Lentils are the most-traded pulse from Ukraine and the one buyers most often mis-specify, because the trade names carry an implication of botanical difference that they do not reliably deliver. This pillar sets out what actually distinguishes the three colour classes and what belongs in a lentil contract.
The three classes in practice
Red. Sold dehulled, usually split, occasionally whole-dehulled as football lentils. The cotyledon is orange to salmon and it cooks fast and breaks down completely. That disintegration is a feature: red lentils are the base of dhal, of thickened soups and of purees where no visible particle is wanted. Because the seed coat is removed, red lentils have no hull fibre and cook without soaking.
Green. Whole, seed coat intact, ranging from a pale to a deep green with mottling in some types. Firmer bite than brown after cooking, and generally the premium retail line. Salad, side dish and premium ready-meal use.
Brown. Whole, seed coat intact, khaki through to deep reddish brown. The volume workhorse: canning, industrial soup, food service and private-label retail. Holds shape well and needs no dehulling, which makes it the cheapest of the three to prepare.
Processing: where the cost is created
Whole green and brown lentils go through cleaning, gravity separation, destoning, optical sorting and calibration, and that is the whole story. The only losses are the removals.
Red lentils add two operations. Dehulling abrades the seed coat off, and splitting separates the two cotyledons. Both are yield losses, and the second one also generates fines and broken material that has to find an outlet. A dehulling and splitting line is therefore only economic where there is somewhere for the by-product to go, usually flour or feed. This is set out in the split and whole comparison and in the cleaning and grading guide.
Calibre and why it drives complaints
Lentils are graded by sieve fraction, quoted against a test sieve specification so that two laboratories get the same answer. Large and small fractions of the same colour class serve different buyers: canners generally want the visible large particle, soup blenders often prefer the small fraction for faster hydration.
Mixed calibre is the leading cause of the complaint that a lot “cooks unevenly”. Small seeds finish before large ones, so a spread of sizes produces some seeds broken down and some still firm in the same pot. Kitchens read this as a quality defect in the lentil; it is a grading defect. Tightening the calibre specification resolves more cook complaints than any other single change.
Colour, and what a colour sorter can and cannot do
Optical sorting removes off-colour seed, discoloured seed, residual foreign material and stones that survived gravity separation. It cannot make a two-shade lot into one shade; it can only remove the tail. A lot that looks like two lots mixed was blended before it reached the sorter, and sorting it to a single shade means rejecting a large share of good product.
Sensitivity is a trade-off. Tightening the reject threshold removes more defect and also removes more sound seed, so a very tight colour clause directly reduces yield and therefore cost. Rank your parameters: state which one governs if two conflict at intake.
Cook behaviour as a contract clause
Cook time, hydration ratio and unhydrated seed count are testable and belong in the specification, with the test protocol and the water hardness fixed. Two honest cook tests on the same lot can disagree if one is run on hard water and one on soft. The cooking specification guide sets out a protocol that both parties can run.
Storage history matters as much as age. Warm humid storage produces a hard-to-cook condition that no analysis on the certificate will reveal. Declare harvest year, ask about the store, and treat cook time as a tested parameter.
Where each class ends up
Canning. The single largest industrial outlet for whole lentils. The canner runs a fixed retort schedule and needs a lentil that reaches full hydration inside it without breaking up, which is why consistency beats absolute cook speed. Calibre uniformity is the parameter that does the most work here.
Soup and ready meals. Whole lentils where the recipe wants visible particles, red lentils where it wants body and thickening without particles. The sensitivity in this channel is to stones and hard foreign matter, because a hard foreign body in a finished retail product is a recall rather than a complaint.
Food service. Catering and institutional buyers prepare from dry in 25 kg bags. Their dominant complaint is uneven cook, and their dominant commercial metric is hydration ratio, because that determines portions per bag.
Retail private label. Packers take cleaned bulk and pack to their own artwork. Visual tolerances are the tightest of any channel and rejection is usually visual rather than analytical: one dark seed in a transparent pack is a consumer complaint.
Milling and protein. Off-calibre, split-out and lightly damaged fractions go to flour and to protein extraction. That outlet is what makes tight visual grading viable on the main stream, and it is covered in the flour and isolate guide.
Season and contracting
Ukrainian lentils are harvested from July into August and cleaned to order, so shipping runs year round from store. Calibre choice is widest between September and December and narrows through the spring as carry-over stock is picked over. Buyers needing a tight large-calibre fraction should contract early in the marketing year; the seasonality guide covers the calendar in full.
Contracts normally run under GAFTA standard forms with the quality vocabulary borrowed from Codex. Shipment is ambient dry cargo, with European destinations moving by road and Turkish, Levant and North African volume through the Black Sea via Istanbul or Mersin.