Spinach Tops the Iron List. It Is Also 435g Per 100 Calories

Raw spinach leads all 182 foods we can rank on iron per 100 calories, at 11.78mg. It earns that by carrying 23 calories in 100g — so the hundred calories behind the headline weigh 435g. Most of a kilo bag, uncooked, in one sitting. Both halves of it fall out of one USDA FoodData Central row: 2.71mg of iron and 23 calories per 100g.

We published the iron-per-calorie ranking eight days ago and it was arithmetically correct. It was also missing the column that decides whether any of it is usable. A per-calorie leaderboard silently rewards water, and water is heavy. Here is the same ranking with the mass printed beside it.

The top ten on iron per 100 calories, with what 100 calories weighs

Seven of those ten need more than 150g to reach 100 calories, and four need more than 400g. The median across all 182 foods is 84g. Raw spinach is 5.2 times the median. Romaine lettuce, eighth on the ranking, is 588g per 100 calories — the leaderboard is quietly listing the wateriest things in the dataset and calling them iron-dense.

Two rows in that table are not like the others. Cooked breaded clams reach 100 calories in 50g and unsweetened cocoa powder in 44g, while sitting fifth and sixth on iron per calorie. They are the only entries in the top ten that a person could plausibly eat in one go, and the ranking buries them below four leafy greens.

Why the per-100g ranking is not the fix

The obvious correction — rank by iron per 100g instead — puts roasted whole sesame seeds first at 14.8mg. Sesame seeds are 565 calories per 100g, so they are 17th on the per-calorie ranking at 2.62mg. Swapping the denominator does not remove the distortion; it flips which direction the distortion runs. Per 100g flatters anything dry and dense. Per 100 calories flatters anything wet and dilute.

Raw spinach is the same food seen from both ends: 1st of 182 on iron per 100 calories, 36th on iron per 100g. Nothing changed but the divisor. If a ranking can move a food 35 places by choosing a denominator, the denominator is the finding, not the food.

Fix the target instead: what 5mg of iron actually costs

Hold the iron constant and let both the grams and the calories vary. That is the only version of this question with a usable answer, because it is the shape of the question people actually have. Five milligrams is an arbitrary target chosen so the arithmetic is readable — our rows carry food composition, not daily requirements, and we are not going to imply one.

The cheapest 5mg by calories is raw spinach at 42 calories, and it costs you 185g. The cheapest 5mg by mass is roasted sesame seeds at 34g, and it costs you 191 calories — 4.5 times as many. Neither of those foods wins both columns, and across the whole 182 nothing does.

What comes closest are the two the per-calorie ranking buried. Breaded cooked clams supply 5mg in 36g and 73 calories; unsweetened cocoa powder in 36g and 82 calories. Roughly two tablespoons of either, on both counts. That is a different answer from "spinach" and it only appears once you print both columns at once.

What this data cannot tell you

Absorption is not in these rows, and it is a large part of the real answer for iron specifically. The iron in the shellfish above and the iron in the spinach are not the same chemical form, and the body does not take them up at the same rate. We hold a milligrams-per-100g column, and that column is blind to it. Anyone comparing a plant source against an animal one on this table alone is comparing the wrong quantity.

Cooking is missing too, in both directions. Every leafy row here is raw, and raw is the state nobody eats 435g of spinach in. Wilted spinach is largely the same iron in a fraction of the volume, but our dataset has no cooked-spinach row to give you the number, so we are not estimating one. The clam row runs the other way: it is "cooked, breaded and fried", and the breading and the frying oil are inside its 202 calories.

Seven of our 189 rows are excluded from everything on this page: one whose published energy value we have flagged internally as false, five whose carbohydrate column is below zero and therefore broken, and one that simply has no iron figure. Hence 182. Not one of the seven is quoted, ranked or divided by anywhere above — we would rather hand you a shorter list than a contaminated one.

The general version

Any "most X per calorie" list is also a list of the foods with the fewest calories, and for water-heavy foods those are the same list. You can spot it without doing arithmetic: if the top of a per-calorie ranking is four salad leaves, the ranking is measuring dilution.

The repair is one extra column. Print the grams beside the ratio and the list re-sorts itself into things a person can eat. Print the calories to hit a fixed target and it re-sorts again, into the answer you were looking for in the first place.

Which is a fine thing to know once. Doing it on a Tuesday evening, for the actual bowl in front of you, is where it falls apart — and that is the part that decides whether any of this reaches your week. Sunn takes that off you inside WhatsApp: photograph the bowl, or write "big spinach salad and two eggs", and the numbers land back in the thread. Nothing to install, nothing to look up.

Elsewhere on the site: the original iron-per-calorie ranking this post corrects, every food we hold sorted by iron, the full breakdowns for raw spinach and unsweetened cocoa powder, and arugula against spinach — two of the four leafy greens in that top ten, side by side.