In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
When lemon appears in a MOMS recipe at 2.5 grams, does the slot we file it under — "Functional" or "Flavor" — correspond to any measurable difference in the shake, or is it only a naming habit?
Hypothesis
We expected dose to be the dividing line. The working assumption inside the recipe database was that "Functional" marks an ingredient present at a quantity chosen for a physiological effect, and "Flavor" marks one present for taste. If that assumption held, the Functional lemons should sit at a higher gram weight than the Flavor lemons, and the split should be visible in the numbers alone, without any other information about the recipe. We expected a clean threshold somewhere around 4 to 5 grams.
Method
Two published dose thresholds framed the question. Cook and Monsen (American Journal of Clinical Nutrition, 1977) ran multiple radioiron absorption tests in 63 male subjects and found that the increase in non-heme iron absorption from a semisynthetic meal was directly proportional to the amount of ascorbic acid added across a range of 25 to 1,000 mg, with absorption ratios of 1.65 and 9.57 at those two extremes. 25 mg is the smallest dose in that range; below it, the study says nothing. Jiang and Fu (Food Chemistry, 1998) partially purified polyphenol oxidase from litchi peel and found that 100 mmol per litre citric acid, combined with 10 mmol per litre glutathione, gave 80 to 85 percent inhibition of PPO activity and good control of browning. Internally, we pulled all 201 recipes and extracted every ingredient line containing lemon, together with the slot it sits in and its gram weight. 32 recipes contain lemon; none contain it in both slots. We then tested three candidate rules against the split, and computed the macro profile of each of the two groups.
Results
The split is exactly even: 16 recipes file lemon under Functional, 16 under Flavor. The dose does not separate them. 25 of the 32 lemon entries are exactly 2.5 g. Functional doses run 2 g to 4 g, with 13 of 16 at 2.5 g. Flavor doses run 2 g to 15 g, with 12 of 16 at 2.5 g. The two distributions sit almost entirely on top of each other. The clearest case is a matched pair. MI-W2-IMMUNE-001-V1 "Vital Flow" and MI-W2-IMMUNE-003-V1 "Resilient Glow" share a body type (MI), a week (2), a goal (Immune & Circulation), a liquid (Rice Milk 150 ml) and a serving size (16 oz). Both contain lemon at 2.5 g. In Vital Flow it is the entire Flavor line, and the Functional line reads Chia Seeds 5 g. In Resilient Glow it sits in the Functional line beside Turmeric 0.5 g and Chia Seeds 8.5 g, while Agave 2.5 g takes the Flavor line. Same ingredient, same weight, same body type, same week, same goal — different column. A second pair repeats it. WAY-W2-COOL-003-V1 "Frosted Clarity" carries Lemon 2.5 g in Functional; WAY-W2-COOL-011-V1 "Frost & Flow" carries Lemon 2.5 g in Flavor. Both are body type WAY, week 2, goal Cool Down & Detox. Against the published thresholds, 2.5 g of lemon reaches neither one. At roughly 39 mg of vitamin C per 100 g of juice, 2.5 g carries about 1 mg — roughly 25 times below the 25 mg floor of the range Cook and Monsen tested. Reaching that floor from lemon alone would take about 65 g of juice, close to a lemon and a half. On the acid side, 2.5 g of lemon contributes roughly 0.12 g of citric acid; dissolved through a 16 oz (473 ml) shake that is about 1.3 mmol per litre, roughly 75 times below the 100 mmol per litre Jiang and Fu used to control browning.
What Failed
The hypothesis failed first, and then two replacement hypotheses failed after it. Dose does not predict the slot. Worse, the single largest lemon dose in the database runs the wrong way: WI-W1-VITAL-001-V1 "Radiant Recharge" carries 15 g of lemon — six times the modal dose — and it is filed under Flavor, with Turmeric 3.8 g and Cashew 9 g occupying the Functional line. We then tested a second rule: lemon moves to Functional when a sweetener already occupies the Flavor line. That predicts 17 of 32 correctly, which is a coin flip. We tested a third: lemon moves to Functional when the Functional line would otherwise be empty. That does better at 20 of 32, but three recipes break it outright — WY-W1-DETOX-014-V1, WY-W4-RENEWAL-003-V1 and WY-W4-RENEWAL-005-V1 all place lemon in Functional and leave the Flavor line completely empty. Goal does not separate them either. Three goals — Light & Vital Renewal, Immune & Circulation, and Cool Down & Detox — contain recipes on both sides of the split. One number is tempting and we are not going to use it. The Functional-lemon group averages a Score of 14.6 and the Flavor-lemon group 16.0, and the two groups also differ in calories (357 vs 377) and carbohydrate (29.2 g vs 36.5 g). None of that can be caused by the lemon: 2.5 g of lemon is under one calorie. Those gaps come from the other eight or nine ingredients in each shake, and reading them as an effect of the slot would be backwards. A limitation we cannot resolve from the sheet: the database records "Lemon 2.5 g" without specifying juice, zest or whole fruit. The vitamin C and citric acid arithmetic above assumes juice, which is the most generous of the three readings for the functional interpretation.
What We Learned
At 2.5 g, the label "Functional" is a claim we cannot support. Neither of the two documented mechanisms we could name for lemon — enhancing non-heme iron absorption, or suppressing enzymatic browning — comes within an order of magnitude of its studied dose at this quantity. What 2.5 g of lemon reliably does is taste like something. It is roughly half a teaspoon of juice, and acid at that level is clearly perceptible against 300 ml of soy milk and a date. So the honest reading is not that the split is arbitrary in both directions. It is that one of the two labels is defensible and the other is not. "Flavor" describes what we can measure. "Functional" describes what we hoped was happening. This is a naming problem rather than a formulation problem — no shake is worse for it, and nothing needs to be reformulated. But a column that means "we think this ingredient does something" and a column that means "this ingredient does something we can cite" are different columns, and ours had quietly become the first while being read as the second.
What Changed at MOMS
Lemon at or below 4 g is now filed under Flavor. The Functional column now requires a dose we can point to a published range for: if the gram weight sits below the lowest dose in the study we would cite, the ingredient belongs in Flavor, and the recipe note says so explicitly. We are applying the same test backward across the other small-dose entries in the database. Turmeric at 0.2 g and cinnamon at 0.2 g are the next two under review, and we expect both to move. Where we actually want lemon to do the iron-absorption work, the dose has to rise to roughly 65 g of juice, which changes the taste of the shake substantially. That is a formulation decision to make deliberately, in a small number of iron-focused recipes, rather than something to imply with a column heading in 16 of them.
References
Cook JD, Monsen ER. Vitamin C, the common cold, and iron absorption. The American Journal of Clinical Nutrition, 30(2):235-241, 1977. | Jiang Y, Fu J. Inhibition of polyphenol oxidase and the browning control of litchi fruit by glutathione and citric acid. Food Chemistry, 62(1):49-52, 1998.
Keep reading
Next experiment
Not sure which shake fits you? We match the formulation to your body type.
Find your body type