Discovering lesser-known local legumes


Switzerland grew 4769 hectares of grain legumes in 2024, against 78,634 hectares of bread cereals. Roughly a third of that area goes to human food, the rest to animal feed. Put differently, every lentil, chickpea, lupin, faba bean and protein pea destined for Swiss plates fits into about 1600 hectares.

That figure sets the scale of the subject. Old varieties are not coming back because a trend brought them back; they are coming back from an extremely low base, in a market with no tariff protection, where imports for human food rose from roughly 6000 tonnes in 2013 to roughly 9000 tonnes in 2023.

Two of these varieties also call for specific precautions that lists of heritage seeds rarely mention: lupin is one of the fourteen allergens Swiss law requires to be declared, and grass pea contains a neurotoxic amino acid where the dose decides everything. Both points come early on this page because they determine what ends up on the plate.

What actually defines a legume

Legumes belong to the Fabaceae family and form root nodules with soil bacteria of the Rhizobium genus, which fix atmospheric nitrogen. That symbiosis, rather than any nutritional property, is what defines the group agronomically: the plant makes its own nitrogen instead of drawing it from a fertiliser.

This explains the two solid agronomic arguments in their favour. They need no nitrogen fertilisation, and the fixed nitrogen remains partly available to the following crop, allowing its fertilisation to be reduced. Most tolerate drought and develop a root system that loosens the soil.

The downside is equally documented: weed control remains a substantial technical obstacle in legume cultivation, and intercropping with a cereal, which partly reduces it and limits lodging, complicates harvesting. The limiting factor in Switzerland is neither agronomic nor climatic. According to the Fribourg agricultural advisory service Grangeneuve, it is the market: buyers are scarce and there is no tariff protection against imported legumes.

There is also a varietal point that rarely gets mentioned. The available seed varieties were bred for cultivation in other countries and are not adapted to the Swiss climate. The Swiss Farmers’ Union notes that the harsher climate makes yields fluctuate heavily and renders production risky for farmers. That explains why many only commit small areas despite the agronomic advantages: according to the same source, roughly one farm in seventeen grows legumes, on an average of 3 hectares.

The local varieties, and what to know before cooking them

The borlotti bean

This red and white mottled bean is common in Italy and Ticino. Its creamy texture holds up to long braises, which makes it a base for soups and stews. Like every Phaseolus vulgaris, it needs full cooking, for the reason set out below.

Sweet lupin: a legume to treat as an allergen

Sweet lupin is the trickiest case on this list, and that has to be said before praising its protein content.

Lupins and lupin products appear at point 13 of Annex 6 of the Swiss Federal Department of Home Affairs ordinance on food information, the Swiss list of fourteen ingredients subject to mandatory allergen declaration. That declaration must stand out in the ingredient list through typeface, style, background colour or another appropriate means.

The reason is documented. A study in the Annals of Allergy, Asthma and Immunology assessed 43 people with peanut or lupin allergy through skin prick tests, specific IgE and oral food challenges: 44 per cent of peanut-allergic subjects had lupin allergy confirmed by challenge. Anaphylaxis was the most frequent manifestation after lupin consumption, with a minimal eliciting dose of 1 gram of lupin flour. The authors note that neither specific IgE levels nor wheal size distinguished mere sensitisation from confirmed allergy, so diagnosis may require an oral challenge. The French allergy vigilance network also published an alert in Allergy in 2024 on lupin as a hidden food anaphylaxis allergen.

In practice: for someone with a peanut allergy, lupin flour is not a harmless alternative to try out of curiosity, and the question belongs with their allergist. For everyone else, it is an ingredient whose presence needs reading on the label, including in breads and pasta where it is not expected.

A second distinction matters. Bitter lupins contain quinolizidine alkaloids, whose content varies by variety and growing conditions. Only so-called sweet varieties, selected for low content, are intended for consumption; traditional bitter seeds require prolonged debittering. Seeds sold for food in Switzerland are sweet varieties.

Grass pea: a question of dose

Grass pea, Lathyrus sativus, appears on lists of forgotten seeds for its resilience on poor soils. It also contains a non-protein amino acid, beta-ODAP, responsible for neurolathyrism, an upper motor neuron disorder presenting as spastic paraparesis.

What the data show is that the risk depends entirely on the overall diet. A community study in BMC Neurology, covering 631 households and 3350 people in a growing area of north-eastern Ethiopia, measured a prevalence of 2.4 per cent in the population and 9.2 per cent of households. The associated factors are telling: having eaten grass pea alone, having consumed immature seeds, and a grass pea to cereal ratio above 3 to 1.

By contrast, a cross-sectional study across three districts of Chhattisgarh, India, covering 1500 households and dietary recall in 5769 homes, found no adverse effects among daily consumers eating an average of 64 grams a day of split grass pea alongside millets, cereals and vegetables, at a mean beta-ODAP content of 0.630 grams per 100 grams.

European contents are not comparable either. Liquid chromatography analysis of three Italian cultivars harvested in Umbria and Puglia, published in 2026 in Food Science and Nutrition, measured between 0.01 and 0.037 milligrams of beta-ODAP per gram of powder, between 170 and 630 times less than the average content in the Indian study.

What that means for a Swiss reader: an occasional grass pea soup within a varied diet, using low-content European cultivars, matches none of the situations described in the neurolathyrism literature, which concerns populations dependent on this single seed during food shortages. The precaution that remains justified is not to make it an exclusive daily staple, and not to eat immature seeds.

The beluga lentil

Small, black and glossy, it holds its shape through cooking and keeps its grain, which suits salads and patties. Nutritionally it does not differ appreciably from other whole lentils.

Tarbais bean and pink lentil

The tarbais, from south-west France, has a thin seed coat that withstands long cooking. Pink or red lentils, hulled, break down quickly and thicken soups and dals without soaking. Comparing values in the Swiss database shows that hulling has a cost: hulled lentils provide 18.2 grams of fibre per 100 grams against 23.2 for whole lentils, and 88 micrograms of folate against 150.

The reference portion, and what it actually delivers

Swiss dietary recommendations set one daily portion of a protein-rich food, corresponding to 60 grams of dry legumes or 120 grams of tofu. That is the quantity the figures should be tied to, not the 100 grams of the tables.

Cooking yields calculated from the Swiss nutrient database give, for 60 grams dry:

LegumeCooked weightProteinFibreIronFolate
Whole lentils166 g14.6 g13.9 g4.8 mg90 µg
Hulled lentils147 g16.2 g10.9 g3.4 mg53 µg
Chickpeas153 g11.2 g12.6 g3.2 mg204 µg
Beans, all species163 g12.5 g13.6 g3.8 mg246 µg
Dry peas142 g12.8 g7.0 g2.6 mg120 µg
Soya153 g22.9 g13.2 g4.0 mg150 µg

The most interesting number in that table is not protein, it is fibre. A portion of whole lentils provides 13.9 grams of fibre, 46 per cent of the 30 gram daily benchmark. No other everyday food delivers as much in a realistic portion.

Iron deserves a careful reading. The 4.8 milligrams in a portion of whole lentils represent 32 per cent of the 15 milligram daily benchmark for women, but this is non-haem iron, whose absorbed fraction is low and variable and depends heavily on what else is in the meal. The page on plant iron absorption covers this in detail.

The gap between dry peas and lentils on fibre, 7.0 against 13.9 grams for the same portion, also shows that these seeds are not nutritionally interchangeable, contrary to what lists that group them together suggest. That fibre density also explains their place in the discussion of the glycaemic index in plant-based eating.

The cereal and legume pairing rule, as it is understood today

The idea that legumes and cereals must be paired at every meal to obtain a complete amino acid profile has circulated since the 1970s. It has been qualified.

A review published in 2025 in the Journal of Nutrition notes that protein quality is assessed by chemical scores such as DIAAS, but that these methods do not capture the metabolic activity of food-derived amino acids, and that overreliance on a single metric leads to generic recommendations lacking individual context. The authors observe that diets high in whole plant-derived proteins may require greater total protein and energy intakes to compensate for lower protein quality, and that for incomplete plant proteins, consuming complementary sources remains relevant.

The practical nuance is this: complementarity operates across the day and the diet as a whole, not across each plate. Nothing requires serving rice with every portion of lentils. What counts is the variety of sources over the week and sufficient total protein intake, which the complete guide to plant proteins sets out.

What studies find, and what they do not

Legumes are credited with many benefits. The state of the evidence is more mixed than the lists of arguments suggest.

On mortality, a systematic review with dose-response meta-analysis in Advances in Nutrition, covering 32 cohorts, 1,141,793 participants and 93,373 deaths, found an association between high legume consumption and reduced all-cause mortality, with a hazard ratio of 0.94, plus reduced stroke mortality at 0.91. No significant association appeared for cardiovascular, coronary or cancer mortality. A 50 gram daily increase was associated with a 6 per cent reduction in all-cause mortality. The authors themselves rate the certainty of evidence as low to moderate.

On cardiovascular disease and type 2 diabetes, a systematic review in Food and Nutrition Research including 47 studies reaches an even more sober result: meta-analyses of cohort studies suggest null associations for cardiovascular disease, coronary heart disease, stroke and type 2 diabetes. The randomised trials, meanwhile, show measurable effects on intermediate markers: total cholesterol lowered by 0.22 mmol/l, LDL by 0.19 mmol/l, fasting glucose by 0.19 mmol/l. Heterogeneity is high.

These results align with two older meta-analyses. One, covering ten randomised trials and 268 participants, measured a fall in total cholesterol of 11.8 mg/dl and in LDL of 8.0 mg/dl with a diet rich in non-soy legumes. The other, covering 41 trials and published in Diabetologia, found lower fasting glucose and insulin with pulses alone, and lower glycated proteins when pulses formed part of a low glycaemic index diet.

The honest reading: the effect on blood lipids and glucose is measured repeatedly in controlled trials, the effect on clinical events remains uncertain, and the association with all-cause mortality is real but modest. That is reason enough to eat them regularly, without treating them as a remedy.

The safety point recipes leave out: bean lectin

The common bean, Phaseolus vulgaris, contains phytohaemagglutinin, a lectin whose consumption in raw or inadequately cooked seeds causes acute symptoms of food poisoning.

A study published in Foods in 2025 quantified this lectin in dark red kidney beans, the type most concerned. ELISA assay estimated active lectin at 223 milligrams per gram of dry weight, and the amount contained in four raw soaked seeds at 544 milligrams, an order of magnitude associated with poisoning. Commercially cooked canned beans, by contrast, showed a reduction of more than 99 per cent, at 4.9 micrograms per gram: an entire can, 94 grams of dry matter, provides 0.46 milligrams, roughly a thousand times less than the amount associated with poisoning.

The practical consequence is clear and applies to every local dry bean variety, borlotti and tarbais included. Soaking followed by gentle low-temperature cooking, in a slow cooker for instance, is the scenario to avoid: a proper rolling boil is needed. Commercial canned beans do not pose this problem.

Soaking and cooking, method

Soaking shortens cooking time and lowers the content of the oligosaccharides responsible for intestinal fermentation, which are water-soluble. Eight to twelve hours suffice for beans, chickpeas and dry peas. Whole lentils do not need it, hulled lentils even less. The page on the benefits of soaking and sprouting covers what happens during those hours.

The method that follows from the data above:

  1. Rinse the seeds, cover with three times their volume of cold water, leave eight to twelve hours at room temperature, or in the fridge beyond twelve hours.
  2. Discard the soaking water and rinse. Part of the oligosaccharides has moved into it.
  3. Cover with unsalted water, bring to a proper rolling boil and hold it for at least ten minutes before lowering the heat, for beans in particular.
  4. Simmer until the grain crushes without resistance between two fingers, generally 20 to 45 minutes depending on the seed and its age.
  5. Salt at the end of cooking.

On that last point, traditional advice says salt toughens the skin. A simpler approach is to wait because the final volume, and therefore the salt concentration, is only known at that stage. A stock salted at the start invariably ends up too salty.

Canned remains a legitimate option, provided the label gets read. One hundred grams of canned peas provide 270 milligrams of sodium, 0.68 grams of salt; a 200 gram portion provides 1.35 grams, 27 per cent of the 5 gram daily limit set by the WHO. The no-added-salt version still contains 240 milligrams per 100 grams. Rinsing canned seeds removes part of that salt. The Swiss context makes the question concrete: the 2022/23 national salt study measured an average intake of 8.7 grams a day, around 75 per cent above the recommendation, with only 14 per cent of the population below 5 grams.

To organise all this without spending every evening on it, the page on legume batch cooking covers soaking, cooking and freezing in series.

Digestibility: what works and what is habit

Three actions have a documented or mechanically plausible effect. Discarding the soaking water removes part of the water-soluble oligosaccharides. Rinsing cans does the same. Increasing quantities gradually gives the gut microbiota time to adapt, which is the most decisive factor in practice.

For people who need to avoid several allergens at once, the allergen-free cooking guide covers compatible substitutions.

Carminative spices, cumin, fennel, coriander and ginger, are traditionally recommended. Their culinary value is real; evidence for an effect on intestinal gas from legumes is thin. The same applies to kombu added to the cooking water.

For people following a low-FODMAP protocol, the question is handled differently, with measured portions rather than tricks: the page on low-FODMAP eating and vegetarianism gives the published quantities.

Five common mistakes

Treating lupin flour as just another flour. It is a mandatory declared allergen in Switzerland, with a minimal eliciting dose measured at 1 gram of flour in sensitised individuals, and confirmed allergy in 44 per cent of peanut-allergic subjects tested.

Cooking dry beans too gently. Long cooking at low temperature, without a proper boil, is precisely the scenario associated with phytohaemagglutinin poisoning.

Believing cereals and legumes must be paired at every meal. Amino acid complementarity operates across the day and across the variety of sources, not on the plate.

Counting the declared iron as absorbed iron. Tables give total iron; the fraction actually absorbed from plant sources is low and depends on the meal.

Buying canned without reading the sodium. Two hundred grams of canned peas provide 27 per cent of the daily salt limit before any cooking.

Where to find them in Switzerland

The channels stay narrow, which follows directly from the 1600 hectares destined for human food. Producer markets, bulk shops and specialist grocers are the most reliable access points, with the advantage of being able to ask about origin and variety. Group orders and producer associations give access to rare varieties, often sold direct, since Grangeneuve notes that harvests are mostly sold through that channel for want of structured buyers.

On the label, three points deserve attention: the presence of lupin, highlighted where it applies; the sodium for cans; and the harvest year for dry seeds, since an old seed takes appreciably longer to cook. The page on reading a nutrition label explains how to convert declared sodium into salt.

The ideal pantry for a beginner vegetarian places these purchases in a coherent whole.

Buying and storing

Dry legumes come in 250 to 500 gram packets in supermarkets, and usually cheaper loose in bulk shops. Loose sales carry no nutrition table, so origin and variety have to be asked for at the counter.

Harvest year is worth insisting on. A bean from the last harvest cooks in the lower half of the stated range; a three-year-old one can still be hard after two hours, regardless of soaking. A batch that refuses to soften is almost never a cooking problem, it is old stock.

Storage comes down to three points. An airtight container away from light and heat is enough; dry seeds keep one to two years that way, while steadily becoming harder to cook. Cooked legumes keep about three days in the fridge and freeze well in portions, with or without their cooking liquid. Milled products such as chickpea or lupin flour go rancid faster than whole seeds and belong in a cool place once opened.

For cans, a second figure matters alongside sodium: the drained weight. A 400 gram can holds between 230 and 265 grams drained depending on the brand, the rest being liquid. Anyone working from a recipe calling for 250 grams of cooked legumes will only just manage with one can, which shopping lists routinely underestimate.

Adapting by profile

With a peanut allergy. Lupin should be set aside until an allergist has ruled, given the measured cross-reactivity.

For a sensitive gut. Start with hulled lentils, which contain less fibre, and build up gradually rather than adding spices meant to compensate.

For athletes and high protein intakes. Soya stands out clearly, with 22.9 grams of protein per reference portion against 11.2 for chickpeas. The other local legumes are excellent fibre sources before they are dense protein sources, and the comparison of tofu, tempeh and seitan places processed products in the same frame.

On a tight budget. Dry seeds in bulk remain among the cheapest protein sources, and batch cooking amortises the time. Heritage varieties, though, cost considerably more than ordinary lentils, because of the small cultivated area.

For gardeners. Legumes require no nitrogen fertilisation and leave part of the fixed nitrogen to the following crop, which makes them a useful head of rotation in a vegetable patch.

Limitations of the available data

The nutritional values cited come from the Swiss nutrient database, which lists generic categories such as beans, all species. They do not describe any particular local variety, whose composition may depart from the average.

The cooking yields in the table are calculated from the dry and cooked pairs in that same database, assuming protein is conserved. They are orders of magnitude, not direct measurements.

The neurolathyrism data come from Ethiopian and Indian populations whose diet depends on grass pea during food shortages. Transposing them to occasional consumption in Europe rests on the dose-response relationship, not on direct European observation.

The mortality studies are observational, with certainty of evidence rated low to moderate by their authors, and the included cohorts generally show low legume consumption.

Finally, the 4769 hectare figure and the one-third share destined for human food come from an agricultural advisory synthesis drawing on swissgranum and Agristat 2024, not from a primary survey consulted directly.

Frequently asked questions

Do lentils really need soaking? Not whole lentils, and hulled ones even less. Soaking mainly concerns beans, chickpeas and dry peas, where it shortens cooking time and lowers oligosaccharide content.

Are canned legumes less nutritious? The differences concern mainly sodium and some water-soluble vitamins. Rinsed, they are a perfectly valid option, and industrial cooking removes the lectin question.

Can you eat legumes every day? Swiss dietary recommendations set one daily portion of a protein-rich food and suggest legumes at least once a week. Daily consumption is compatible with those benchmarks, with varied sources.

Does soaking destroy nutrients? It moves part of the minerals and oligosaccharides into the water. The balance stays favourable, because the same operation improves digestive tolerance and shortens cooking.

Are legumes fattening? They provide between 111 and 157 kilocalories per 100 grams once cooked, with high fibre density, which places them among the most satiating foods per calorie.

Is grass pea banned in Switzerland? It is not banned. The documented precaution is not to make it an exclusive staple and to avoid immature seeds, since the described neurolathyrism cases concern populations that depended on it almost exclusively.

What is the difference between red and pink lentils? They are hulled lentils, without the seed coat. That hulling, rather than the variety, explains their fast cooking, their breakdown and their lower fibre and folate content.

Are legume flours as good as the seeds? They supply essentially the same nutrients, but milling raises the rate of starch digestion. For lupin flour, the allergen question outweighs the nutritional one.

Sources