Sources and evidence

The citation register, and an honest account of how far each citation has actually been checked.

Last reviewed: 9 August 2026

What BakeHQ stores, and what it does not

The register holds citation metadata and a locator — what a document is, and where in it the relevant part sits. It does not hold the text. There is no field for an excerpt, a quotation or a body, and a validation rule rejects any record that acquires one, because a citation register that stores source text is an unlicensed copy of other people’s writing.

What the verification statuses mean

This is the uncomfortable part and it is stated plainly. There is a real difference between knowing a document exists and having read the passage a claim rests on, and most citations in most reference writing are the first kind.

Citation checked, passage not read — the work exists and is described correctly: author, publisher, edition, subject matter. It does not mean anybody has opened it to the page being relied on.
Could not obtain the source — verification was attempted and the document could not be obtained. This is deliberately distinct from “citation checked”, which does not record whether anybody tried.
Passage not found where cited — the document was obtained and the cited passage was not where the locator said it was.
Passage verified — somebody read the named passage and it says what BakeHQ claims it says.
Source supports a narrower claim — somebody read it and it supports something narrower than BakeHQ claims. The claim is what gets narrowed.
Contradicted on checking — somebody read it and it does not support the claim. Such records are kept rather than deleted.

Of the 33 sources in the register, 25 have had the passage itself read, and 5 could not be obtained when verification was attempted. Every printed book BakeHQ cites is in that second group: no physical or licensed copy has been consulted, and those entries now say so rather than resting at “citation checked” as though nobody had tried.

Where a source still carries load unread, the entry below names the chapter a verifier must read and the statement that passage has to support.

Claims that carry no citation, and why

Not every statement is the kind a citation applies to. That a technique exists to do a particular thing is a definition; that removing the water stops fermentation restates what the fermentation page already cites. Attaching a source to those would inflate the count without adding evidence.

Until Sprint 7 all of them were simply labelled “no source cited”, which gave a definition and a genuine gap the same badge and let the pile grow unexamined. Each such claim now records why it carries no source, a validation rule refuses to publish one that does not, and the reason appears on the page beside the claim.

Follows from evidence recorded elsewhere on BakeHQ
True by definition rather than by measurement
Established practice, not a measured result
Not yet sourced — a known gap — the only one of the four that is a debt. It is counted, and it is meant to fall rather than accumulate.

Splitting them apart was worth doing: of the 105 claims carrying no citation, 104 need none, and 1 is genuinely outstanding.

Independence

A striking amount of what circulates as baking science is downstream of a handful of textbooks and popular-science authors, restated across thousands of pages without attribution. A register that counts URLs counts reach; one that collapses to distinct originals counts evidence.

Any source declaring itself republished, aggregated or translated must name what it is downstream of, and a validation rule enforces it.

Outstanding verification (8)

These sources currently support claims that would be strengthened — or would have to be narrowed — once the named passage is read. Each entry below states what must be confirmed and what to do if the passage says something narrower than BakeHQ does.

The register (33 sources)

On Food and Cooking: The Science and Lore of the Kitchen

Harold McGee, Scribner, 2004 · Reference work

Verification: Could not obtain the source
May carry a claim alone: No
ISBN: 9780684800011

Sprint 2 attempted verification and could not obtain the text. No licensed or physical copy was available, and no lawful full-text source was reachable. Bibliographic details remain confirmed; the passage remains unread. Recorded as access_failed rather than citation_checked so the register distinguishes 'nobody tried’ from 'tried and could not'.

Outstanding verification

Where: Chapter 10, 'Cereal Doughs and Batters'

Must confirm: That gluten forms from glutenin and gliadin on hydration and mechanical work, and that starch gelatinisation in wheat systems occurs over a temperature range rather than at a single point.

If it says less: Narrow the BakeHQ claim to whatever the passage actually states, and move any temperature figure to the more specific cereal-chemistry source rather than restating it from a general reference.

A general reference work rather than a primary source. Excellent for mechanism, weaker as a sole authority for specific numbers, which is why it may not carry a claim alone.

Baking Problems Solved

Stanley P. Cauvain and Linda S. Young, Woodhead Publishing, 2017 · Professional baking text

Verification: Could not obtain the source
May carry a claim alone: No
ISBN: 9780081007655

Sprint 2 attempted verification and could not obtain the text. Second edition, 2017; bibliographic details remain confirmed. A question-and-answer reference aimed at commercial bakers, so its answers are framed for industrial production and their domestic relevance is an inference BakeHQ makes rather than one the book states.

Outstanding verification

Where: Sections on bread faults: low volume, collapse and open crumb structure

Must confirm: That the named causes BakeHQ lists for low loaf volume and for collapse are among those this text attributes to those faults in commercial bread production.

If it says less: Where the text addresses only commercial production, mark the BakeHQ cause as applying to that context and do not generalise it to domestic baking without a second source.

Written for industrial and craft bakeries. Domestic relevance is usually strong but is not automatic, and BakeHQ should say which context a claim came from.

Technology of Breadmaking

Stanley P. Cauvain, Springer, 2015 · Professional baking text

Verification: Citation checked, passage not read
May carry a claim alone: No
ISBN: 9783319146867

Third edition, 2015. Bibliographic details confirmed.

The standard technical account of commercial breadmaking processes.

How Baking Works: Exploring the Fundamentals of Baking Science

Paula Figoni, John Wiley & Sons, 2010 · Food science textbook

Verification: Could not obtain the source
May carry a claim alone: No
ISBN: 9780470392676

Sprint 2 attempted verification and could not obtain the text. Third edition, 2010; bibliographic details remain confirmed. A teaching text, so it states consensus positions clearly and rarely cites primary experimental work for them — which caps how far a claim resting on it alone should be promoted.

Outstanding verification

Where: Chapters on flour, sweeteners and fats

Must confirm: That the functional roles BakeHQ attributes to sugar (tenderising by competing for water, promoting browning) and to fat (shortening gluten strands, contributing tenderness) are stated as functions of those ingredients in this text.

If it says less: Reduce the affected ingredient-function edges from professional_consensus to common_practice and record the discrepancy in the record’s note.

A teaching text, so it states consensus positions clearly but rarely cites primary experimental work for them.

Principles of Cereal Science and Technology

Jan A. Delcour and R. Carl Hoseney, AACC International, 2010 · Food science textbook

Verification: Could not obtain the source
May carry a claim alone: Yes
ISBN: 9781891127632

Sprint 2 attempted verification and could not obtain the text. Third edition, 2010; bibliographic details remain confirmed. The claims this source was carrying about starch gelatinisation have since been independently verified against an accessible peer-reviewed paper (renzetti-2025-gelatinisation), so they no longer depend on an unread passage.

Outstanding verification

Where: Chapters on starch and on wheat proteins

Must confirm: The gelatinisation temperature range given for wheat starch, and the statement that gelatinisation is a range-bounded transition dependent on water availability.

If it says less: Replace the BakeHQ figure with the exact range the text gives, including its stated moisture conditions. Do not carry a range across from another cereal.

Primary technical reference for cereal chemistry. Permitted as sole support for compositional and physical-chemical claims within its scope.

Handbook on Sourdough Biotechnology

Marco Gobbetti and Michael Gänzle (eds.), Springer, 2013 · Food science textbook

Verification: Could not obtain the source
May carry a claim alone: Yes
ISBN: 9781461454243

Sprint 2 attempted verification and could not obtain the text. First edition, 2013; bibliographic details remain confirmed. The sourdough microbiota and acidification claims resting on it are therefore still unverified at passage level, and remain the largest open verification item in the register.

Outstanding verification

Where: Chapters on the sourdough microbiota and on acidification

Must confirm: That mature sourdough is characterised by lactic acid bacteria together with yeasts, and that the balance of lactic and acetic acid — and therefore perceived sourness — is influenced by hydration and temperature.

If it says less: State the specific conditions the text reports rather than the general rule, and downgrade the sourness claim to professional_consensus.

An edited scientific volume. The strongest available support in the catalogue for sourdough microbiology.

Safer Food, Better Business for Caterers

Food Standards Agency · Government guidance

Verification: Citation checked, passage not read
May carry a claim alone: Yes
Link: www.food.gov.uk

UK Food Standards Agency guidance pack for caterers. Referenced for food-safety framing only.

Crown copyright, generally available under the Open Government Licence. Cited, never reproduced.

Used only where BakeHQ touches food safety. It is not a baking-science source and must not be cited as one.

The Bread and Flour Regulations 1998

United Kingdom, Statutory Instrument 1998 No. 141, 1998 · Standards body

Verification: Citation checked, passage not read
May carry a claim alone: Yes
Link: www.legislation.gov.uk

UK statutory instrument. Referenced for the legal composition and fortification requirements applying to flour sold in the UK.

Crown copyright, available under the Open Government Licence.

A legal instrument, not a scientific source. Supports what flour must contain in the UK, and nothing about how it behaves. Note that the Regulations have been subject to consultation and amendment; check currency before relying on a specific provision.

Hydrogen bond density and glass-transition temperature govern gelatinization and gel rheology in cereal and tuber starches

S. Renzetti, J. Henket, E. Raaijmakers, I. van den Hoek and R. van der Sman, Current Research in Food Science, 2025 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text obtained and read on 9 August 2026. Confirms that wheat starch gelatinisation is a transition with distinct onset, peak and end temperatures rather than a single point, reports a wheat peak of 61.5 °C at a starch mass fraction of about 0.2, and models gelatinisation temperature as decreasing systematically with increasing water content — which is the same relationship BakeHQ states from the other direction, that limited water raises the temperature.

Volume 10, article 101101. Replaces an unread textbook chapter as the support for BakeHQ’s two gelatinisation claims.

Effect of storage temperature on starch retrogradation of bread staling

J. F. Aguirre, C. A. Osella, C. R. Carrara, H. D. Sánchez and M. del P. Buera, Starch/Stärke, 2011 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes

Bibliographic record and abstract obtained and read on 9 August 2026; the full text is paywalled and was NOT read. The abstract states that at 4 °C the rate of starch retrogradation is the highest of the conditions studied, that water activity fell over time at both 25 °C and 4 °C while remaining almost constant at −18 °C for 23 days, and that at −18 °C only crystal enlargement occurred whereas warmer storage produced both growth and new crystal formation. Volume 63, issue 9, pages 587–593.

Outstanding verification

Where: Full text, results and discussion

Must confirm: The magnitude of the difference in retrogradation rate between 4 °C and 25 °C, and whether the finding holds for the bread formulations BakeHQ describes rather than only the one studied.

If it says less: Keep the directional claim, which the abstract supports, and drop any implication about how much faster refrigeration stales bread.

Verified at abstract level only. The directional claim is supported; no quantitative claim should be made from this without reading the full text.

Diverse Microbial Composition of Sourdoughs From Different Origins

A. Comasio, M. Verce, S. Van Kerrebroeck and L. De Vuyst, Frontiers in Microbiology, 2020 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text obtained and read on 9 August 2026. Confirms three things BakeHQ now relies on: the LAB-to-yeast ratio in sourdough is 'mostly 10:1 to 100:1'; more than 90 LAB species and more than 40 yeast species have been isolated from sourdoughs worldwide; and 'neither the nature of the flour nor the geographical region of the producers impact the microbial community structure of mature sourdoughs', while dough yield, pH and temperature are of direct impact on which LAB species are present. The paper does NOT state a typical pH for mature sourdough, and BakeHQ makes no pH claim from it.

Volume 11, article 1212. The record of what this source does not say is deliberate: a widely-repeated figure for stable sourdough pH was checked against it and was not there.

Composition and Properties of Aquafaba: Water Recovered from Commercially Canned Chickpeas

Y. Y. Shim, R. Mustafa, J. Shen, K. Ratanapariyanuch and M. J. T. Reaney, Journal of Visualized Experiments, 2018 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text obtained and read on 9 August 2026. Confirms that the functional properties of aquafaba from ten commercial products — particularly foam volume and foam stability — 'varied significantly', that after one hour liquid had separated from every brand’s foam except one, that the viscosity coefficient of variation across products was 160%, and that freeze-dried aquafaba protein content ranged 23.3–26.8% between brands. It contains NO direct experimental comparison between aquafaba and egg white, so it cannot support any equivalence claim. The authors state that input material has been 'incompletely characterized and standardized'.

Outstanding verification

Where: A controlled comparison of aquafaba against egg white in a defined bake

Must confirm: Whether aquafaba matches egg white for aeration in a real formulation, and by how much it falls short if it does not.

If it says less: Keep the claim that aquafaba can aerate and that its performance varies by brand; make no claim about equivalence to egg white.

Issue 132, article 56305. Supports the variability and composition claims. Does not support equivalence with egg white, which is the claim most aquafaba writing actually makes.

Enhancing Gluten-Free Bread Production: Impact of Hydroxypropyl Methylcellulose, Psyllium Husk Fiber, and Xanthan Gum on Dough Characteristics and Bread Quality

R. Torres-Pérez, E. Martínez-García, M. M. Siguero-Tudela, P. García-Segovia, J. Martínez-Monzó and M. Igual, Foods, 2024 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text obtained and read on 9 August 2026. Confirms, with figures: raising psyllium increased crumb hardness (64.7 N against a 58.9 N control) and gumminess (40.7 N against 35.3 N) and lowered cell count (8 against 11 cells/cm², i.e. larger cells); REMOVING xanthan gum lowered hardness to 46.3 N, lowered gumminess to 27.8 N and RAISED cell count to 12 cells/cm², giving a finer structure, while producing a less elastic dough. Raising HPMC reduced extrusion force without significantly affecting bread quality. States that gluten 'forms a unique network that creates a viscoelastic dough capable of retaining gases'. The paper does NOT report specific volume in tabulated form, and BakeHQ claims nothing about it from this source.

Outstanding verification

Where: Replication across other gluten-free flour bases and formulations

Must confirm: Whether the direction of the xanthan finding holds outside this formulation, or is specific to the flour base and the other hydrocolloids present.

If it says less: Keep the claim that more hydrocolloid is not reliably better and that xanthan raises gumminess in at least one measured formulation; drop any implication that removing xanthan improves every gluten-free bread.

Volume 13, issue 11, article 1691. One formulation studied under controlled conditions; the authors note these do not fully replicate industrial conditions and that water quantity was judged visually.

Synergistic Effects of Hydrocolloid Combinations on Gluten-Free Batter and Bread Characteristics

M. Parsamajd, M. Fazaeli, M. Majdinasab and M.-T. Golmakani, Food Science & Nutrition, 2025 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text obtained and read on 9 August 2026. Supplies the mechanism BakeHQ’s gas-retention function rests on: hydrocolloids’ 'high molecular weight enables the retention of gas cells during fermentation and baking, mimicking gluten’s role in the dough matrix'. Also states that gluten contributes 'extensibility, elasticity, water absorption, and gas-holding capacity’ — four separable jobs, which is the basis of BakeHQ’s framing. Reports that increasing xanthan gum DECREASES specific volume of gluten-free bread through increased elasticity and reduced extensibility, and that combinations are not additive: an HPMC-xanthan blend gave the lowest hardness while an HPMC-guar blend gave the hardest, least resilient bread. Hydrocolloids were used at 2% on flour weight. The paper does not discuss psyllium.

Volume 13, issue 10, article e71107. Authors state the main limitation is the absence of sensory evaluation, and that pilot-scale validation was not performed.

Oats and the gluten free diet

Coeliac UK, 2026 · Government guidance

Verification: Passage verified
May carry a claim alone: Yes
Link: www.coeliac.org.uk

Page obtained and read on 9 August 2026. Confirms: oats contain avenin, 'a protein similar to gluten'; 'most people with coeliac disease can tolerate gluten free oats with no problems’ while 'a very small number’ may still be sensitive; 'many standard oats are produced in the same place as wheat, barley and rye, which makes them unsafe because they can be contaminated'; and 'uncontaminated oats that test at 20 parts per million (ppm) gluten or less may be labelled gluten free'. On personal suitability it says 'it is up to you to decide’ and directs anyone with ongoing symptoms to review it with their health professional. NO PERCENTAGE for avenin sensitivity appears on the page, and BakeHQ states none.

A patient-facing charity guidance page, classified as government_guidance because it functions as national guidance rather than as primary research. It is authority on labelling and on what people are advised, and is not evidence about baking behaviour.

Effects of Physical and Chemical Factors on the Structure of Gluten, Gliadins and Glutenins as Studied with Spectroscopic Methods

K. Kłosok, R. Welc, E. Fornal and A. Nawrocka, Molecules, 2021 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 10 August 2026. Section 2 states that “glutenins are responsible for dough strength and elasticity” and that “gliadins contribute to the viscosity and extensibility of the dough”, and describes gluten as “a continuous and viscoelastic network that is formed during mixing process”. Section 4.1 adds that the network forms “as a result of hydration, depolymerisation and re-polymerisation processes”, which is the hydration-and-working half of BakeHQ’s claim.

Volume 26, article 508. Replaces two unread textbooks as the support for the glutenin/gliadin division of labour — the claim the Explorer’s gluten-network mechanism rests on.

The Thermodynamic and Gelation Properties of Ovalbumin and Lysozyme

L. Wang, R. Li, S. Lv and Y. Liu, Gels, 2025 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 10 August 2026. Section 2.5 reports that “Compared to the control (water), LYZ and OVA exhibited significantly higher Td and ΔH values in the presence of sugars” across 0.05–0.5 M, concentration-dependently, with lysozyme rising to 80.08 °C in sucrose against a water control of 77.46 °C. SCOPE IS NARROWER THAN THE CLAIM: these are isolated egg proteins in solution measured by DSC, not whole egg in a batter. The direction is measured; the magnitude in a cake is not.

Volume 11, article 470. Replaces an unread teaching text as the support for sugar delaying the egg set.

Maillard Reaction in Flour Product Processing: Mechanism, Impact on Quality, and Mitigation Strategies of Harmful Products

Y. Qi, W. Wang, T. Yang and W. Ding, Foods, 2025 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 10 August 2026. Section 3.1.1 describes “carbonyl groups (C=O) of reducing sugars… undergo a nucleophilic addition reaction with the amino groups (-NH2)”, establishing both required reactants. Section 3.2.4 states that “Within the pH range of 3 to 9, the reaction rate generally increases with pH” and that an “acidic environment promotes hydrolysis… inhibiting further Maillard progression”. Section 3.2.3 treats caramelisation as a separate process “competing with the Maillard reaction for substrates”. NOT FOUND: an explicit sentence that sugar alone caramelises without an amino group; BakeHQ states that as following from the reactant requirement rather than as quoted.

Volume 14, article 2721. Chosen over a general food-chemistry review because it addresses flour products specifically.

Genetic evidence that high noninduced maltase and maltose permease activities, governed by MALx3-encoded transcriptional regulators, determine efficiency of gas production by baker’s yeast in unsugared dough

V. J. Higgins, M. Braidwood, P. Bell and P. Bissinger, Applied and Environmental Microbiology, 1999 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 10 August 2026. The Introduction states directly: “In a dough consisting of flour, water, yeast, and salt (unsugared or plain dough), the most abundant available sugar is maltose, produced by the action of amylases on damaged starch.” That is BakeHQ’s claim almost word for word, and it comes from a paper whose whole subject is gas production in unsugared dough.

Volume 65, pages 680–685. Replaces an unread cereal-science textbook as the support for lean dough fermenting without added sugar.

FoodData Central, SR Legacy 171265 — Milk, whole, 3.25% milkfat, with added vitamin D

U.S. Department of Agriculture, Agricultural Research Service, 2019 · Standards body

Verification: Passage verified
May carry a claim alone: Yes
Link: fdc.nal.usda.gov

Queried directly against the FoodData Central API on 10 August 2026 rather than read from a nutrition aggregator. Water is reported as 88.13 g per 100 g, derived from 28 data points with a stated minimum of 87.69 and maximum of 89.20. THE RANGE AND THE SAMPLE COUNT ARE THE REASON THIS SOURCE WAS USED: they let BakeHQ state that whole milk is roughly seven-eighths water without implying a constant, which is exactly the false precision the hydration model exists to avoid.

Used for the one water fraction BakeHQ states numerically. Secondary aggregators reproducing USDA figures were deliberately not cited — a restatement of a measurement is not the measurement.

The Chorleywood Bread Process

Campden BRI, 2026 · Trade publication

Verification: Passage verified
May carry a claim alone: Yes
Link: www.campdenbri.co.uk

Read on 10 August 2026. THE ORIGINATING INSTITUTION’S OWN ACCOUNT — the process was developed in 1961 by what was then the British Baking Industries Research Association, which became Campden BRI. States mixing energies for standard UK bread flours of “around 11 Wh/kg dough”; a typical high-shear mixing time “around three minutes… two to four minutes depending on flour strength, compared with 10-15 minutes for many other types of dough mixer”; that it “enables a higher proportion of UK wheat to be used in the flour grist”; and that the resulting bread “has a finer (and whiter) crumb structure and is softer because of the thinner walls between gas cells. It will stay softer for longer.” A WIDELY REPEATED CLAIM WAS CHECKED AND NOT FOUND: that the energy must be delivered within five minutes. Campden BRI’s own page states no such time requirement, and BakeHQ does not repeat it.

An industry research body writing about a process it invented. Treated as authoritative for what the process IS and for its own figures, and NOT as neutral evidence on whether the bread it produces is good — that distinction is made explicitly wherever this source is cited.

Technological and Nutritional Aspects of Bread Production: An Overview of Current Status and Future Challenges

M. Mesta-Corral, R. Gómez-García, N. Balagurusamy and C. Torres-León, Foods, 2024 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 10 August 2026. Section 4.3 names the emulsifiers used in bread — “diacetyl tartaric acid ester of monoglycerides (DATEM), sodium stearoyl lactate (SSL), polysorbates, and mono- and diglycerides” — and states that they “assist dough strengthening and crumb softening by interacting with the starch”. The same section states that “α-amylases have been used to improve the fermentation process, increase the volume of bread, and decrease the hardening of bread”. NOT FOUND: any discussion of the Chorleywood Bread Process, which was searched for specifically.

Volume 13, article 2062. Supports the technological function of bread emulsifiers and amylases, and nothing about their health effects.

Clean Label in Bread

M. C. A. Vargas and S. Simsek, Foods, 2021 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 10 August 2026. Section 2 states that the primary function of dough strengtheners and preservatives is “to increase machine, and mixing, and handling tolerance”, and enumerates improver components including vital wheat gluten, oxidising agents, reducing agents, enzymes and emulsifiers. Section 2.3 describes ascorbic acid as “an intermediate oxidizer” that “promotes dough development at a high mixing speed… improves the dough strength, reduces dough stickiness, and improves gas retention”. Section 5.1.2 states that “heat-stable amylases help reduce staling in bread” and that “enzymes are seen as more natural and fit clean-label standards”. SCOPE NOTED: the paper does not give the molecular mechanism by which amylases slow staling, and BakeHQ does not assert one.

Volume 10, article 2054. Chosen because its subject is precisely the consumer-facing question the demand asks — what these ingredients are and why they are there.

Evaluation of Plant-Based Milks in Vegan Muffins: Functional, Structural, Rheological and Nutritional Characterization

K. T. Günan and P. Y. Ömeroğlu, Foods, 2025 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 10 August 2026. Eight plant milks — “soy, hazelnut, walnut, quinoa, flaxseed, coconut, oat, and almond” — were compared against cow’s milk in muffins. THE HEADLINE FINDING CUTS AGAINST COMMON ADVICE: “Physical parameters such as baking loss, volume index, and symmetry revealed no significant structural differences (p > 0.05), confirming that milk substitution did not affect baking performance.” Batter rheology DID differ — coconut, oat and flaxseed milks gave “higher shear stress and viscosity values”, almond and hazelnut “lower shear stress and viscosity”. All plant formulations were lower in protein than the dairy control (5.40–5.92 vs 8.51 g/100 g). SCOPE IS ONE PRODUCT: these are muffins, and the paper draws no causal link between the compositional differences and baking performance.

Volume 14, article 3989. Cited for what it measured — a muffin — and deliberately not extended to bread, pastry or custard, where protein and fat differences plausibly matter far more.

A Systematic Comparison of the Intrinsic Properties of Wheat and Oat Bran Fractions and Their Effects on Dough and Bread Properties: Elucidation of Chemical Mechanisms, Water Binding, and Steric Hindrance

S. Renzetti, M. Theunissen and K. Horrevorts, Foods, 2021 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 13 August 2026, and IT CORRECTED A CLAIM BAKEHQ HAD ALREADY PUBLISHED. The conclusions state that “the interplay between water binding, mainly provided by the insoluble fraction, and the plasticizing properties of the soluble bran fraction control the detrimental effects on bread volume”, and that “steric hindrance and gluten dilution seemed to play a minor role”. The experimental design is what makes this decisive: glass beads of comparable particle size but with no water-binding capacity produced minimal change, so the authors conclude the differences were “mainly responsible… rather than steric hindrance”. Brans were milled to a comparable 200–300 μm distribution.

Volume 10, article 2311. Cited specifically because it isolates the mechanism experimentally rather than describing the effect — which is what let it overturn the physical-cutting explanation BakeHQ had been repeating.

Rye Bread Defects: Analysis of Composition and Further Influence Factors as Determinants of Dry-Baking

Foods, 2020 · Peer-reviewed paper

Verification: Source supports a narrower claim
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 13 August 2026. SUPPORTS THE STRUCTURAL CLAIM AND NOT THE ENZYME ONE. The introduction describes rye’s hemicelluloses — arabinoxylan and beta-glucan — as “colloidally dissolved in water”, forming a viscous liquid phase in place of wheat’s protein network, which is what BakeHQ says about how a rye dough holds together. WHAT WAS SEARCHED FOR AND NOT FOUND: a direct causal statement linking amylase activity to a specific crumb defect. The paper discusses falling number at length and presents it as a parameter that fails to predict rye baking quality reliably, which is close to the opposite of a clean amylase-to-gumminess link. BakeHQ therefore keeps the pentosan claim as supported and records the amylase claim as professional consensus rather than upgrading it on this source.

Recorded as source_narrower deliberately: it supports part of what BakeHQ says about rye and specifically does not support the rest, and the register should show that rather than an unqualified tick.

Effect of wheat grain protein composition on end-use quality

A. Sharma, S. Garg and I. Sheikh, Journal of Food Science and Technology, 2020 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 14 August 2026, BEFORE the flour content was written. The classification section states that “bread making quality is highly affected by the type and amount of gluten proteins”, distinguishing type from amount explicitly. The glutenins subsection states that “the functional and rheological properties of dough relies on the ratio of glutenin/gliadin, ratio of high/low glutenin polypeptide, gliadin binding strength to glutenins”. The paper also states that “higher protein content will improve the bread making quality”, so it supports BOTH quantity and composition mattering — “the viscoelasticity of the dough is ensured by the quantity and quality of the wheat storage proteins”. BakeHQ therefore does NOT claim that protein percentage is irrelevant; it claims that percentage alone does not determine strength.

Cited for the quantity-versus-quality distinction. It deliberately does not support the stronger folk claim in either direction — neither “protein percentage is everything” nor “protein percentage tells you nothing”.

The Properties of Damaged Starch Granules: The Relationship Between Granule Structure and Water–Starch Polymer Interactions

A. G. Teobaldi, E. J. Carrillo Parra and G. N. Barrera, Foods, 2024 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 14 August 2026. The introduction gives the figures that settle a common misconception: “Damaged granules absorb between 200 and 430% of their weight in water at room temperature, while native granules absorb between 39 and 87%”. It also states that damaged starch increases “the flour’s water absorption capacity”, and that the condition “makes damaged starch granules more accessible to amylases hydrolysis, compared to native ones, which results in the production of dextrins”. THE RANGES ARE WIDE AND ARE QUOTED AS RANGES — the useful finding is the order-of-magnitude gap between damaged and intact, not any single figure.

The evidence behind BakeHQ’s position that flour absorption is a milling property as much as a protein one.

Impact of Baking Powder and Leavening Acids on Batter and Pound Cake Properties

E. A. Asamoah, A. Le-Bail and A. Oge, Foods, 2023 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: pmc.ncbi.nlm.nih.gov

Open-access full text read on 14 August 2026, to replace two unread textbooks under one of BakeHQ’s most load-bearing claims. Section 3.2.1 states that “batter expansion involves the previously incorporated air bubbles during mixing, which now act as a nuclei for CO2 produced from BP to diffuse into, causing them to expand and grow in size”. SCOPE: measured in a chemically leavened pound cake batter, not a yeasted dough. The claim BakeHQ makes — that leavening gases expand pre-existing cells rather than creating new ones — is supported here for chemical leavening specifically, and the same mechanism in fermented dough remains professional consensus rather than verified by this source.

Volume 12, article 946. Upgrades a Tier A claim from two unread textbooks to a read passage, with the scope limit stated rather than glossed.

The Study of Bread Staling Using Visible and Near-Infrared Reflectance Spectroscopy

F. Xie, Department of Grain Science and Industry, Kansas State University, 2002 · Doctoral thesis

Verification: Passage verified
May carry a claim alone: Yes
Link: www.ars.usda.gov

Full text read on 15 August 2026 from the copy hosted by USDA ARS. TWO SEPARATE PASSAGES WERE USED, and the distinction between them matters. The literature review states that “retrograded AP is more sensitive to heat than retrograded amylose” and that “if bread is reheated, AP crystallinity is disrupted again and bread is re-freshened” — that part is the author restating Zobel and Kulp (1996), which BakeHQ has NOT read. The author’s OWN calorimetry in section 3.4 is what BakeHQ actually relies on: “DSC easily tracked starch retrogradation in bread. Amylopectin melting peak temperature ranged from 50 °C to 60 °C for bread stored at low temperature and from 60 °C to 70 °C for bread stored at high temperature.” Section 3.3 records that the amylopectin enthalpy was integrated between 50 °C and 80 °C and the amylose–lipid complex between 105 °C and 120 °C. THAT SECOND FIGURE IS WHY THE CLAIM IS NOW NARROWER: domestic reheating passes through the amylopectin melting range and nowhere near the amylose–lipid range, so it can only undo part of the firming.

Cited for measured melting temperatures of retrograded amylopectin in real bread crumb, not for the reheating anecdote in its literature review. The two are recorded separately above so a later reader can tell which is which.

Moisture migration in multicomponent food systems: Investigating the effect of temperature

P. Mavrou, University of Surrey, 2022 · Doctoral thesis

Verification: Passage verified
May carry a claim alone: Yes
Link: openresearch.surrey.ac.uk

Full text read on 15 August 2026 from the University of Surrey open research repository. Section 1.1 states that “moisture migration occurs when two food materials of unequal water activity are in contact”, and that “moisture migrates from the material with the higher water activity to the material with lower water activity until equilibrium is reached”. It attributes the driving force explicitly: “the driving force in this case is a difference in chemical potential which for foods is commonly expressed by their water activity”. Section 2 makes the modelling consequence unambiguous — the transport equation’s gradients “should be expressed in terms of the driving force which is water activity”. Section 6.4 shows the point BakeHQ had been getting loosely: at equilibrium the moisture CONTENT is discontinuous across the interface while the water activity is continuous, because “both materials are considered to be at equilibrium at the start of the simulation (equal water activity) and the driving force for moisture migration is the difference in water activity”. Section 5 records the companion observation that a compound “absorbs significantly smaller amounts of moisture at the same water activity”.

The evidence behind BakeHQ’s correction from “more available water” to water activity, and behind the statement that two foods can hold very different amounts of water and still not exchange any.

Effect of Heat on Egg White Proteins

Z. Akkouche, L. Aissat and K. Madani, 3BS Laboratory, A. Mira University of Bejaia; International Conference on Applied Life Sciences, 2012 · Conference paper

Verification: Passage verified
May carry a claim alone: Yes
Link: cdn.intechopen.com

Full text read on 15 August 2026. THIS SOURCE WAS CHOSEN BECAUSE IT TESTED THE ACTUAL QUESTION rather than asserting the answer. Most texts state that coagulation is irreversible; this one measured whether it reverses. Section 3.1 reports that a Student test “indicates no significant differences between the transmittance measured immediately after heat and after 24 hours storage at 4 °C”, and concludes that “the phenomenon of protein denaturation in albumen is irreversible and interactions occurring between them are covalent (disulfide bridge) and non covalent, (hydrophobic and electrostatic interactions)”. Section 3.2 adds that “the heat applied to solutions of egg white lead to irreversible loss of solubility”, with only 23.44 ± 0.21 % of protein still soluble after 20 min at 89 °C. SCOPE, STATED RATHER THAN GLOSSED: this is egg white in solution measured by turbidity and solubility, not the mechanical behaviour of a set gel, and it says nothing about gluten.

Cooling to 4 °C for 24 hours is a direct test of reversal on cooling, which is precisely what BakeHQ’s claim asserts. Recorded as a conference paper, which cannot promote a claim to established science alone.

Lipid shortenings: a review

B. S. Ghotra, S. D. Dyal and S. S. Narine, Food Research International, 2002 · Peer-reviewed paper

Verification: Passage verified
May carry a claim alone: Yes
Link: talcottlab.tamu.edu

Full text read on 15 August 2026, BEFORE any of the Sprint 15 fat content was written. FOUR PASSAGES CARRY LOAD. Section 3 defines the shortening effect mechanically: “the term ‘shortening’ refers to the ability of a fat to lubricate, weaken, or shorten the structure of food components”, and “if shortening is present, the fat breaks the continuity of the protein and starch structure”, which “enables the lubrication of gluten particles”. Section 5.6 defines plasticity: “the term ‘plastic’ implies a solid, non-fluid, non-pourable, and non-pumpable shortening at room temperature”, and records that bread shortening is “about 3% based on flour weight”. Section 9 ties plastic range to the solid-fat-content curve: “an oil with a steeper SFC profile typically has a very narrow plastic range whereas a fat system with a flat SFC profile typically has a wide plastic range”. Section 5.5 and section 10.3 together give the crystal-form finding: beta crystals form “large plates of solid fat” that “are much less effective in entrapping dispersed air”, while “small beta-prime crystals are preferred due to the increased ability to incorporate air bubbles”, and a beta-prime to beta transition on standing produces “a loss in creaming ability of the shortening”.

Volume 35, pages 1015–1048. The evidence behind BakeHQ’s position that fat STATE and crystal form, not the name on the packet, decide what a fat can do — a fat can lose its creaming ability without its composition changing at all.

Corrections

If a citation here is wrong, or a passage does not say what BakeHQ claims, that is worth knowing more than almost anything else on the site. See corrections.