Foundations · Article 27

Weights & Measures

A reference article · 6100 words · 49 cited sources · last reviewed September 2026

Ratios & Formula Frameworks argues that volume cannot be scaled or diagnosed, and Tools & Equipment argues that the scale is the highest-value tool in the kitchen. Neither gives a table. This is the table, and the apparatus needed to read it: which cup, which fill, which chart, and where the sources disagree enough that one number would be a guess dressed as a fact.

A measuring cup does not measure flour. It measures the space flour is occupying at the moment you level it off, and how much flour is in that space depends on what you just did to it. This is not a small effect hiding at the edge of precision: the same cup, the same flour, the same baker can differ by a third depending on whether the cup was filled with a spoon or plunged into the bin. Everything difficult about converting a volume recipe into a weight recipe follows from that, and so does the answer, which is to say out loud which filling method a number assumes instead of quietly picking one.

Which cup

The US customary cup is one-sixteenth of a US liquid gallon. The gallon is defined as exactly 231 cubic inches, so the cup is 14.4375 cubic inches, or exactly 236.5882365 mL. Its tablespoon is a sixteenth of that (0.5 fl oz, 14.79 mL) and its teaspoon a third of a tablespoon (4.93 mL). This is the cup behind USDA’s household measures: the FoodData Central record for bottled water gives 1 cup as 237 g, which is 236.6 mL of room-temperature water rounded to the gram (U.S. Department of Agriculture, Agricultural Research Service, n.d.).

The nutrition-label cup is a rounded version of the same unit. For a Nutrition Facts panel, FDA guidance states that “1 cup means 240 mL, 1 tablespoon means 15 mL, 1 teaspoon means 5 mL, 1 fluid ounce means 30 mL” (U.S. Food and Drug Administration, 2019). That is 1.4 percent larger than the cup in the drawer, which matters only when a serving weight is lifted off a package and used as an ingredient weight.

The metric cup is 250 mL, the standard in Australian, New Zealand and metric Canadian recipe writing, 5.7 percent larger again. Metric recipes also use a 20 mL tablespoon rather than the American 14.79 mL one, so a tablespoon crossing that border changes by 35 percent, far more than the cup does. A recipe that says “2 tablespoons of cocoa” means 11 g in Sydney and 8 g in Seattle.

Why a cup of flour is not a fixed weight

Bulk density is the bridge between volume and mass, and for a fine powder it is not a constant of the material. The pharmaceutical literature states the problem more bluntly than the baking literature does. Measuring bulk and tapped density of common excipients across laboratories to one standard method, Akseli and colleagues open with the point: “the bulk properties of a powder are dependent on the preparation, treatment, and storage of the sample, that is, how it was handled… the slightest disturbance of the powder bed may result in a changed bulk density. Thus, the bulk density of a powder is often difficult to measure with good reproducibility and, in reporting the results, it is essential to specify how the determination was made” (Akseli et al., 2019). That last clause is the design rule for every table below.

Two standard numbers size the effect. Loose or aerated bulk density is the density of powder settled under gravity without disturbance; tapped bulk density is the same powder after a standard number of consolidating taps. The Hausner ratio is tapped divided by loose, and the Carr compressibility index is the same information as a percentage. Both are routine flowability indicators, and both measure directly how far a cup weight can travel (Abdullah & Geldart, 1999; Santomaso, Lazzaro, & Canu, 2003).

For wheat flour the numbers are large. Zhao and colleagues measured Carr indices of 25.77 for a hard red spring flour and 46.60 for a soft red winter flour, the hard-wheat figure rising to 38.48 after 24 hours under 20 kPa of consolidation (Zhao, Phalswal, Shetty, & Ambrose, 2021). A Carr index of 46.6 percent is a Hausner ratio of 1.87: the consolidated powder is nearly twice as dense as the aerated powder. Measured directly on another cereal flour, hammer-milled teff gave an aerated bulk density of 548 kg/m³ against a tapped density of 804 kg/m³, and pin-milled teff reached a Hausner ratio of 1.62 with an angle of repose of 71.6° (Barretto, Buenavista, Pandiselvam, & Siliveru, 2022). Ranked against maize semolina, sugar and skim-milk powder, all of which are classed “easy flowing,” flour is classed “cohesive” and cocoa “extremely cohesive” (Benković, Srečec, Špoljarić, Mršić, & Bauman, 2013). Flour’s static angle of repose sits above 40°, the highest in one comparative test set, driven by cohesion rather than particle shape (Macho et al., 2020).

Four variables move the number, and a domestic kitchen has all four.

Handling is the largest, and the only one the baker controls. Plunging a cup through a bin applies the consolidating stress that the tapped-density test applies on purpose. Guan and Zhang measured the consequence in a model flour bin: cohesion rose 72 percent as moisture went from 8.6 to 14.2 percent, and compaction raised the hopper opening needed for arch-free flow by about half (Guan & Zhang, 2009). Shen and colleagues, compressing edible powders under low pressure, found compressed volume rising logarithmically with applied pressure, and wetter powder compressing more (Shen, Li, & Xu, 2022). The mechanism that makes an industrial hopper arch is the one that makes a scooped cup heavy.

Moisture. Flour equilibrates with the air in the room. Rivera and colleagues exposed sieved hard and soft winter wheat fractions to 50, 60 and 70 percent relative humidity and found flowability degrading as humidity rose, with Hausner ratio and compressibility index both climbing (Rivera, Zhao, Owonikoko, & Siliveru, 2023). Hasmadi found tapped density itself barely moving across added-water levels of 12.5 to 30 percent while caking strength rose sharply (Hasmadi, 2021). Chang and colleagues, fitting model starch-and-protein powders against water activity, found loose bulk density falling as water activity rose while Hausner ratio, angle of repose and cohesion all rose (Chang, Kim, Kim, & Jung, 1998). This is why the same cup of the same flour weighs differently in a humid August and a dry January.

Particle size and wheat class. Siliveru and colleagues found flour shifting from “easy flowing” to “very cohesive” at 12 percent moisture and 45 µm particle size, with bulk and tapped density correlating strongly with damaged starch, protein and fat content (Siliveru, Ambrose, & Vadlani, 2017). Bian and colleagues, comparing hard red winter with soft white winter flour at nearly identical mean particle size, found the hard flour’s bulk and tapped density significantly higher (Bian, Sittipod, Garg, & Ambrose, 2015). Milling stream matters too: flours from the fifteen streams of a pilot roller mill differed enough that the later reduction streams behaved like gluten-free flours (Nkurikiye et al., 2024). Cake, bread and pastry flour cannot honestly share one cup weight.

Storage under its own weight. Apparent density of salt, sugar, flour, starch and protein powders follows a power function of applied stress across the storage range (Lanzerstorfer, 2020). The bottom of a large bag is denser than the top.

None of this literature reports cup weights. It reports why they move, and it sets the envelope: a factor approaching two between the loosest and most consolidated state of one flour. The kitchen figures sit comfortably inside it.

The three fills, quantified

Sifted means passed through a sieve and spooned in without settling, and gives the lowest weight. Spooned and levelled, which King Arthur calls fluff, sprinkle and scrape, means loosening the flour, spooning it in until it overflows, and levelling with a straight edge without tapping or pressing (King Arthur Baking, 2023). Bob’s Red Mill gives the same instruction and warns against the alternative: do not “pack the flour into the cup or scoop directly from the bag” (Bob’s Red Mill, n.d.). Dip and sweep, the America’s Test Kitchen standard, means plunging the cup in and levelling the heaped result: “Dip the measuring cup into the container and scoop up the ingredients in a heaping mound. Use a straight edge to sweep off the excess” (America’s Test Kitchen, n.d.-a). Packed is what happens when a cup is pressed, tapped, or filled from the settled bottom of a bag: intended for brown sugar, an error for flour, and the top of the range.

Three independent kitchens converge on about 120 g for a light spoon-and-level and about 140 g for dip-and-sweep, with packed figures clustering at 155 to 160 g. King Arthur’s own recommended and condensed figures are 33 percent apart, and that 120-to-160 g band is the honest range for all-purpose flour specifically. Counting Gisslen’s sifted bread flour at the bottom, wheat flour as a class spans 113 to 160 g, 42 percent.

The test kitchens say the same thing from the other side. America’s Test Kitchen reports that a 1-cup measure “can vary by as much as 20 percent, depending upon the light- or heavy-handedness of the baker,” and that their 5 oz standard “was arrived at by having dozens of volunteers measure out 1 cup of flour and weigh it and then taking the average” (America’s Test Kitchen, n.d.-c). Gisslen sets it as a student exercise: sift and lightly spoon a cup, then scoop and lightly pack one, weigh both, “note the difference. No wonder home recipes can be so inconsistent!” (Gisslen, 2004). Figoni gives the mechanism: “When flour settles, there is less air between particles. With less air, density is greater and more flour is needed to fill a container… if flour is sifted before it is measured, there is more air between particles. Density is lower, and less flour is needed to fill a cup” (Figoni, 2011).

The tables

Every weight in grams. The metric column is the US figure multiplied by 250 ÷ 236.588, which is arithmetic rather than measurement. “Method” is the fill the source assumes; not stated means the source publishes none, and those figures should be read as a middling fill.

Flours, starches and cocoa

Bread flour is the largest unexplained gap. King Arthur gives it the same 120 g as all-purpose; USDA gives 137 g against 125 g. The two charts disagree not just on the figure but on whether bread flour is denser than all-purpose at all. The peer-reviewed work says hard-wheat flour does pack denser than soft at equal particle size (Bian et al., 2015), which supports USDA’s direction, and King Arthur’s equal figure is more likely a deliberate simplification for recipe writing. Cake flour carries the widest legitimate method spread, 100 g sifted to 128 g unsifted, on the one ingredient whose entire job is a tender crumb. Cornstarch shows the method effect most cleanly: Gisslen’s sifted and unsifted pair, 113 and 128 g, brackets every other source, and USDA’s method-unstated figure lands exactly on the unsifted one.

Sugars and syrups

Granulated sugar is the best-behaved ingredient in the library: four independent sources inside 1 percent, because a free-flowing crystalline solid barely responds to handling. The lone outlier is USDA’s newer Foundation record, which gives 188 g and sits 6 percent below USDA’s own older figure. Brown sugar is the reverse case, spanning 198 to 220 g packed because “packed” is not a defined pressure; USDA’s loose-to-packed pair is 145 against 220 g, and America’s Test Kitchen puts the same difference at “about 2 ounces… almost 30 percent of the weight of a cup” (America’s Test Kitchen, n.d.-d). Confectioners’ sugar is the only sugar where sifting is routinely specified, and the sources disagree about the size of the effect: Gisslen’s sifted figure and King Arthur’s unsifted figure are the same 113 g, which is as much as the published data supports.

Fats

Butter’s cup weight is a convention, and an honest one. A US stick is a quarter-pound, 113.4 g, and its wrapper carries eight tablespoon marks and a half-cup mark; every chart’s 227 g cup is two sticks and every 14.2 g tablespoon is a stick divided by eight. Cut at the printed lines and it is exact. It is not a density, so it does not convert a volume of melted butter, and no published density curve for butter exists to do that with.

The lard figures disagree by 10 percent for a related reason. USDA’s 205 g is 0.867 g/mL, plausible for rendered pork fat; King Arthur’s 226 g is the butter convention applied to a different fat. Shortening’s single published figure, 184 g, is 0.778 g/mL, low for a triglyceride and consistent with commercial shortening being aerated with nitrogen during manufacture.

Dairy and other liquids

Every figure here is a real density multiplied by 236.6 mL, and the ordering makes physical sense. Buttermilk and yogurt carry the most non-fat solids and are densest; heavy cream at roughly 36 percent fat is barely above water, because butterfat is lighter than the serum it displaces; sour cream sits below water. Watson and Tittsler found milk density more closely tied to non-fat solids than to fat percentage, which is exactly this pattern (Watson & Tittsler, 1961), and their cream equation, density = 1.03123 − (0.000770 × milk-fat percent) at 9.85 °C, returns 237 g per cup for 36 percent cream against USDA’s 238 (Watson & Tittsler, 1962). The whipped-cream row is the one pastry cooks need: whipping roughly halves the density, so a cup of whipped cream is about 120 g, a figure no conversion of liquid cream will produce.

Eggs

Eggs are the one baking ingredient whose recipe unit is a count, which makes the size standard the conversion table. The Agricultural Marketing Service grades US shell eggs by minimum net weight per dozen rather than by individual weight, so a carton marked Large guarantees a dozen weighing at least 24 oz and individual eggs vary around that floor (U.S. Department of Agriculture, Agricultural Marketing Service, 2000).

The shelled column is the in-shell minimum less 10 percent for shell; the Food Safety and Inspection Service gives a shell egg as “approximately 10% shell, 58% white, and 32% yolk” (U.S. Department of Agriculture, Food Safety and Inspection Service, 2021). That arithmetic reconciles the two USDA scales that otherwise look contradictory: 56.7 g in shell less 10 percent is 51 g, and USDA’s nutrient record for a raw large egg gives 50 g, because the nutrient database weighs the edible portion. The same split predicts 32.9 g of white and 18.1 g of yolk against the database’s 33 and 17 g.

European eggs are graded by individual weight in shell: XL at 73 g and over, L from 63 to under 73 g, M from 53 to under 63 g, S under 53 g (European Commission, 2023). A US Large egg, at its 56.7 g minimum, is therefore a European Medium. A European L egg carries roughly 10 to 25 percent more egg than a US Large, which is enough to move a custard or a génoise, and is a standing reason to convert continental recipes by weight rather than by count.

Leaveners and salt

Spoon measures govern here, and for salt the brand governs the spoon.

Table salt is the best-attested figure in this article: three independent sources, one of them a manufacturer’s nutrition panel, give 6.0 g per teaspoon. The two kosher salts are baking’s clearest case of a brand changing a measurement, and the cause is crystal geometry. Cargill states that Diamond Crystal’s Alberger pan-evaporation process gives “a unique crystal shape” with “rapid solubility, strong adherence, and lower bulk density” (Cargill, n.d.); America’s Test Kitchen describes the contrast directly, Morton’s crystals “made by crushing the granules between high-pressurized rollers” against Diamond Crystal’s “smaller, pyramid-shaped flakes” from pan evaporation (America’s Test Kitchen, n.d.-e). Hollow pyramids trap air; crushed flakes pack.

The soda and powder figures are reconciled in Chemical Leavening and appear here only for completeness. That article adopts 4 g per teaspoon for baking powder and 4.6 g for baking soda, noting that King Arthur rounds half a teaspoon of soda to 3 g and so implies 6 g per teaspoon, 30 percent above the USDA figure, on an ingredient with a narrow useful dose.

Nuts, chocolate, oats and dried fruit

Nuts and dried fruit agree between sources within 4 percent, because large irregular pieces pack by geometry rather than cohesion and a hand cannot compress them much. Two rows misbehave. The chocolate-chip sources disagree on direction, USDA giving large chips 182 g and minis 173 g while King Arthur gives standard chips 170 g and minis 177 g; packing theory favours King Arthur, since smaller uniform particles fill interstitial voids better, but the difference is inside brand-to-brand variation in chip geometry. Rolled oats carry the widest brand spread in the article, 81 g for USDA’s generic record, 89 g for King Arthur’s generic row and 113 g for King Arthur’s own branded oats, all reading “1 cup rolled oats,” because flake thickness is a rolling-mill setting. Here the package weight beats any chart.

Yeast

One standard US envelope is 7 g, a quarter-ounce, and holds about 2¼ teaspoons, which puts both active dry and instant yeast near 3.1 g per teaspoon (Red Star Yeast, n.d.-a). Red Star treats the two as interchangeable by volume, and puts 0.6 oz — about 17 g — of fresh compressed yeast as the equivalent of one dry envelope (Red Star Yeast, n.d.-b; Fleischmann’s Yeast, n.d.). Lesaffre’s technical sheet works in weight instead: replace active dry with three-quarters the amount of instant, and replace compressed yeast at 33 to 40 percent of its weight (Lesaffre, n.d.). Those two agree, and both agree with the instant 1 : active dry 1.25 : fresh 3 ratio that Ratios & Formula Frameworks already carries.

The size of the error

A 66 percent hydration bread formula written as “5 cups flour, 2⅔ cups water” is, at 120 g and 237 g per cup, 600 g of flour to 632 g of water: 105 percent hydration, which is not bread. Even read correctly it swings. Five cups of flour is 600 g spooned, 710 g dipped and swept and 800 g packed, so one written recipe runs from about 79 percent hydration down to 59 percent across three bakers with three habits and nothing on the page changed. That is the range from a slack ciabatta to a stiff bagel dough.

No peer-reviewed study has measured cup-to-cup variance using flour as the test material. The closest evidence is other dry powders measured the same way, and it points consistently. Pooled across twelve studies, people measuring dry kibbled food by cup ranged from an 18 percent under-estimate to an 80 percent over-estimate, with between-subject coefficients of variation up to 28 percent (German et al., 2011); a later study of the same task found individual accuracy from −47.8 to +152.2 percent, worse at smaller volumes (Coe, Rankovic, Edwards, & Parr, 2019). Scooping infant-formula powder carries roughly twice the error of measuring water, 9.0 against 4.4 percent mean absolute error (Rosenkranz et al., 2024), and across 42 pre-weighed foods only 29 percent of household cup and spoon estimates landed within 25 percent of true weight (Gibson et al., 2016). Weighing is measurably more precise and more accurate than volume for the same preparation (Boothby, Parr, & Bartges, 2022). Larger error for powders than liquids, larger error at small volumes, and wide variation between people: exactly what the powder physics predicts.

What is missing on purpose

Some figures are absent because nobody has published them, and inventing them would be worse than leaving a gap.

There is no primary cup weight for caster or superfine sugar. It is granulated sugar milled finer and packs somewhat denser, but no measured figure appears in USDA, King Arthur, America’s Test Kitchen or Gisslen. No source anywhere gives cocoa both sifted and unsifted; two research passes across Bob’s Red Mill, Valrhona, Guittard, Callebaut, Cacao Barry, Hershey’s, Ghirardelli and BAKERpedia found none, and none publishes a natural-versus-Dutch-process density difference either. Since cocoa is the most cohesive powder in the cupboard, the sifting effect is probably larger than flour’s, which makes this a real hole in any chocolate-cake conversion. There is no density curve for melted butter: searches returned dilatometric and solid-fat-content work (De Man & Wood, 1959) and texture studies, but no density against temperature, so “one cup of melted butter” has no sourced answer. Cargill publishes no gram figure for Diamond Crystal, so that row rests on two secondary sources. And confectioner’s glucose has no kitchen figure at all; its density depends on dextrose equivalent and solids, and the relationship is published (Wartman, Spawn, & Eliason, 1984) without a cup weight ever being derived from it.

Five USDA records — confectioners’ sugar, baking powder, shortening, corn syrup and cake flour — were read from FoodData Central mirrors rather than the database itself, because the public demonstration key rate-limited before they were reached. They are marked as such in the source corpus. Gisslen’s figures were verified in the fourth edition only. Two of Beranbaum’s figures reach this article at second hand through a publication quoting her.