Excerpts: 3
- Citations: 1
- Zettels: 3
Completion record
Historical completion statement superseded by the September28 audit below: all204 supplied sheets have now been examined with chapter-specific analysis and visual/table checks. Missing or illegible supplementary cells remain in both supplied and author copies; no complete-dataset certification is made.
September 28 Full Reading and Evidence Audit
Arielle Jurchak Johnson, 2014 UC Davis doctoral dissertation. All 204 supplied sheets were examined, including front matter, every experimental chapter, references, supplementary tables and conclusions. Printed page 1 is PDF sheet 11; all locators below use PDF sheets. All 17 page grids were inspected and scientific figures and encoded tables were opened individually.
Scope correction: the review of the readable material is finished, but the source is not a fully legible dataset. Table S3.1 on PDF102 omits the final Scrappy Celery column for B33–B69; Table S7.1 on PDF198 clips the final koji column's numeric endings; PDF199 prints “###” for MV34/koji-kelp ethyl octanoate. Figure 7.4C on PDF191 has overlapping/clipped labels. A fresh download of the author's public dissertation reproduces the three table defects. No missing values were guessed. The earlier “No completion hold remains” statement is superseded: clean tables or original data remain necessary for exhaustive numerical reuse.
What Each Chapter Contributes
Introduction, PDF1–48: links chemical composition to perception, experimental sensory methods, multivariate analysis and culinary practice. It supplies an excellent vocabulary for explaining why a chromatogram is not a tasting note. Distinguish descriptive intensity, similarity, recognition and liking. The historical receptor-count and trillion-odor estimates are 2014 literature claims, not current teaching constants. Food-pairing criticism identifies limitations of recipe and ingredient databases; it does not show that shared compounds can never help a pairing.
Lavender GRO, PDF49–68: a Deans switch selectively diverts separated compounds, then a cryotrap recombines retained fractions for smelling. Ten mixtures were evaluated in triplicate by three experienced women: nine observations per condition, not nine independent assessors. Whole extract and a broad retained fraction elicited fresh-lavender language; smaller fractions changed the description. Figures 2.1–2.3 (PDF51–55) explain the apparatus and cuts; supplementary chromatograms (65) show excluded windows. This is a useful omission/recombination demonstration, limited by extraction selectivity, coelution, trapping and the tiny panel. A point displaced from the correspondence-map origin does not itself prove interaction. Failure to find a representativeness difference is not an equivalence test; Figure2.5 (58) has no error bars.
Sixteen commercial bitters, PDF69–105: 14 trained assessors, blinded black glasses/red light, replicated assessments and a Williams design provide considerably stronger product-description evidence than casual tasting notes. Thirty attributes were assessed; nutmeg was nonsignificant. The first two sensory PCA dimensions account for about71%, while two PLS components describe23% chemical and60% sensory variance. These are fitted descriptions, not external prediction accuracy or causal molecule-to-aroma assignments. The author explicitly acknowledges correlation versus causation (94–95). The 148 chromatographic entries include unidentified and questionable duplicate entries, so they are not 148 securely identified unique chemicals. Semi-quantitation uses 2-undecanone equivalents, not analyte-specific absolute concentrations. Products and formulations are historical samples.
Angostura omissions, PDF106–122: three assessors, four repetitions and five mixtures compare the whole extract, three single-compound omissions and a triple omission. Linalool, alpha-terpinyl acetate and caryophyllene removals change descriptor counts across several aroma terms. Table4.2 (114) uses yellow for decreases and magenta for increases. Twelve possible responses are repeated observations from three people. Small count changes without inferential uncertainty do not prove masking, enhancement or no effect. Pairwise omissions are absent, preventing a full factorial interaction decomposition. Putative OAVs7 and1.3 use relative headspace estimates and water thresholds;1.3 is not evidence of a subthreshold contribution in the actual mixture. Correspondence-map reflection is not a calculated causal effect.
Model Old Fashioneds, PDF123–145: a four-whiskey by four-bitters factorial study directly connects this dissertation to whiskey. Elijah Craig, Evan Williams, Rittenhouse and Old Overholt were combined with four bitters;14 trained assessors evaluated diluted, room-temperature mixtures. The design reveals combination-dependent oak and cola ratings (Figure5.2,133). It does not establish consumer preference, best recipes, or a general premium-price advantage: one example per price/type category and three Heaven Hill whiskeys constrain generalization. There are no whiskey-only controls here, sugar is omitted, and the sensory mixture at20%ABV/25°C differs from the chemical extraction at10% ethanol/40°C with salt. The author says the spiciness tendency was not significant (141). Treat the experimental drinks as model aroma mixtures rather than complete served cocktails.
Novel vinegars, PDF146–174: eight samples from seven substrate types combine descriptive analysis, volatile profiling and organic-acid measurement. Ten assessors used23 terms, with citrus/tropical fruit nonsignificant. Different extraction, sugar, initial ethanol, fruit concentration and common-starter conditions complicate causal attribution to raw materials. Table6.5 (159) distinguishes acetate from malate, citrate, oxalate and lactate: sourness is not evidence that acetic fermentation is complete. Lactic acid alone does not prove a newly established cofermentation. Shared volatiles may derive from the common starter. These experiments do not validate a preservation procedure or demonstrate consumer liking.
Malt vinegars, PDF175–200:16 culinary workers sorted16 vinegars once, then named their groups. DISTATIS, label correspondence analysis and MFA connect similarities, language and chemistry. Broad distinctions are practical, but no repeated sorting reliability or same-sample comparison against descriptive analysis is established; the author acknowledges reduced precision for small differences (195). Botanical additions and five malt/koji bases are not fully crossed. Chemical PCA is strongly influenced by juniper berry. Absence of detectable hop volatiles has several possible causes, not uniquely poor extraction. Comparing this three-month passive process with chapter6's five-day aeration does not isolate a time effect.
General conclusions, PDF201–204: the strongest contribution is a research strategy: combine instrumental and perceptual evidence, experimentally perturb mixtures, and choose sensory methods to fit the question and resources. It does not yield a universal chemical recipe for flavor, a validated predictive model of whiskey blends, or a ranking of what people enjoy.
Corrections and Reuse Limits
- Method transcription: PDF54 prints a25mm column internal diameter; PDF110 prints50µm internal diameter/0.32µm film. These are actual printed fields, not extraction artifacts. Chapter4's oven steps total approximately43.83min rather than49; chapters3/5/7 repeat an oven program totaling approximately78.33min rather than47. Clarify the original method before replication.
- Chemical identity: Table5.4 (134) gives whiskey lactone the same CAS as benzeneethanol, and sabinene the same CAS as beta-pinene; its isobutyl-alcohol identifier also conflicts with earlier tables. Other repeated name/CAS entries occur in chapter3. Authentic-standard flags do not repair mislabeled identifiers. Do not bulk-import these identities.
- Vocabulary and sample codes: Table3.2 reuses C1/C2 for citrus and celery. Table5.2 (125) gives an initial “banana” row an alfalfa-hay reference and later repeats banana;26 versus25 terms also needs clarification. PDF133 calls premium bourbon B2 although the sample key uses B1.
- Statistics and figures: PDF132 versus139 disagree about sensory standardization; Figure5.3 gives PC3 chemical variance13% while prose gives20%. Figure7.1B's horizontal-axis label conflicts with its caption. “amber.cascade” in figures is not explained by the plain-amber sample key. The malt chapter says75 compounds but tables list74. Map proximity, overlapping ellipses and nonsignificant tests must not be converted into certainty or equivalence.
- Supplementary inconsistencies: S6.1 lists licorice acetic acid as not detected despite Table6.5's highest acetate15.9g/L and prose claiming detection in all samples. Different assays preclude assuming the correction. Asparagus has81.390 for both acetic acid and ethyl octanoate; this deserves checking, not silent alteration. TableS3.1 has all-not-detected entries among supposedly detected compounds. Printed rounded concentrations do not exactly reproduce every claimed fold change.
- Safety and scope: historical bitters ingredients and exploratory vinegar production are research context, not approved ingestion or preservation instructions. No public recipe or course protocol has been changed.
Proposed Academy Applications and Connections
- Aroma mixtures lesson: redraw the GRO pathway as separation → selective omission → recombination → assessor response; use lavender and Angostura as bounded examples. Ask learners what an omitted compound can change besides its familiar isolated odor.
- Cocktail interaction exercise: compare selected whiskey/bitters combinations with coded order, fixed dilution/temperature and explicit whiskey-only controls. Record aroma intensity separately from liking; frame any exercise as an Academy proposal, not a replication already performed.
- Reading scientific maps: use chapter5 to contrast factorial interaction tests with PCA/PLS visualization. Include variance coverage, scaling, sample size and the difference between fitted relationships and prediction.
- Evidence audit exercise: compare the sample key, chemical identifiers and supplementary tables before trusting a polished figure. Missing and contradictory values remain visible in the evidence record.
- Cross-source synthesis: connect this work to
Your Tasting Brain — Full-Source Literature Note for perceptual context,
Preservation and Verification
Original Movies PDF retained unchanged:8,266,512bytes;SHA2569f191a5d1512062d12bd6f1e20d09b79d3d23e8625b4925323473af7c4ebbc3d. Existing native Notion attachment and all relations retained. Author-copy comparison is a defect check, not a second full reading; its10,421,564bytes/SHA2561442c4eb1eaf624d6753f278a3eb34a491dfd86f27a27ece7bf2fe399e6f6b0d differ from the supplied file. Proton exact-byte identity and fresh native-attachment hash remain unverified. API read-back is recorded separately from browser rendering. The remaining table gaps do not prevent processing the next source.
