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 27-page report arguing that glass geometry can reduce ethanol interference and improve spirits evaluation.
Author argument
Manska applies airflow, evaporation, and olfactory arguments to advocate an engineered vessel over traditional copita-derived forms, claiming improved aroma delivery, reduced ethanol numbing, and competition adoption.
Researcher synthesis
The source raises a legitimate methodological question—glass geometry alters the vapor presented to the nose—but product-specific superiority claims should be treated as hypotheses until replicated in blinded independent studies. Standardizing one vessel within a trial remains essential.
Evidence assessment
Published technical report with useful physical reasoning, diagrams, and testable design variables.
Limitations and open questions
Limitations: Single author is affiliated with the commercial manufacturer; evidence is largely product advocacy, physical inference, and adoption history rather than independent blinded comparison; performance and award claims need corroboration.
Open questions: What independent blinded crossover studies compare common whiskey glasses at controlled fill, proof, temperature, and nosing behavior?
Connected records
- Contributors: 1
- Verified Excerpts: 3
- Citations: 1
- Zettels: 2
Completion record
September27: all27 main pages,8tables/allcells,9figures/alllabels,2equations,16refs read. Later AppendixXIV screens624–625 directly checked;2DOI mismatches confirmed. S1homepage bounded text read; S2/S3video sameURL/throttled/unviewed. Main complete; full package remains partial. Original/relationspreserved; cloudhash/browserunverified.
Later author correction — 2025
Manska's later book, Appendix XIV (supplied PDF screens 624–625), revises his earlier ethanol receptor-binding explanation, diffusion treatment and surface-tension account. Preserve the earlier report as a historical version, but do not reuse those mechanisms without this correction. The author's continuing confidence in the vessel is not independent validation of its performance.
Later critical analysis: 
Full article audit — Manska (2018), nosing-vessel design
SRC-404 / LIT-328; DOI10.3390/beverages4040093. September27,2026.
All27 supplied pages read sequentially, all8 tables/every cell, all9 figures/labels/dimensions, both equations, 16 references, funding/conflicts/patents. Rendered sheets2,3,4,6,7,8,9,11,17,18,19,20,21,22,23 inspected. Main article complete; promotional supplementary videos remain inaccessible and are not counted as viewed.
Argument, provenance and methods
Arsilica cofounder and R&D head George Manska advocates the NEAT glass. The author/company funded the work and sell the design. This conflict is explicitly disclosed, not proof that every observation is wrong. It does demand separation of observed preference, hypothesized physical mechanism and diagnostic superiority.
Pages1–10 introduce ethanol interference and criticisms of traditional glasses. Pages10–17 describe small expert panels (usually8 from23), event preference challenges and judging simulations. Pages17–23 derive dimensions and diagrams; pages23–26 give a use protocol, commercial adoption claims, conclusions and patents; pages26–27 disclose funding and list references.
Study-by-study assessment
Table1 gives pure-water/ethanol properties, not measurements of whisky headspace composition. Bulk ABV, pure-component vapor pressure and receptor exposure are different quantities.
Table2's64/36 tumbler,90/10 copita and79/21 snifter ethanol/character-aroma percentages are subjective estimates, not GC-MS composition. Vodka results were excluded for high variance. Fixed tumbler→Glencairn→copita→snifter order confounds glass with exposure/order. Styling, handling, fill and nose position differ.
Table3's17 event rows sum correctly to2,914:272+1,830 males and14+798 females; preferences12.94% versus1.72%. These are attendees who agreed to participate, not a representative population or threshold experiment. Engineered glass always came first, brands were visible, and approximately30% of invited women declined. Glass preference cannot identify ethanol preference alone or prove innate sensory sensitivity. Standard deviations between events are not automatically confidence intervals for population effects. The date range includes2010 despite prose2011–2015, and San Diego August2014 appears twice with distinct counts; retain as unexplained rather than silently deduplicate.
Table4 reports pooled unique descriptors (bourbon10vs7;Scotch7vs5;gin5vs3), not per-person accuracy. C-STEM always preceded combined orthonasal/retronasal evaluation. Repetition, extra observation and priming are confounded. Page14's second bourbon trial adds4 to7=11, unlike table's10. Counting more descriptors is not proving they are correct.
Table5 has four panels of3–5judges and a fixed vodka→gin→whiskey→whiskey→break→rum order. Changing products, flight sizes, elapsed time and break simultaneously prevents identification of an ethanol-specific causal mechanism. Spitting and appearing unimpaired are not measured blood ethanol. No matched NEAT crossover trial supports the contrast with100 reported events; absence of complaints is not a measured zero incidence.
Table6 (8panelists×6distances) directly conflicts with the claim that2.5cm captures all aromas:6/8report additional descriptors at2cm and one at1cm. Filling upward changes volume/headspace as well as distance; prior notes remain visible and distances always decrease. No blinded molecular identification maps these descriptors to molecular mass.
Table7 lists23 compounds. The cis/trans whisky-lactone entries incorrectly have different masses264.26/208.21; NIST lists cis whisky lactone156.2221. Ethyl propanoate211.28 is inconsistent with FAO102.13; 2-methyl propanol is labeled an aldehyde. Recalculating square-root ratios from erroneous masses cannot validate actual spatial separation. Table7 does not include concentration, partitioning, air diffusion coefficients, convection or measured concentration fields.
Table8's34.5mm/17mm triangle yields38.45mm as printed, but source samples/distributions for facial measurements are unspecified. An average does not guarantee fit for all faces.
Diagrams and specification contradictions
Figures1–2 show glass types and test vessels. Figures3–6 illustrate the proposed mass-dependent cloud/separation mechanism, not measured multicomponent maps. Figure6's pi dilution factor is not established by a concentration experiment.
Figure7 shows the facial geometry assumption. Figure8 labels best fill60mL(1½oz), inconsistent with repeated45mLtext and its own fluid-ounce conversion. Figure9 specifies79±1mm rim,56±1mm neck,80±1mm bowl,67±1mm height,48.8±1mm neck height,55±1mm base,6±2mm bottom,1.2±.2mm wall. Text section2.11 instead gives51mm neck. A51/79 diameter ratio gives41.68% remaining area, or58.32% reduction, not40% reduction.
The kinetic-speed and Graham-ratio equations are not a demonstrated model for this open, macroscopic, convecting glass. The proposed receptor lockout, gravity/mass aroma sorting and universal ban on added water must not be taught as established mechanisms.
Citation and later-correction checks
Page8 cites Spence2015 as support for75–95% smell contribution, then selects90% for spirits without measurement. Spence's actual article questions the evidentiary basis of that numerical claim. No fixed aroma/taste/mouthfeel percentages should be inferred here.
The later supplied Distilled Spirits—Sensory Science PDF has740 screens. AppendixXIV on screens624–625 (reader printedpage349) was directly read and visually inspected in this audit, confirming the existing library link. It retracts the receptor-binding account and surface-tension explanation and acknowledges molecular-mass errors, while maintaining the author's confidence in the design. This is an author correction, not independent replication. Do not treat its replacement explanation as established merely because it replaces an error.
Two references in that correction have mismatched DOIs:10.1016/j.foodres.2020.109412 identifies Nguyen et al., fat crystallization of palm-oil/milk-fat blends;10.1093/chemse/27.3.215 identifies Kinoshita et al., mouse olfactory-pathway labeling, not the stated Hummel/Livermore paper. Crossref metadata and primary publication records confirm the mismatch. Dalton DOI10.1093/chemse/25.6.733 did not resolve in this check; recorded unresolved, not declared nonexistent.
Only this appendix was reread now; the740-screen book is not newly certified by this article audit.
Supplementary boundary
The article's S1 is the commercial NEAT website, not a fixed dataset. The current homepage text was read as a bounded vendor page on September27: sales claims and products, not independent validation or the2018 site snapshot. No claim to exhaustively read the website or inspect every page/media.
S2 and S3 have different titles but the exact same YouTube URL in the supplied PDF:5UpDRxxx2Bg. Fetch and oEmbed were throttled. Neither video has timestamp/visual coverage; underlying manufacturing footage identity is unresolved. Publisher article fetch returned429. These access gaps remain open while other library work continues.
Academy contribution and cross-source synthesis
Use as an advanced evidence-audit case and source of testable service variables; do not adopt its glass endorsement, universal2.5cm threshold,90%flavor arithmetic, sex-based ability ranking or no-water rule.
Compare with Karlsson–Friedman and Ashmore dilution studies: a simulation, measured headspace response, trained-panel description and consumer preference answer different questions. Their bounded results do not validate the2018 surface-tension shutdown claim.
Proposed Academy study: randomized counterbalanced within-person glass comparison, coded spirits, controlled proof/volume/temperature/rest, standardized instructions, recorded nose distance and sniff behavior, replicated sessions. Separate preference, irritation, descriptor count, discrimination and repeatability; use predefined outcomes and an independent administrator. Glass geometry prevents perfect vessel blinding, so measure expectations and acknowledge the limitation.
Keep the original diagrams as attributed hypotheses for critique; a neutral Academy diagram should map controlled variables to measured outcomes rather than depict unverified molecular sorting.
Existing-record corrections
First evidence paraphrase correctly attributes the author's argument. Second needs narrowing from generic standardization advice to the actual protocol on23–24. Third falsely presents the author as withholding universal-optimality claims; replace with the genuine tension between strong endorsement and admitted need to quantify onset/causes. Preserve all identities, originals and links. Main content review is complete; supplementary media, cloud-byte identity and browser rendering are separate unresolved layers.
