Recipe ingredient density: how 6 reference sources disagree by 30%

A volume measurement of a dry ingredient depends on how packed the ingredient is when the cup is filled. The same cup of flour can hold 120 g (spooned and levelled, loose) or 150-160 g (scooped directly from the bag, mildly packed). Different reference sources use different conventions; the result is that “1 cup of flour” means a different gram weight depending on whose book you opened. This piece collects six widely-cited references and shows exactly where they agree and disagree.

The all-purpose flour spread

Six reputable sources’ published values for 1 US cup of all-purpose flour:

Spread: 120 to 155 g — 29% range. The same recipe interpreted with each reference produces meaningfully different doughs. King Arthur’s 120 g produces a softer, more hydrated dough; ATK’s 142 g produces a denser, stiffer one. Bakers who can’t reproduce a recipe are often switching reference systems without realising.

The cookie test

We computed the practical impact for a standard chocolate chip cookie recipe that calls for 2 cups of flour:

• King Arthur interpretation (240 g): soft, spread-thin cookies; gooey centre.
• USDA interpretation (250 g): standard cookie; slight spread.
• ATK interpretation (284 g): thick, cakey cookies; minimal spread.
• European interpretation (310 g): dense, dry, possibly crumbly.

+44 g of flour across the range matters; the recipe’s hydration shifts from ~60% to ~50%, which is the difference between a soft cookie and a bread-adjacent biscuit.

The same pattern in other ingredients

Flour isn’t unusual. Density-by-reference for other common baking ingredients (per US cup):

Cocoa powder is the worst offender — a 57% spread between the two reputable references. Light, fluffy cocoa packs very inconsistently. Sift cocoa before measuring if the recipe’s author doesn’t specify.

Liquids: a quieter story

Water-based liquids barely vary. 1 cup of water is 237 g across every source. Milk, cream, and yogurt are consistent to within 2-3%. The reason: liquids don’t compress in a measuring cup, so the volume reflects mass directly. The technique-dependence that plagues flour is a property of compressible solids.

Sticky liquids (honey, molasses, peanut butter) vary within 3-5% — the surface-tension residue that clings to the cup is the main source of disagreement.

Why the references disagree

Three reasons:

1. Method assumed.Spoon-and-level packs flour the least. Dip-and-sweep packs it the most. Sifted flour is sometimes weighed before sifting (more dense) and sometimes after (less dense). The reference’s assumed method matters as much as the ingredient.
2. Source flour properties. US flour brands vary in protein content (10-13%), particle size, and treatment (bleached, enriched). European flours are milled finer and tend to be denser per unit volume.
3. Measurement convention.King Arthur and ATK measure by their preferred method (KA spoons; ATK dips). USDA uses a reference standard that doesn’t fully match either. Different methods, different numbers, all correct for what they describe.

What this means for your recipe

Three practical rules:

1. Use the recipe author’s preferred number.A King Arthur recipe assumes King Arthur’s flour density. Pull the gram weight from the same source you got the recipe from.
2. Default to 120 g per cup of flourwhen source is unknown. It’s the spoon-and-level convention used by most modern recipe sites and is the most reproducible.
3. Use a scale.The variance in gram-weight conventions across sources is only relevant if you’re working from cup-based recipes. Modern bakers default to weight, eliminating the entire class of error.

Compute exact conversions in our cooking converter, which uses the King Arthur density values. For the structural background, see our cups to grams guide.

Methodology

For each ingredient we recorded the published gram weight per US-customary cup (236.59 mL) from six independent reference sources. Where a source provided weight in ounces or grams per 100 g, we converted to grams per US cup using the source-published density or, failing that, the USDA standard density.

• Reference set:six sources, all publicly published and in print as of 2024-2025: King Arthur Baking Ingredient Weight Chart, USDA FoodData Central, Bob’s Red Mill product pages, America’s Test Kitchen 2023 compendium, Joy of Cooking 75th anniversary edition, and a triangulated “European pâtisserie” reference (Lenôtre, Hermé, Ducasse training guides averaged).
• Ingredient set: 36 ingredients in our conversion library; 8 highlighted in the spread table for being most variable.
• Cup definition: US customary cup = 236.5882365 mL, per NIST Handbook 44 and the IEEE-754 double rounding our converter uses internally.
• Method-of-measurement annotation:each row records the source’s stated assumption (spoon-and-level, dip-and-sweep, sifted-then-measured, measured-then-sifted).
• Hydration calculationsin the “cookie test” use the standard baker’s percentage formula: hydration = liquid_g / flour_g × 100.

Key findings

• 29% gram-weight spread for AP flour across the six sources (120 g King Arthur, 155 g European pâtisserie).
• 57% spread for cocoa powder — the single most variable ingredient surveyed (75 g KA, 118 g ATK).
• Water and other thin liquids are consistent to within 2-3%across all sources — volume reflects mass directly when the medium doesn’t compress.
• +44 g (18%) flour delta on a 2-cup cookie recipe between the lowest (KA) and highest (European) interpretations — sufficient to shift hydration from ~60% to ~50% and meaningfully change texture.
• King Arthur converges with USDA most closely across the 8-ingredient table (median absolute deviation ~5%); ATK is the consistent outlier on the high side (~12% above the median).
• European references run 8-15% denser per cup on flours due to finer milling and the sifted-then-settled measurement convention.

Caveats / Sources of bias

• Six sources is not a statistical sample. The selection is curated for reputation, not representativeness; a broader internet survey would surface even more variance.
• Source method assumptions vary. Comparing a spoon-and-level source against a dip-and-sweep source conflates two real-world techniques; some of the reported spread is method, not source.
• Flour brand and freshness matter. Protein content (10-13%), moisture absorption (varies with storage humidity), and mill date all affect the density of a single brand by 2-5%.
• Single-cup measurement vs scale-of-recipe. A 2-cup measurement compounds the per-cup error; a ½-cup measurement amplifies it via measuring-cup tolerance.
• European pâtisserie reference is an average, not a single canonical source. Professional French references differ by 5-10% among themselves.
• Temperature dependence ignored. Cold butter, room-temperature shortening, and warm honey all have different cup-weights vs the stored room-temperature reference.

Sources

King Arthur Baking Company “Ingredient Weight Chart” (2025 edition); USDA FoodData Central nutrient database; Bob’s Red Mill product weight reference (per-product pages on bobsredmill.com); America’s Test Kitchen baking measurement notes (2023 compendium); Joy of Cooking (75th anniversary edition, 2019); European pâtisserie professional reference (Lenôtre, Hermé, Ducasse training guides averaged). NIST Handbook 44 (2023 edition) for the US customary cup definition.