Caldo de Ervilha with Cheese Toast

What it is

Dried split peas, simmered with onion, carrot, garlic and smoked paprika until they give up and collapse into the pot. Half of it gets blitzed smooth and half stays chunky, then cheiro-verde (parsley and scallions) goes on top, with a slice of cheese toast on the side for dunking.

Panelinha files its split pea soup under winter, grandma's cooking and the classics, and its version is a pea and bacon soup made with chicken stock that even comes with popcorn. I'm lacto-vegetarian, so mine keeps neither the bacon nor the stock: smoked paprika brings the smoke, plain water does the rest, and the peas bring the body.

It's Protein Champion number five, and split peas are an easy one to love. No soaking (NDSU Extension says dried split peas don't need it), one pot, and about an hour and a half from chopping board to bowl.

Why you should eat it

Here's the short version, before anyone files pea soup under starters. USDA FoodData Central lists 8.34 g of protein and 8.3 g of fiber per 100 g of cooked split peas, almost gram for gram, with 118 kcal and 21.1 g of carbohydrate in that same 100 g. Each bowl here starts from about 74 g of dried peas, and by USDA's raw record that's about 17 g of protein and 16 g of fiber before the toast lands. The peas bring next to no salt (2 mg of sodium per 100 g cooked), so most of the salt on this plate comes from the shaker, the bread and the cheese, and most of the fat from the cheese and the olive oil. That's a lot of fiber for one sitting if pulses are new to you, so go easy at first (the tips say how). And pea protein has been tested in real people, gut and muscle both. Those stories wait below while the pot simmers.

Tips

• Stir more often near the end: as the peas collapse, the caldo thickens and loves to catch on the bottom of the pot. NDSU Extension suggests checking the water every 5 minutes or so near the end of cooking.
• New to pulses? This bowl is a lot of fiber: USDA lists 22.2 g of fiber per 100 g of dry split peas (172428), and if your gut has opinions about that, it's allowed to. In case three, 2 men left after their first test meal with gut discomfort (which meal isn't reported); in case four, gut feelings were the same across all four breakfasts. My suggestion, not the trials': start with half a bowl and build up.
• Season at the end and taste first: the bread and the cheese bring salt of their own. If you're managing your blood pressure, your doctor is the right person to set your salt.
• Make the pot on Sunday and your next three lunches are waiting in the fridge (3 to 4 days, USDA FSIS). Only the toast needs making fresh: three minutes under the grill.

Substitutions

• Dried green split peas: Dried yellow split peas Cook them the same way and let the squash test decide. The card is built on the green split pea record, so it won't match exactly.
• Sourdough or French-style bread: Wholemeal bread More fiber: USDA lists 6 g per 100 g of commercial whole-wheat bread (172688) against 2.2 g for French or sourdough (172675). USDA's whole-wheat bread records carry no amino acid values, so 'While it simmers' uses the white loaf, and the card won't match.
• Mozzarella: Any good melting cheese made with non-animal rennet (check the label), or leave it off The card uses USDA's part-skim mozzarella record, so another cheese shifts the numbers, and leaving it off takes protein, fat and salt with it.
• Stick blender: A jug blender, in batches Panelinha's advice for hot soup: blend in batches and hold the lid down with a kitchen towel against the steam. Same caldo, same card.

Storage

Caldo: into the fridge within 2 hours, divided into shallow containers so it cools fast. It keeps 3 to 4 days there or 3 to 4 months in the freezer. Reheat it to a rolling boil (USDA FSIS). It thickens a lot as it sits, so loosen it with a splash of hot water as it warms. The toast doesn't keep: make it fresh.

While it simmers: the pea lends, the bread pays back

Lid half on, peas doing their slow collapse. You've got forty minutes, so pull up a chair. Start with a word: 'pulse', the family name for beans, lentils and dried peas, comes from the Latin puls, a thick soup (NDSU Extension). The family is named after what's in your pot.

Protein is a chain of smaller pieces called amino acids. Nine of them your body can't make, so food brings them, and every food brings its own mix.

Here's the mix on this plate, from USDA FoodData Central. Lysine makes up about 7.2 g of every 100 g of protein in cooked split peas, and 2.5 g in toasted French bread or sourdough. The sulfur pair, methionine and cysteine, goes the other way: 3.3 g in the bread, 2.5 g in the peas. These are shares of the protein in the food, before digestion has its say.

Now a yardstick. The FAO, the UN's food and farming agency, says a gram of protein for older children, teens and adults should carry 48 mg of lysine (4.8 g per 100 g of protein) and 23 mg of the sulfur pair (2.3 g). The bread reaches about half the lysine line, and the peas clear it by half again. Tonight, the peas are the lender.

Teacher hat on. Two kids at the art table, same colours in both boxes: one has a fistful of blue crayons and a stub of red, the other plenty of red and a stub of blue. Nobody's missing a colour. The peas are the first kid, the bread the second. Push the boxes together and the whole table can draw.

The picture stops in two places. First, the size of the boxes. The FAO's own worked example mixes 400 g of wheat with 100 g of peas and a little milk powder, and still comes out short on lysine, because the wheat outweighs the peas. This bowl leans the other way: by my arithmetic the peas bring about half its protein and the bread a fifth, and together they land at about 5.9 g of lysine per 100 g of protein, comfortably over the line.

Second, the peas' stub of red. In case one, purified pea protein measured in people landed right on the FAO line for the sulfur pair: enough, with nothing to spare. Whole peas may digest less well than a powder, so the bread's bit extra of the sulfur pair is handy insurance.

So, in one line: the peas lend the bread its lysine, the bread tops up the peas' sulfur pair, and it works because the peas get the bigger share of the bowl.

Case one: how much pea protein your gut really takes in

Healthy adults in Paris, average age 36: 26 started and 15 finished with usable samples, 8 on pea and 7 on casein, milk's main protein. Each had a thin tube passed through the nose to the end of the small intestine, where digestion wraps up, so the team could collect what was left of the meal.

Over 4 hours they ate 9 small portions of mashed potato carrying 35 g of pea protein isolate or 31 g of casein isolate, matched for nitrogen. The peas had been grown with a labelled fertiliser, so their amino acids could be told apart from the body's own.

The gut took in 93.6% of the pea's amino acids and 96.8% of the casein's, a gap too small for the statistics to call, though a few, lysine among them, came in lower for pea. The share of the protein the body held on to, rather than losing or breaking it down, was the same: 71.6% against 71.2%. On DIAAS, a score that checks each essential amino acid after digestion against the FAO's reference, pea scored 1.00, enough for adults, with the sulfur pair as its lowest point. Casein scored 1.45.

For your plate: the protein inside a pea gets taken in well, and its one thin spot is the sulfur pair, the bread's job here.

Where it stops: this was a purified powder, not a bowl of split peas, and the authors note that plant proteins tend to digest better purified than whole. The FAO's worked example credits whole peas with digesting 69% of their sulfur pair, a figure measured in pigs. The mini-meal pattern may have overestimated digestibility, the authors add, and 15 people is a small group. Roquette funded it, and three of the authors work there. Careful, small and company-funded.

Case two: pea protein against milk protein, at the muscle

Twenty-four healthy young men in Maastricht, average age 24, active but not on a structured training plan. Twelve per group, split at random, and neither the men nor the researchers knew who drank what.

After an overnight fast, at rest, each drank 30 g of pea protein concentrate or milk protein concentrate. A labelled amino acid ran into a vein, and small samples of thigh muscle over 5 hours showed how fast new muscle protein was being built.

The answer: the same, to the third decimal, 0.053% per hour for both (P = 0.96). The blood differed on the way there: essential amino acids rose about 48% more after milk, and leucine peaked about 25% higher. After the pea drink, methionine rose for 90 minutes, then dipped below where it started; the pea drink carried 0.2 g of it, the milk drink 0.7 g.

For your plate: the protein pulled out of peas built muscle protein as fast as milk protein did, even with a smaller wave in the blood. And methionine, one half of the sulfur pair, was the thin spot again.

Where it stops: these were powders, and the authors write that whole plant foods carry fiber and other compounds that can lower digestibility. Young men only, one dose, at rest rather than after a workout. A bowl here carries about 17 g of protein from the peas (my arithmetic, USDA), plus the bread and the cheese: a meal, not a shake. The industry partners (Tereos, Cargill and Kellogg) helped design the study and made the drinks, though the authors state they had no role in the data or the write-up, and two authors have had grants or fees from FrieslandCampina, Tereos Syral and PepsiCo. A clean result, in a narrow test.

Case three: split pea mash against a meat breakfast

Forty-three healthy young men in Copenhagen finished this one, average age 24. On three mornings, at least two weeks apart, each ate a breakfast of the same energy: veal and pork patties with mashed potato; fava bean patties with a mash of 90 g of dried split peas, matching the meat's share of protein; or a lighter bean, pea and potato mix with about half the protein. They rated their hunger and fullness for 3 hours, then sat down to all the pasta they wanted.

The bean and pea breakfast won the round: less hunger and more fullness than after either of the others, and 12% less eaten at lunch than after the meat. Most of that held after the team allowed for how much the men liked each meal, and they did not like the legume one much: 29 out of 100 for palatability, against 54 for the meat. I'd like to think a caldo with cheese toast would score higher. My bias, not data.

For your plate: a real split pea mash kept these men fuller than meat did. Your bowl starts from about 74 g of dried split peas, a little less than their 90 g, and without the fava beans.

Where it stops: the legume breakfast carried about four times the fiber of the meat one, so the authors say this doesn't show that plant protein itself is more filling. One breakfast, young healthy men only, and 2 men left after their first test meal with gut discomfort (the paper doesn't say which meal). The Nordea Foundation and the Danish Agriculture & Food Council paid for it, and one author has been a consultant or adviser to Global Dairy Platform, McCain, McDonald's and Weight Watchers. Which brings us to the rematch.

Case four: the rematch, with the fiber evened out

Same Copenhagen group, the following year: 33 healthy men aged 18 to 50 (average 26.5), vegetarians left out. Four breakfasts, at least a week apart, all with the same energy and share of protein. One was case three's fava bean patties and split pea mash, with 28.4 g of fiber. Two were meat or egg patties with a fiber made from pea shells stirred in to match (28.8 g and 28.5 g). The fourth was the egg meal with nothing added, at 6.0 g. Then the same 3 hours of ratings and the all-you-want pasta lunch.

This time, nothing separated them. Hunger, fullness and appetite showed no difference, and neither did lunch or the rest of the day's eating. The legume meal still lost on taste (39.7 against 61.3 for the meat), and gut feelings were the same across all four.

For your plate: once the fiber was evened out, case three's edge disappeared, and the authors name the fiber matching as the probable reason. In a split pea bowl, that fiber comes in the same spoon as the protein.

Where it stops: one honest wrinkle. The egg breakfast with no added fiber didn't differ either, so this trial can't cleanly prove fiber was the whole story. That reading is the authors', and I find it fair, but it's a hint. Men only, in a lab, eating test meals the authors admit weren't a typical Danish breakfast. The Danish Agriculture and Food Council paid for it, DLG Food donated the pea fiber, and the same author lists advisory work for food and drink organisations and Weight Watchers, plus research grants that include dairy funders.