Tea time, rebuilt

Chapter 24 — Tea Time, Rebuilt

This chapter will save you more energy than any other chapter in this book, and it is worth saying so plainly before anything else. Not the rice chapter, not the coconut chapter, not the sadya chapter. This one. The reason is that four o'clock in Kerala is not a meal, and so nobody counts it.

Do the arithmetic out loud. One pazham pori from a tea shop — about 80 g of ripe nendran in a flour batter, deep-fried, absorbing roughly 14 g of oil — comes to about 307 kcal, 14 g of fat and 12 g of saturated fat. One 150 ml glass of chaya made the way it is actually made, with three-quarters milk and two level teaspoons of sugar, comes to about 119 kcal, of which 12 g is sugar — 112 ml of whole milk at 73 kcal per 100 ml (IFCT L002), 8 g of sugar at 4 kcal a gram, and about 5 kcal of tea solids. Together: about 425 kcal, eaten standing up, between lunch and dinner, on most days of most years. Over a year that is roughly 155,000 kcal — the equivalent of about 78 days of a 2,000 kcal diet, taken in fifteen-minute instalments that no one records and no one describes as eating. Two snacks instead of one, which is entirely normal, puts it past 730 kcal a day. Nothing else in Kerala food comes close to that as a single, invisible, daily load.

This is not a moral failing and it is not a reason to be ashamed of a tea shop. The chayakkada is arguably the most important public institution 20th-century Kerala produced. Its glass case of pathiri and unnakkaya and bonda, its enamel kettle, its shared newspaper and its arguments are where the state's literacy became a conversation rather than a statistic. Kerala's snack repertoire is the most developed in India because it had somewhere to live — three somewheres, each with its own cuisine and its own hour: the chayakkada by day, the shaap in the afternoon, the thattukada at night. The state's bakery culture starts, datably, with Mambally Bapu's bakery in Thalassery in 1880 and Western-style cake in 1883. In Malabar the Mappila kitchens built a repertoire of hundreds of items — unnakkaya, mutta mala, pazham nirachathu, vattayappam — held together by unbroken mother-to-daughter transmission inside the tharavaad, which is exactly why it is so technically elaborate. None of this deserves to be replaced with a protein bar.

What it does deserve is honesty about the cost, and then a set of dishes that keeps the repertoire and drops the frying oil. This chapter works along three lines.

Steaming and baking instead of deep-frying. The largest lever, and purely mechanical: a batter dropped into hot oil absorbs oil, and a batter put into an oven does not. The saving on one ethakka appam is about 117 kcal and 11 g of saturated fat, available without touching the banana, the flour or the sugar. Kerala already steams better than almost any cuisine on earth — puttu, idiyappam, idli, ada, vattayappam and kozhukkatta are all steamed — so this is not an imported technique. It is simply the one the tea shop does not use, because frying is faster over a single flame and holds better in a glass case.

The banana doing the work of the jaggery. A ripe nendran carries about 25 g of available carbohydrate and 22 g of sugars per 100 g, and it tastes of something jaggery cannot imitate. Most traditional tea-time sweets add jaggery on top of that, where it is doing very little the fruit is not already doing. Cut it by half to two-thirds and the dish tastes more of banana. Where this book keeps jaggery it says how much and why.

A serious strategy for the tea itself. Sugar in tea is the easiest saving most readers will ever make, and the one most often attempted badly — all at once, on a Monday, abandoned by Thursday. R106 gives a proper Kerala chaya method, the arithmetic of two teaspoons four times a day over a year, and a staged reduction that works because taste preference adapts when it is moved slowly. There is no moralising in it and no advice about sweeteners, because the evidence base behind this book contains nothing about them.

One last thing about portions. The recipes here state a weight or a count for a reason. A tea-time snack has no plate and no serving spoon, which is exactly why it is so easy to eat three. Weigh the first one. After a fortnight your hand will know.

A note on the panels in this chapter. Figures are computed from this book's nutrition tables — ICMR-NIN Indian Food Composition Tables 2017 and USDA FoodData Central SR Legacy — divided by the stated portions and labelled estimated. Ingredients with no entry in those tables (jaggery, raisins, ginger, green chilli, cardamom, tea, whole spices) are computed from the closest documented food, named under each panel: jaggery costed as sucrose at 98 g of carbohydrate and about 390 kcal per 100 g, the working assumption this book states in its reference tables, and raisins costed as sucrose at 400 kcal and 100 g of sugars per 100 g, which overstates them; ginger as tapioca (IFCT F015); green chilli as bitter gourd (D004); tea, cardamom and whole spices as Bengal gram dal (B001). Sesame oil energy is taken as 892 kcal per 100 g from the USDA oil figure, with saturated fat at 16.25% of fatty acids from IFCT 2017 Table 12. "Net carbs" is tabulated available carbohydrate; "carbohydrate" is available carbohydrate plus fibre; sugars are estimated as the free-sugar fraction of available carbohydrate, since IFCT does not report sugars separately.

R100 · Baked Ethakka Appam [LOW SAT FAT] [VEGAN] [UNDER 250 KCAL]

Pazham pori out of the oven, and what that costs

ICONS: B · L

THE RE-ENGINEERING: The frying is gone and nothing else has changed except the jaggery, which halves. A fried portion absorbs about 14 g of oil; a baked one absorbs the 2.5 g you brush on. Energy falls from about 332 kcal to 215 kcal a portion and saturated fat from about 12 g to 0.6 g.

Say what is lost first, because it is real. Deep-frying makes a shell that shatters, because water leaving the batter at 180°C (350°F) blows it into a rigid lace. An oven cannot do that. What you get instead is a crust that is crisp at the edges and tender-chewy in the middle, closer to a good baked ada than to a pazham pori. It is a different pleasure, not a diminished one, and the banana tastes far more of banana without frying oil over it.

Serves 4 · two pieces, about 120 g per portion Active 20 minutes · Total 45 minutes Difficulty: Easy

EQUIPMENT: Oven; heavy baking tray; baking parchment; pastry brush. Substitute: an air fryer at 190°C (375°F) for 12–14 minutes gives a slightly crisper result in two batches. A tawa or heavy frying pan with a lid also works — 4–5 minutes a side over a low flame, with a brushing of oil — and gives the best crust of the three.

BEFORE YOU START: Heat the oven and the tray together. Batter meeting a hot tray sets instantly and does not spread into a puddle, which is the single difference between a good baked version and a bad one.

INGREDIENTS

• 400 g (2 medium) ripe nendran plantains, peeled and cut lengthways into 4 slices each
• 50 g (6 tbsp) rice flour
• 20 g (2½ tbsp) refined wheat flour (maida)
• 15 g (1 tbsp, packed) jaggery, grated
• 2 g (⅓ tsp) ground turmeric
• 3 g (½ tsp) ground cardamom
• 1 g (¼ tsp) salt
• 90 ml water
• 10 g (2 tsp) sesame (gingelly) oil, for brushing

SOURCING NOTE: Nendran is the plantain this dish is built on — long, angular, ripening to deep yellow, sold as "Kerala banana" or "nendran" frozen and fresh in South Asian groceries. Where it is unavailable, a large ripe plantain from a Caribbean or African grocer is the right substitute; ordinary dessert bananas are too soft and will collapse. The fruit must be ripe enough to smell sweet through the skin but still firm enough to slice cleanly — a fully soft banana turns to purée on the tray.

ALLERGENS: Sesame, gluten (wheat).

METHOD

1. Heat the oven to 200°C (400°F) with the baking tray inside it.
2. Whisk the rice flour, wheat flour, jaggery, turmeric, cardamom and salt with the water into a batter the consistency of single cream — thick enough to coat a slice and thin enough to drip off in a steady thread, not a blob. Rest it 10 minutes so the flours hydrate.
3. Slice the plantains lengthways into four flat pieces each, about 8 mm thick. Thinner and they dry out; thicker and the centre stays starchy.
4. Line the hot tray with parchment, working quickly, and brush the parchment with half the sesame oil.
5. Dip each slice into the batter, let the excess run off for two or three seconds, and lay it on the hot tray with space between the pieces. Crowding makes them steam.
6. Brush the tops with the remaining oil and bake 12–14 minutes, until the batter has set opaque and is beginning to blister and colour at the edges.
7. Turn each piece with a thin spatula and bake a further 8–10 minutes, until both sides are golden with darker freckles and the underside lifts cleanly without sticking. Golden, not brown — the batter contains sugar and goes from golden to bitter in about two minutes.
8. Rest 3–4 minutes on the tray before lifting. Straight from the oven the crust is soft; after three minutes it firms up as the surface steam leaves.

WHY THIS WORKS

Deep-frying transfers heat through oil, and the oil that carries it stays in the food — 14 g a portion, at 9 kcal a gram. An oven transfers heat through air, so the same starch gelatinisation and browning happen with only the fat you brush on. The turmeric matters: it supplies the golden colour the eye reads as fried, and appearance is part of how a snack satisfies.

PER SERVING, two pieces (estimated)

AND THE TRADITIONAL FRIED VERSION, same fruit and batter, 30 g jaggery, 14 g of coconut oil absorbed per portion (estimated)

*A saving of 117 kcal and 11.4 g of saturated fat per portion. Substitutions: jaggery costed as sucrose at 98 g of carbohydrate and about 390 kcal per 100 g, the working assumption this book states in its reference tables; cardamom as Bengal gram dal (IFCT B001); sesame oil as noted in the chapter panel note. Oil absorption at 14 g per portion is an estimate from ordinary batter-fried practice, not a measured figure.

IF YOU ARE WATCHING… Blood sugar: this is still 43 g of net carbohydrate and 26 g of sugars a portion, and baking does not change that. Ripe nendran is a sugary fruit and no cooking method makes it otherwise. Eat one piece rather than two, eat it after a meal containing a pulse rather than on an empty stomach at four o'clock, and treat it as the carbohydrate of that meal rather than an addition to it. Cholesterol: the reason to make this version. Saturated fat drops from about 12 g to 0.6 g a portion, which over a daily habit is the largest single saturated-fat reduction available anywhere in this book. Blood pressure: 100 mg a portion, and it can be nil — the salt is there for flavour balance, not structure.

SWAP IT

• Ragi flour for the refined wheat flour → Consequence: earthier, darker, less crisp, and fibre rises by about 2 g a portion.
• No jaggery in the batter → Consequence: sugars fall from 26 g to about 22 g and the crust browns less, because there is less sugar to caramelise. With a properly ripe fruit you will not miss it.
• Green plantain instead of ripe, sliced thin → Consequence: a different dish entirely, savoury rather than sweet, at about 17.6 g of available carbohydrate per 100 g against ripe nendran's 25 g. Serve it with R093 rather than with tea.
• Fried, honestly, in coconut oil → Consequence: the crust you remember, and the second panel above. If you are going to fry, fry once a fortnight and weigh the oil that goes in and the oil that is left, so you know what you ate.

SCALE IT

At eight servings use two trays on two shelves and swap them at the turn, or bake in two batches. A crowded tray steams and the pieces go pale and soft.

MAKE AHEAD · STORE · FREEZE Best within 20 minutes of leaving the oven. Fridge 2 days in a covered box, refreshed in a 180°C (350°F) oven for 5–6 minutes — a microwave turns the crust to rubber. Freeze 1 month baked and cooled, reheated from frozen at 180°C (350°F) for 10–12 minutes. Make ahead by mixing the batter up to 4 hours in advance and keeping it covered in the fridge; slice the fruit only at the last moment, because cut plantain darkens within 15 minutes.

LEFTOVERS Chop cold pieces and fold them through R101's batter for the next morning's pancakes, or mash them with a spoonful of curd and a pinch of cardamom for a child's breakfast.

SERVE WITH R106 Chai, unsweetened or at half a teaspoon · R111 Sulaimani · R105 Roasted Kadala Mix, if one sweet piece is not enough

TROUBLESHOOTING Soggy and pale: the tray was cold, or the pieces were crowded. Both are fixable and both are the usual fault. Batter slid off: the plantain was wet. Pat the slices dry before dipping. Bitter edges: the sugar in the batter caught. Pull them at golden; two minutes past golden is too far. Starchy in the centre: the slices were too thick, or the fruit was underripe.

NOTES

R101 · Ragi Banana Pancake [LOW SODIUM] [LOW SAT FAT] [VEGETARIAN]

Finger millet, banana and no added sugar at all

ICONS: B · F · L

THE RE-ENGINEERING: No sugar and no jaggery — 200 g of ripe banana supplies all the sweetness the dish needs. Ragi replaces most of the flour, taking fibre from about 1 g to 5 g a serving, and the pan takes 8 g of oil across eight pancakes rather than the 30 g a batter this size usually gets.

Ragi is not a Kerala grain by inheritance so much as by adoption, and its honest advantages are two: about 11 g of fibre per 100 g against rice's 2.8 g, and a great deal of calcium. Its protein, at 7.2 g per 100 g, is actually lower than rice's. What it also has is a slightly bitter, malty flavour that ripe banana handles better than anything else, which is why this pairing turns up wherever ragi is eaten. Ten minutes of mixing, ten of cooking, and it holds a child through to dinner.

Serves 4 · two pancakes, about 125 g per portion Active 20 minutes · Total 35 minutes (plus 15 minutes resting the batter — recommended) Difficulty: Easy

EQUIPMENT: Heavy non-stick or well-seasoned cast-iron tawa, 20 cm; ladle. Substitute: an appachatti gives thicker, softer, bowl-shaped pancakes and takes a minute longer each.

BEFORE YOU START: Mash the banana thoroughly before the flour goes in. Lumps of banana in a ragi batter do not break up later, and they scorch on the pan.

INGREDIENTS

• 200 g (2 medium) ripe bananas, well mashed
• 150 ml whole milk
• 1 egg (about 50 g), lightly beaten
• 120 g (1 cup) ragi flour
• 30 g (3½ tbsp) rice flour
• 20 g (¼ cup) fresh grated coconut
• 3 g (½ tsp) ground cardamom
• 1 g (¼ tsp) salt
• 8 g (1½ tsp) sesame (gingelly) oil, for the pan

SOURCING NOTE: Buy ragi flour from a South Asian grocer rather than a health-food shop, where the same thing costs three times as much under the name "finger millet flour". It goes rancid faster than wheat flour because of its higher fat content, so buy small bags and keep them in the fridge. For the banana, ordinary ripe dessert bananas — poovam, robusta, or any Cavendish — are correct here; nendran is too dense. Ragi flour older than three months tastes noticeably bitter, and no amount of banana hides it.

ALLERGENS: Egg, milk, sesame.

METHOD

1. Mash the bananas to a smooth purée with a fork, 1–2 minutes. A few fibres are fine; lumps are not.
2. Whisk in the milk and the egg until uniform.
3. Fold in the ragi flour, rice flour, coconut, cardamom and salt, mixing only until no dry flour remains, 30–45 seconds. Over-mixing a millet batter makes it gluey.
4. Rest the batter 15 minutes. Ragi absorbs water slowly, and a rested batter spreads evenly where a fresh one sits in a mound. It should fall from the ladle in a thick ribbon; add 20 ml more milk if it does not.
5. Heat the tawa over medium heat until a drop of water flicked onto it dances rather than sits, 2–3 minutes. Brush with a scant teaspoon of the oil.
6. Cook each pancake from a ladleful, spread to about 12 cm, until the surface has gone matt, bubbles have risen and burst across the whole face, and the edges lift when nudged, 2½–3 minutes. Ragi will not colour like wheat, so judge by the surface rather than by the colour.
7. Turn and cook a further 1½–2 minutes, until the second side is set and freckled and the pancake feels springy rather than doughy in the middle.
8. Stack the cooked pancakes under a cloth rather than on a rack; ragi pancakes stiffen as they cool and steam keeps them tender.

WHY THIS WORKS

Ragi has no gluten, so there is no network to trap gas — the structure comes from egg protein setting and from the banana's pectin, which is why both are here and why the batter must not be beaten. The banana also supplies all the sugar, and that matters beyond the arithmetic: fruit sugars arrive with fibre attached, where 15 g of jaggery arrives with none.

PER SERVING, two pancakes (estimated)

*Substitutions: banana computed as IFCT E012, ripe robusta; cardamom as Bengal gram dal (B001); sesame oil as noted in the chapter panel note.

IF YOU ARE WATCHING… Blood sugar: 40 g of net carbohydrate a serving, which is not small — this is a substantial snack, closer to a light breakfast. Two points of honesty. Millet's glycaemic advantage belongs to the coarsely milled grain; finely milled millet flour cooked into a soft batter loses much of it, and one study of finger-millet formulations measured a glycaemic index of 93. And low-GI eating as a pattern delivers about 0.3–0.5% off HbA1c, which is larger than any spice in this book but is a pattern effect, not a property of this pancake. Eat one pancake with a boiled egg rather than two on their own. Cholesterol: 4 g of saturated fat, mostly from the whole milk and the coconut. Semi-skimmed milk takes it to about 2.5 g. Blood pressure: 130 mg a serving.

SWAP IT

• No egg → 40 ml more milk plus 10 g more rice flour. Consequence: flatter, softer, more fragile to turn, and protein falls from 7 g to about 5 g a serving. The dish becomes vegetarian-to-vegan with a plant milk.
• Buttermilk for the milk → Consequence: tangier and lighter, and the acid gives a slightly better rise. Sodium rises by about 40 mg a serving, because cultured buttermilk carries around 148 mg per 100 g before anything is added.
• Half ragi, half wheat flour → Consequence: more familiar and easier to turn, and fibre falls from 5 g to about 3 g. (Allergen: this swap introduces gluten (wheat).)
• Add 20 g of chopped roasted peanuts to the batter → Consequence: better texture and about 30 kcal more a serving. (Allergen: this swap introduces peanuts.)

SCALE IT

Oil does not scale linearly; a hot pan needs less per pancake as it goes on.

MAKE AHEAD · STORE · FREEZE Batter keeps in the fridge 12 hours and thickens as it stands, so loosen it with 20–30 ml of milk before cooking. Cooked pancakes: fridge 2 days, freeze 2 months interleaved with parchment, reheated from frozen on a dry tawa for 2–3 minutes a side. Do not microwave them from frozen — they go leathery.

LEFTOVERS Cut cold pancakes into strips and eat them with curd and a spoonful of R098, which is better than it sounds. Or layer two with a spoonful of grated coconut between and warm them through on the tawa.

SERVE WITH R106 Chai · R107 Sambharam · R099 Carrot and Coconut Raita · a boiled egg, for the protein this dish is short of

TROUBLESHOOTING Gluey: the batter was over-mixed, or not rested. Falling apart on the turn: too thin, or turned too early. Wait for the edges to lift by themselves. Bitter: the ragi flour is old. There is no way to hide it. Dense and heavy: the batter was too thick. It should ribbon off the ladle, not drop off it.

NOTES

R102 · Pazham Nirachathu, Steamed [LOW SODIUM] [VEGAN] [UNDER 250 KCAL]

Stuffed nendran, from the steamer instead of the pan

ICONS: L

THE RE-ENGINEERING: No batter and no frying. The traditional version is slit, stuffed, dipped in an egg-and-flour batter and shallow-fried in ghee or coconut oil; this one is stuffed, wrapped in banana leaf and steamed. Jaggery in the filling drops from a usual 50 g to 20 g, because a properly ripe nendran does not need it.

This is a Malabar tea-time dish and an iftar dish, and it is one of the very few sweets in the Kerala repertoire that needs almost no re-engineering at all — the banana and the coconut were always the whole point, and the batter was only ever there to hold the filling in. A banana leaf holds it in better. Twenty minutes, one steamer, no oil.

Serves 4 · one stuffed half, about 135 g per portion Active 15 minutes · Total 35 minutes Difficulty: Easy

EQUIPMENT: Steamer, or a puttu kutti with the top removed, or a colander over a pan with a tight lid; banana leaf, or baking parchment. Substitute: parchment works exactly as well mechanically and gives up the faint vegetal aroma the leaf contributes — a real loss, but a small one.

BEFORE YOU START: Get the steamer to a rolling boil before the parcels go in. Parcels put into a cold steamer sit in condensation and the leaf tears.

INGREDIENTS

• 450 g (2 large) ripe nendran plantains, unpeeled
• 60 g (⅔ cup) fresh grated coconut
• 20 g (1½ tbsp, packed) jaggery, grated
• 15 g (1½ tbsp) raisins, roughly chopped
• 3 g (½ tsp) ground cardamom
• 2 pieces banana leaf, about 25 × 20 cm each, or parchment

SOURCING NOTE: The plantains must be ripe — the skin blackening at the tips, the fruit yielding to a thumb — because underripe nendran stays starchy through steaming and the dish tastes of raw flour. Frozen banana leaf is sold in every South Asian and South East Asian grocery and is better than fresh for this, because freezing has already broken the leaf's fibres so it folds without splitting. Pass fresh leaf briefly over a flame until it darkens and goes glossy, 5–10 seconds a side, to make it pliable.

ALLERGENS: Sulphites.

METHOD

1. Steam the whole unpeeled plantains over rolling water until a skewer slides into the thickest part with slight resistance, 10–12 minutes. They should be cooked through but still firm enough to slit and handle.
2. Toast the coconut in a dry pan over a low flame until it is pale gold and smells of biscuit, 3–4 minutes, then stir in the jaggery, raisins and cardamom off the heat. The residual heat melts the jaggery into the coconut without cooking it to caramel.
3. Peel the plantains while warm — the skin lifts away cleanly at this stage and fights you when cold — and halve each lengthways without cutting all the way through, so each half opens like a book.
4. Fill each half with a quarter of the coconut mixture, pressing it into the slit and closing the fruit around it.
5. Wrap each stuffed half in banana leaf or parchment, folding the ends under so the parcel holds itself closed.
6. Steam the parcels over rolling water 10–12 minutes, until the parcel smells strongly of cardamom and coconut and the leaf has gone from bright to olive-green.
7. Rest 5 minutes in the closed parcels before opening. The filling is molten straight from the steamer and will set to a fudgy consistency in five minutes.

WHY THIS WORKS

Steaming cooks the banana's starch fully without adding anything, and the closed parcel traps the coconut's and cardamom's volatile aromatics against the fruit instead of letting them escape into a kitchen. The toasting at step 2 matters more than it looks: Maillard products from browned coconut read as richness, which is what lets 20 g of jaggery do the job of 50 g.

PER SERVING (estimated)

*Substitutions: jaggery costed as sucrose at 98 g of carbohydrate and about 390 kcal per 100 g, the working assumption this book states in its reference tables, and raisins costed as sucrose at 400 kcal and 100 g of sugars per 100 g, which overstates them slightly; cardamom as Bengal gram dal (IFCT B001).

IF YOU ARE WATCHING… Blood sugar: be clear-eyed about this one. It is 38 g of net carbohydrate and 35 g of sugars a portion, nearly all of it from the fruit, and steaming does not change that. It is a sweet, and this book's position on sweets is about frequency rather than substitution. Halve the portion, eat it after a meal that contained a pulse, and do not eat it at four o'clock on an empty stomach. Cholesterol: 6 g of saturated fat, all of it from 60 g of fresh coconut. The traditional fried version adds 8–12 g of coconut oil on top of that. Cutting the coconut to 40 g takes this to about 4 g. Blood pressure: under 10 mg. There is no added salt in this dish.

SWAP IT

• Dates for the jaggery and raisins, 40 g stoned and chopped → Consequence: stickier, more caramel, a little fibre, and sugars land around 33 g rather than 35 g. A modest change, honestly reported.
• No jaggery at all, with a very ripe fruit → Consequence: sugars fall to about 30 g and the filling is looser. Perfectly good; most people will not notice.
• Shallow-fried in the traditional batter → Consequence: the crust, plus roughly 90–110 kcal and 8–10 g of saturated fat a portion from the absorbed oil and the egg batter. (Allergen: this swap introduces egg.)
• Green plantain, steamed and stuffed with a savoury coconut mix → Consequence: a different dish, at 17.6 g of available carbohydrate per 100 g against 25 g, and one that belongs beside R093 rather than after it.

SCALE IT

At eight servings steam in two batches; a crowded steamer runs cool and the parcels in the middle stay underdone.

MAKE AHEAD · STORE · FREEZE Assemble the parcels up to 1 day ahead and keep them in the fridge, then steam from cold with 3–4 minutes added. Cooked: fridge 3 days in their parcels; freeze 2 months wrapped, re-steamed from frozen for 15–18 minutes. Reheat by steaming, never in a microwave, which makes the fruit weep.

LEFTOVERS Slice cold leftovers into rounds and warm them on a dry tawa until the cut faces caramelise, 2–3 minutes a side — arguably better than the original. Or mash one into 200 ml of milk with a pinch of cardamom for a banana drink.

SERVE WITH R106 Chai, unsweetened · R111 Sulaimani · R107 Sambharam

TROUBLESHOOTING Starchy and bland: the fruit was underripe. Nothing fixes this. Filling ran out into the steamer: the slit went all the way through, or the parcel was loosely folded. Leaf split: fresh leaf that was not heated first. Pass it over a flame until it goes glossy. Watery: condensation dripped from the lid onto the parcels. Lay a cloth under the lid next time.

NOTES

R103 · Unnakkaya, Rebuilt [LOW SODIUM] [VEGETARIAN] [UNDER 250 KCAL]

The Mappila plantain pod, steamed rather than fried in ghee

ICONS: B · F · L

THE RE-ENGINEERING: Steamed instead of fried in ghee, and the jaggery in the filling cut from a usual 60 g to 20 g. The dough is the same pounded nendran it has always been. What goes is roughly 12–15 g of ghee a portion — about 110 kcal and 8 g of saturated fat — and nothing else.

Unnakkaya is Thalassery's signature, and the name describes the shape: a small tapering pod, deliberately made to look like one. Boiled nendran is pounded to a dough, flattened in the palm, wrapped around a filling of coconut, egg and raisin, sealed into a spindle and fried. The technique is the pleasure — it is a hand-shaped food, and Mappila kitchens taught it mother to daughter inside a single household for generations. Steaming keeps every part of that except the frying pan.

Serves 5 · two pods, about 130 g per portion Active 35 minutes · Total 55 minutes Difficulty: Moderate — the shaping takes a few pods to learn

EQUIPMENT: Steamer; potato masher or a heavy pestle; small pan for the filling; banana leaf or parchment to line the steamer basket. Substitute: a food processor will turn the plantain to glue rather than dough — use a masher. A ricer works well and is faster.

BEFORE YOU START: The dough must be worked while warm. Cold pounded plantain cracks when you shape it and no amount of kneading brings it back, so cook the filling first and have it cool and ready before the plantain comes out of the steamer.

INGREDIENTS

• 500 g (2 large) ripe nendran plantains, unpeeled
• For the filling:
• 60 g (⅔ cup) fresh grated coconut
• 1 egg (about 50 g), lightly beaten
• 20 g (1½ tbsp, packed) jaggery, grated
• 10 g (1 tbsp) raisins, chopped
• 3 g (½ tsp) ground cardamom
• 5 g (1 tsp) sesame (gingelly) oil, for your hands and the steamer basket

SOURCING NOTE: Nendran is not optional here — the dough depends on its dense, dry, low-water flesh, and a Cavendish banana will not hold a shape at all. Frozen nendran, thawed, works. The fruit should be ripe but not soft: yellow with black-tipped ridges, firm to a thumb. A fully soft banana gives a wet dough that will not seal, and the only remedy at that point is to add 20 g of rice flour, which changes the texture.

ALLERGENS: Egg, sesame, sulphites.

METHOD

1. Steam the whole unpeeled plantains over rolling water until a skewer goes through the thickest part with no resistance, 14–16 minutes. Fully soft, unlike R102 — this fruit is going to be pounded.
2. Cook the filling: warm the coconut in a dry pan over a low flame until pale gold and smelling of biscuit, 3–4 minutes. Add the beaten egg and stir constantly until it has set into fine curds through the coconut and no wet egg remains, 2–3 minutes. Take it off the heat, stir in the jaggery, raisins and cardamom, and set it aside to cool completely, 15 minutes.
3. Peel and pound the plantains while hot, discarding the central seed line if it is fibrous, working with the masher until you have a smooth, stiff, slightly tacky dough with no lumps, 3–5 minutes. It should hold a thumbprint.
4. Divide the dough into 10 pieces of about 50 g each and keep them covered with a cloth; the surface dries and cracks within five minutes uncovered.
5. Shape: oil your palms lightly, flatten a piece into a 9 cm oval about 6 mm thick, place a heaped teaspoon of filling down the centre, and fold the dough over it. Seal the long edge by pressing, then roll gently between your palms while tapering both ends into points. Any gap will open in the steamer.
6. Steam the pods on an oiled banana leaf or parchment, in a single layer with space between them, for 12–14 minutes, until the surface has gone from matt to faintly glossy and a pod lifts cleanly from the leaf without sticking.
7. Rest 5 minutes before moving them. Straight from the steamer they are fragile; after five minutes the starch sets and they firm up.

WHY THIS WORKS

Pounding hot plantain works its starch into a cohesive dough — the reason it must be done warm. Setting the egg into the coconut at step 2 does two jobs: it adds about 2 g of protein a portion, and it binds the filling so it cannot run out through a seam. Frying was never structural here; the dough holds itself together.

PER SERVING, two pods (estimated)

AND THE TRADITIONAL VERSION, 60 g jaggery and about 13 g of ghee absorbed per portion (estimated)

*A saving of about 135 kcal and 8 g of saturated fat per portion. Substitutions: jaggery costed as sucrose at 98 g of carbohydrate and about 390 kcal per 100 g, the working assumption this book states in its reference tables; raisins costed as sucrose at 400 kcal and 100 g of sugars per 100 g; cardamom as Bengal gram dal (IFCT B001); ghee costed at 99 g of fat and about 890 kcal per 100 g with a saturated fraction of about 62%, the working assumption this book states in its reference tables. Ghee absorption at 13 g a portion is an estimate from ordinary shallow-frying practice, not a measured figure.

IF YOU ARE WATCHING… Blood sugar: 32 g of net carbohydrate and 29 g of sugars a portion. This is a sweet and it should be treated as one; the rebuilt version is a better sweet, not a free one. One pod rather than two is 105 kcal and 16 g of net carbohydrate, which is a sensible tea-time item. Cholesterol: 5 g of saturated fat rather than 13 g. Over a household that makes unnakkaya for every Ramadan iftar, that is the largest single saving in this chapter after R100. Blood pressure: 20 mg a portion, from the egg. There is no added salt.

SWAP IT

• No egg in the filling → 15 g more coconut. Consequence: the filling is looser and more likely to escape, protein falls from 3 g to about 2 g, and the dish becomes vegan.
• Dates for jaggery and raisins, 40 g chopped → Consequence: a darker, chewier filling with a little fibre; sugars land around 27 g.
• Fried in ghee, honestly → Consequence: the second panel above, and a crisp shell the steamer cannot give. If it is Eid, fry them and count them. (Allergen: this swap introduces milk.)
• Baked at 190°C (375°F) for 18–20 minutes, brushed with 8 g of oil → Consequence: a drier, firmer pod with a little colour; about 25 kcal more a portion than steaming and about 110 kcal less than frying.

SCALE IT

At 20 pods steam in two batches. The dough dries while it waits, so keep the unshaped portions under a damp cloth.

MAKE AHEAD · STORE · FREEZE Shape the pods up to 1 day ahead and keep them covered in the fridge, steaming from cold with 3 minutes added. Cooked: fridge 3 days. Freeze 2 months shaped and raw, on a tray until solid then bagged, steamed from frozen for 18–20 minutes — this is the better way to freeze them. Cooked pods also freeze for 1 month and re-steam in 12–14 minutes. Reheat by steaming only.

LEFTOVERS Slice cold pods into 1 cm rounds and warm them on a dry tawa until the cut faces caramelise, 2 minutes a side. Or crumble two into 200 ml of warm milk with cardamom, which is a serviceable payasam and uses up the ones that split.

SERVE WITH R106 Chai, unsweetened · sulaimani — black tea with lime, the Malabar after-feast tea, at R111 · R107 Sambharam

TROUBLESHOOTING Cracked and split in the steamer: the dough was worked cold, or a seam was not sealed. Both are shaping faults, and the pods still taste right. Wet, unworkable dough: the fruit was overripe. Work in 20 g of rice flour, accepting a slightly pastier texture. Filling leaked: too much filling, or the seam was on the underside. A heaped teaspoon is the limit for a 50 g piece of dough. Gluey: the plantain was processed rather than mashed.

NOTES

R104 · Savoury Kozhukkatta [LOW SAT FAT] [VEGAN] [UNDER 250 KCAL]

Steamed rice dumplings with a moong dal filling

ICONS: P · B · F · L

THE RE-ENGINEERING: The traditional kozhukkatta is sweet — coconut and jaggery inside a rice shell. This one is savoury, and the filling is 120 g of moong dal, which takes the protein from about 2 g a serving to nearly 10 g. Nothing is fried at any stage, and the whole dish uses 8 g of oil.

Steaming is Kerala's great technique and the one that most obviously belongs in a tea-time chapter: puttu, idiyappam, ada, vattayappam and kozhukkatta are all cooked in steam, and none of them absorbs a gram of fat. What a tea-time snack almost never has is protein, and that is why most people are hungry again at six. A dumpling filled with seasoned dal is the highest-protein snack in this chapter, and at 246 kcal for three of them it is also the most filling.

Serves 4 · three dumplings, about 135 g per portion Active 40 minutes · Total 1 hour 10 minutes (plus 4 hours soaking the dal — required) Difficulty: Moderate — the dough needs a feel for it

EQUIPMENT: Steamer or idli stand; heavy pan for the dough; small pan for the filling. Substitute: a colander over a pan with a tight lid steams perfectly well. An idli stand gives flatter dumplings and needs 2 minutes less.

BEFORE YOU START: The dough is made with boiling water and must be kneaded while it is almost too hot to touch. Have the filling completely finished and cooled before you start it, because the dough stiffens as it cools and cannot be re-softened.

INGREDIENTS

• For the filling:
• 120 g (⅔ cup) green gram dal (moong dal), soaked 4 hours and drained
• 4 g (¾ tsp) sesame (gingelly) oil
• 2 g (⅓ tsp) mustard seed
• 10 g (about 2) green chillies, finely chopped
• 5 g (a 1 cm piece) ginger, finely grated
• 3 g (about 10 leaves) curry leaves, torn
• 35 g (⅓ cup) fresh grated coconut
• 1 g (¼ tsp) salt
• For the dough:
• 100 g (¾ cup) rice flour, preferably roasted
• 170 ml boiling water
• 4 g (¾ tsp) sesame (gingelly) oil
• 1 g (¼ tsp) salt

SOURCING NOTE: Use the fine rice flour sold for puttu or kozhukkatta — coarse rice flour will not make a dough that holds. If yours is unroasted, dry-roast it in a pan over a low flame until it smells faintly toasted and feels loose rather than sticky between the fingers, 4–5 minutes, and cool it before use; roasted flour absorbs water more predictably. Moong dal is in every supermarket. Do not use split-and-skinned moong that has been pre-cooked or "instant" — it goes to purée in the steamer.

ALLERGENS: Sesame, mustard.

METHOD

1. Steam the drained moong dal in a covered dish for 10–12 minutes, until a grain crushes between finger and thumb but the mass has not gone to paste. It should still be recognisably individual grains.
2. Temper: heat the 4 g of oil in the small pan until a mustard seed dropped in sizzles at once. Add the mustard and wait until it pops steadily, 30–40 seconds. Add the green chilli, ginger and curry leaf and fry 30–40 seconds, until the ginger smells sweet rather than raw.
3. Fold in the steamed dal, the coconut and the 1 g of salt, and cook 2–3 minutes over a low flame, until the mixture is dry enough to hold together when squeezed in a fist. A wet filling will burst the dough. Cool completely, 15 minutes.
4. Make the dough: bring the 170 ml of water to a full boil with the 4 g of oil and the 1 g of salt. Take it off the heat, tip in all the rice flour at once, and stir hard with a wooden spoon until it comes together into a shaggy mass with no dry flour, 30–45 seconds. Cover and leave 4–5 minutes.
5. Knead the warm dough in the pan with the back of the spoon and then, as soon as you can bear it, with a wet hand, until it is completely smooth and does not crack when you bend a small piece, 2–3 minutes. If it cracks, work in a tablespoon of boiling water and knead again.
6. Shape: divide into 12 pieces of about 23 g. With wet palms, flatten each into a 7 cm disc, place a teaspoon of filling in the middle, and draw the edges up and pinch them closed, then roll gently into a ball or shape into a flattened pouch. Keep everything not being worked under a damp cloth.
7. Steam the dumplings in a single layer over rolling water, 10–12 minutes, until the surface has gone from chalky-matt to faintly translucent and glossy and a dumpling lifts without sticking. Under-steamed kozhukkatta tastes of raw flour and there is no rescue.
8. Rest 4–5 minutes before serving. They firm up as the starch sets and are much easier to handle.

WHY THIS WORKS

Boiling water gelatinises the rice starch at once, which is what makes a glutenless flour behave like a dough. Keeping the dal grains whole rather than mashing them leaves the filling dry enough not to split the shell, and it keeps the pulse doing what pulses do: 50 g of legumes eaten with brown rice lowered glycaemic response 22.9% below white rice, against 19.8% for changing the rice alone.

PER SERVING, three dumplings (estimated)

*Substitutions: green chilli computed as bitter gourd (IFCT D004); ginger as tapioca (F015); mustard seed and curry leaf as Bengal gram dal (B001); sesame oil as noted in the chapter panel note.

IF YOU ARE WATCHING… Blood sugar: 37 g of net carbohydrate, but with 10 g of protein and 5 g of fibre in the same package, which is a very different proposition from 37 g of net carbohydrate in a fried banana. This is the dish in the chapter to eat if you are choosing on glycaemic grounds. Blood pressure: 200 mg a serving, all from 2 g of added salt across 12 dumplings. It tastes properly seasoned at 1.5 g, which brings it to 150 mg. Cholesterol: 4 g of saturated fat, almost entirely from the 35 g of fresh coconut. Sesame oil rather than coconut oil is specified throughout, at 16% saturated against coconut oil's 91%.

SWAP IT

• Cooked vanpayar (red cowpea) for the moong dal → Consequence: earthier and firmer, fibre up from 5 g to about 6 g a serving, and 10 minutes longer to cook. Arguably the better dumpling.
• Cooked kadala (black chickpeas), coarsely crushed → Consequence: the highest-fibre version of all, at about 8 g a serving, and a coarser filling that needs a thicker dough.
• Ragi flour for a third of the rice flour → Consequence: darker, nuttier and more fragile to shape, with fibre up by about 2 g a serving.
• Traditional sweet filling — 60 g coconut with 30 g jaggery → Consequence: the kozhukkatta everyone knows, at about 235 kcal and 3 g of protein a serving, with sugars up from 2 g to about 10 g. A sweet, not a savoury.
• Salt replaced with the 70:30 potassium blend (M07) → sodium falls from 200 mg to about 140 mg a serving. Consequence: it works well here, because the salt is inside a fatty coconut filling where potassium chloride's metallic edge is masked. Potassium-enriched salt is not for readers with chronic kidney disease, or on a RAAS inhibitor (an ACE inhibitor or an ARB) or a potassium-sparing diuretic, without a doctor's approval.

SCALE IT

Water does not scale exactly — flours differ in absorption. Hold back 20 ml and add it during the knead if the dough cracks.

MAKE AHEAD · STORE · FREEZE Make the filling up to 2 days ahead and keep it in the fridge; make the dough on the day, because it stiffens irreversibly overnight. Shaped raw dumplings: fridge 6 hours under a damp cloth. Cooked: fridge 3 days, re-steamed 5–6 minutes. Freeze 2 months shaped and raw, frozen on a tray then bagged, steamed from frozen for 16–18 minutes. Reheat by steaming; a microwave makes the shell tough.

LEFTOVERS Halve cold dumplings and crisp the cut faces on a dry tawa for 2–3 minutes a side, then eat them with R094. Or crumble them into a pan with a little R015 and a beaten egg for a good breakfast hash.

SERVE WITH R094 Tomato Chutney · R093 Coconut Chutney, Rebuilt · R106 Chai · R107 Sambharam

TROUBLESHOOTING Cracked dough: not enough water, or it was kneaded cold. Add a tablespoon of boiling water and knead again while hot. Burst in the steamer: the filling was wet, or overfilled. A teaspoon per 23 g of dough is the limit. Tastes of raw flour: under-steamed. Give them 3 minutes more; there is no other fix. Tough and rubbery: over-steamed, or reheated in a microwave. Sticky to shape: wet your palms rather than flouring them. Flour on the outside bakes into a chalky skin.

NOTES

R105 · Roasted Kadala Mix [HIGH FIBRE] [VEGAN] [UNDER 250 KCAL]

The honest answer to a savoury craving, weighed

ICONS: P · B · L

THE RE-ENGINEERING: This is the deep-fried mixture from the tea-shop jar, rebuilt as an oven-roasted one. Six grams of oil across the whole batch — under a gram a portion — instead of the 30–40 g a fried mixture absorbs. And it states a portion weight, because a mixture with no portion is a mixture you eat until it is gone.

Black chickpeas are the highest-fibre pulse in the Kerala repertoire by a wide margin: 25.2 g per 100 g, against 15.2 g in chana dal and 11.5 g in vanpayar. Soaked and roasted they go from a curry ingredient to something you eat with your fingers, and a 35 g portion carries 6 g of protein and 8 g of fibre for 116 kcal. That is the best satiety-per-calorie snack in this book. The portion is the whole recipe, so weigh it the first six times.

Makes about 245 g · Serves 7 · a weighed 35 g per portion — about three tablespoons, a closed palm's worth Active 20 minutes · Total 1 hour 5 minutes (plus 8 hours soaking — required) Difficulty: Easy

EQUIPMENT: Oven; large baking tray; clean tea towel. Substitute: an air fryer at 180°C (350°F) for 20–24 minutes, shaken every 6 minutes, in two batches. A heavy pan on the hob also works and needs near-constant stirring for 20–25 minutes.

BEFORE YOU START: The chickpeas must be dried thoroughly after soaking. Wet chickpeas steam rather than roast and come out leathery, and no extra oven time fixes it.

INGREDIENTS

• 200 g (1 cup) black chickpeas (kadala), soaked 8 hours in plenty of water, then drained
• 6 g (1¼ tsp) sesame (gingelly) oil
• 4 g (¾ tsp) Kashmiri chilli powder
• 2 g (⅓ tsp) ground turmeric
• 2 g (⅓ tsp) salt
• 30 g (⅓ cup) fresh coconut, sliced into thin chips
• 10 g (about 5 sprigs) curry leaves, stripped
• 3 g (½ tsp) M06 low-sodium all-purpose podi, to finish

SOURCING NOTE: Black chickpeas are sold as kala chana, kadala or "brown chickpeas" in every South Asian grocery and in most supermarkets now. Ordinary white chickpeas work and give a softer, less nutty result with about 10 g less fibre per 100 g. Do not use tinned chickpeas — they are already cooked soft and will not crisp. Fresh coconut chips are worth the trouble; desiccated coconut burns before it colours.

ALLERGENS: Sesame. Gluten (wheat) — only if your asafoetida is compounded with wheat flour, as most brands are (it arrives with M06).

METHOD

1. Soak the chickpeas 8 hours or overnight, in at least three times their volume of water. They will roughly double in weight.
2. Drain and spread them on a tea towel, patting the top dry and leaving them 15–20 minutes until the skins look matt rather than shiny.
3. Heat the oven to 190°C (375°F).
4. Toss the chickpeas with the oil, chilli powder, turmeric and salt until every one is coated and faintly orange.
5. Roast on the tray in a single layer, shaking it every 8–10 minutes, for 35–40 minutes, until the chickpeas are dry and hard to the fingernail, rattle on the tray, and one crushed between the teeth snaps rather than yields. If they still bend, give them 5 minutes more and check again.
6. Add the coconut chips and curry leaves for the last 5–6 minutes only, tossing once. Both burn in the time the chickpeas need, and both are done when the coconut is pale gold and the leaves crumble between finger and thumb.
7. Cool completely on the tray, 20–25 minutes. Warm chickpeas taste chewy and soften a sealed jar; cold ones are properly crisp.
8. Toss with the M06 podi and store. Adding the podi to warm chickpeas makes it clump.

WHY THIS WORKS

Soaking hydrates the starch so it gelatinises during roasting, and long dry heat then drives out enough water for the chickpea to go glassy and brittle — a texture frying reaches faster and at four times the fat. The fibre is the point: 8 g a portion, in a pulse, doing volume, satiety and glycaemic work at once. Pulses at 130 g a day lowered body weight by 0.34 kg in pooled trials.

PER PORTION, 35 g (estimated)

*The unit is a weighed 35 g portion of the finished mixture, because that is the only honest way to report a snack eaten by the handful. Substitutions: chilli powder, turmeric and curry leaf computed as Bengal gram dal (IFCT B001); M06 from its own panel in Chapter 15; sesame oil as noted in the chapter panel note.

IF YOU ARE WATCHING… Blood sugar: 13 g of net carbohydrate against 8 g of fibre, in a pulse. This is the snack in the book with the best case behind it, and it works best eaten around 4pm in place of something fried rather than in addition to it. Blood pressure: 120 mg a portion, from 2 g of salt across the batch. Halve the salt and lean on the M06 and it drops to about 65 mg. Commercial fried mixtures are a different order of magnitude and the reason ICMR-NIN's 2024 guidance names salted snacks specifically. Cholesterol: 2 g of saturated fat a portion, all from the coconut chips. Leave them out and it falls to under 0.5 g, at a real cost to the flavour.

SWAP IT

• White chickpeas for black → Consequence: softer, milder, and fibre falls from 8 g to about 5 g a portion.
• Coconut oil for sesame → Consequence: the tea-shop smell, and saturated fat up from 2 g to about 3 g a portion.
• Add 40 g of roasted peanuts after cooling → Consequence: better as a mixture, more monounsaturated fat, and about 25 kcal more per portion. (Allergen: this swap introduces peanuts.)
• Fried, as the tea shop makes it → Consequence: roughly 30–40 g of absorbed oil across the batch, so about 45 kcal and 4 g more fat a portion, and a crunch the oven does not quite reach.
• Skip the M06 and use 1 g more salt → Consequence: simpler, and sodium rises from 120 mg to about 175 mg a portion for no gain in flavour.

SCALE IT

At 400 g use two trays and swap shelves halfway, or the lower tray stays leathery. Oil does not scale linearly; a fuller tray needs proportionally less.

MAKE AHEAD · STORE · FREEZE Cupboard 2 weeks in a genuinely airtight jar, and they soften noticeably by week three in a humid kitchen. Fridge 6 weeks — the better option in Kerala, or anywhere above 25°C (77°F). Freeze 4 months; they thaw crisp in 10 minutes on the counter and this is the best way to keep a large batch. Refresh any batch that has gone soft in a 160°C (325°F) oven for 6–8 minutes and cool completely.

LEFTOVERS Crush a handful coarsely and scatter it over R089 or R099 for texture, or over a thoran off the heat. Whole, they go into a lunchbox better than anything else in this chapter.

SERVE WITH R106 Chai · R107 Sambharam · R099 Carrot and Coconut Raita · R100 Baked Ethakka Appam, if one sweet thing is wanted alongside

TROUBLESHOOTING Leathery rather than crisp: not dried after soaking, or pulled from the oven too early. Return them for 8–10 minutes. Burnt coconut and leaves: they went in at the start rather than in the last 5–6 minutes. Soft after three days: the jar is not airtight, or they were jarred warm. Refresh in the oven and cool completely. Hard as gravel: over-roasted by 10 minutes or more, or the chickpeas were old. Old pulses never soften properly, in a curry or here.

NOTES

R106 · Chai, and the Sugar Arithmetic [LOW SODIUM] [VEGETARIAN] [UNDER 250 KCAL]

A proper Kerala chaya, and the number nobody adds up

ICONS: Q · L

THE RE-ENGINEERING: The method is untouched — this is how a Kerala tea shop makes tea, and it should be. Two quantities change, and it would be dishonest to name only one of them. The sugar falls from two level teaspoons to one, arrived at over about twelve weeks rather than overnight; and the milk falls from the three-quarters of the glass a tea shop pours to one-third, which is where most of the energy difference actually comes from. The staged sugar plan and the arithmetic of both changes are set out below.

Kerala's chaya is not Indian masala chai and it is not English tea. Tea dust is boiled hard in water with a little crushed ginger, milk goes in, the whole thing is brought up once, and then it is poured from height between two vessels — adichu, "beaten" — until it foams. The pouring is not showmanship: it aerates and cools the tea, and it is why tea-shop chaya has a texture nothing made in a cup has. Learn the method first. The sugar is a separate conversation, and it is the one that matters.

Serves 2 · about 230 ml per glass Active 8 minutes · Total 8 minutes Difficulty: Easy

EQUIPMENT: Small deep pan; fine strainer; two vessels tall enough to pour between — a metal tumbler and a jug will do. Substitute: pour between two mugs from about 30 cm and accept less foam. Do not skip the pouring altogether; it changes the drink.

BEFORE YOU START: Have the strainer, the glasses and the pouring vessels out before the milk goes in. From the moment the milk is in, the tea has about 90 seconds before it boils over or stews.

INGREDIENTS

• 300 ml water
• 6 g (2 tsp) black tea dust, or strong CTC leaf
• 4 g (a 1 cm piece) fresh ginger, crushed
• 1 green cardamom pod, crushed, optional
• 150 ml whole milk
• 8 g (2 level tsp) sugar — see the arithmetic and the plan below

SOURCING NOTE: Kerala tea shops use tea dust, the finest grade, because it extracts fast and hard; a good CTC leaf from Munnar, the Nilgiris or Assam is the closest substitute and needs 30 seconds longer. Orthodox leaf teas and Darjeeling do not work — they are too delicate and go bitter under this treatment. Whole milk is what makes the foam; skimmed milk will not hold it.

ALLERGENS: Milk.

METHOD

1. Boil the water with the ginger and cardamom in the pan for 1–2 minutes, until the water smells clearly of ginger.
2. Add the tea dust and boil hard for 1½–2 minutes, until the liquid has gone deep mahogany and the surface carries a fine tan foam. Kerala chaya wants the tea boiled, not steeped.
3. Pour in the milk and bring it back just to the point where the surface swells and lifts, 60–90 seconds. Take it off the heat there. Milk boiled beyond this point tastes cooked and the tea turns dull brown.
4. Strain into a jug, pressing the leaves once with the back of a spoon.
5. Sweeten in the jug, not in the glass — the sugar dissolves completely and you can see how much went in.
6. Pour the tea between the two vessels from as high as you can manage without splashing, six to eight times, until a good head of foam stands on the surface. It should be hot but drinkable.

WHY THIS WORKS

Boiling extracts tannins and theaflavins fast, which is why chaya tastes strong at a modest quantity of leaf, and the milk proteins then bind some of those tannins so the astringency reads as body rather than bitterness. Pouring from height whips air into those same proteins, producing the foam. It also drops the temperature by several degrees, which is why tea-shop chaya can be drunk immediately.

PER SERVING, one glass with 1 level teaspoon of sugar (estimated)

THE SAME GLASS AT FOUR SUGAR LEVELS (estimated)

*Milk computed from IFCT L002, whole cow's whole milk at 73 kcal per 100 ml; tea, ginger and cardamom from the substitutions in the chapter panel note; sugar as sucrose at 400 kcal and 100 g sugars per 100 g. A level teaspoon of granulated sugar is taken as 4 g.

The tea-shop glass and this one, on the same assumptions. A tea shop's glass is 150 ml and about three-quarters of it is milk; on the figures above that glass comes to about 119 kcal at two level teaspoons of sugar, which is 79 kcal per 100 ml. This recipe's glass is 230 ml and about one-third of it is milk, so at the same two teaspoons it is 100 kcal, or 43 kcal per 100 ml, and at one teaspoon 84 kcal, or 37 per 100 ml. The recipe's glass is half again as large and still lighter, and the reason is the milk rather than the sugar. Both changes are real, and they do different jobs: the milk step does most of the arithmetic in one move, and the sugar step is the one that can be staged so slowly that nobody at the table notices.

THE ARITHMETIC, DONE OUT LOUD

Two level teaspoons of sugar is 8 g. Four glasses a day — which is not an extreme figure in Kerala; it is a normal one, and many households run higher — is 32 g of sugar a day. That is:

Eleven and a half kilograms of sugar a year, in tea, from a habit nobody describes as eating sugar. Go from two teaspoons to one and you remove 5.8 kg of sugar and 23,360 kcal a year — about twelve days' worth of a 2,000 kcal diet — and you have not given anything up except half a spoon.

The company that sugar keeps matters too. In the Nurses' Health Study II, women who increased sugar-sweetened soft drink intake from a weekly to a daily habit gained 4.69 kg over four years against 1.34 kg in those who cut back, and those drinking one or more a day had a relative risk of 1.83 for type 2 diabetes against under one a month; homemade sweetened fruit punch carried a relative risk of 2.00 (Schulze et al., 2004). The relevant point for this chapter is that last one. Sweetened drinks made at home are not exempt from the finding because they were made at home.

THE STAGED PLAN, AND WHY IT WORKS

Do not go from two teaspoons to none. It fails, reliably, and it fails for a reason that is not about willpower: an abrupt drop is detected, resented and reversed, while a small drop is not detected at all.

The best-documented analogue is salt. Salt-taste preference adapts over roughly eight to twelve weeks, which is why the sensory literature recommends cutting salt by 10–15% a month rather than in one step — at that rate the change goes unnoticed. Apply the same logic:

Three practical notes. Measure the sugar with a level teaspoon rather than pouring it — most people's pour is nearer three teaspoons than two, and the first week's saving is often invisible for that reason alone. Sweeten in the pan or the jug, so the whole household's tea steps down together rather than one person negotiating at the table. And make the last glass of the day the unsweetened one, because the evening cup is the one where the sugar is most habitual and least tasted.

On non-sugar sweeteners: the evidence base assembled for this book contains nothing on them. This book therefore makes no recommendation about them in either direction, and readers should not read that silence as either endorsement or warning.

IF YOU ARE WATCHING… Blood sugar: the sugar in tea is free sugar in liquid form, arriving without fibre, fat or protein to slow it, four times a day. Of everything in this book, it is the cheapest thing to change and the item where a reader watching post-meal glucose will see a difference soonest. Blood pressure: 30 mg of sodium a glass, from the milk. Not an issue. Cholesterol: 2 g of saturated fat a glass from whole milk, and 8 g a day across four glasses, which is worth knowing. Semi-skimmed milk halves it and costs you most of the foam; the fermented-dairy evidence does not resolve the full-fat question in either direction, so this is a straightforward choice about the fat, not about the milk.

SWAP IT

• Sulaimani — black tea with lime, no milk, the Malabar after-feast tea → Consequence: about 5 kcal a glass unsweetened, no saturated fat, and it is genuinely a different and very good drink. The single largest change available in this chapter.
• Semi-skimmed milk → saturated fat falls from about 2 g to 1 g a glass. Consequence: a thinner drink that will not hold foam.
• Jaggery for the sugar → Consequence: almost none that matters. Jaggery is unrefined sugar and behaves like sugar; what it gives is a better flavour, and better-tasting sugar is a real advantage only if it lets you use less. If it does not, this is not a change.
• Extra ginger and cardamom, sugar down a full teaspoon in one step → Consequence: aroma partly covers the missing sweetness, and this is the one place where a faster cut sometimes holds.

SCALE IT

Tea does not scale linearly — a larger pot extracts more efficiently, so use proportionally less leaf and taste at 90 seconds.

MAKE AHEAD · STORE · FREEZE Chaya is a made-to-order drink and does not keep; reheated tea is stewed tea. Fridge 1 day as the base of a cold tea, strained and unsweetened. Freeze — no. What you can make ahead is the ginger-cardamom water: boil 15 g of crushed ginger and 4 cardamom pods in 600 ml of water for 5 minutes, strain, and keep it in the fridge 3 days or freeze 2 months in an ice-cube tray, which takes the tea from eight minutes to four.

LEFTOVERS Cold strained chaya, unsweetened, over ice with a squeeze of lime is a better iced tea than most. Two tablespoons of strong strained tea also darkens a rice-flour batter usefully.

SERVE WITH R105 Roasted Kadala Mix · R104 Savoury Kozhukkatta · R100 Baked Ethakka Appam · R111 Sulaimani, which takes this further

TROUBLESHOOTING Bitter and flat: the tea boiled after the milk went in. Take it off the moment the surface lifts. No foam: skimmed milk, or the pour was too short. Pour from higher and more times. Weak: orthodox leaf instead of dust, or not enough boiling at step 2. Skin on the surface: the milk was boiled. Strain it and pour again; next time watch step 3. Tastes wrong at one teaspoon: you stepped down too fast. Go back half a step and hold it for three weeks. The plan works at the pace it is written.

NOTES

Four o'clock, rebuilt. Seven preparations, and the argument behind them is the arithmetic at the top of this chapter: a fried snack and a sweet glass of tea, most days, is the largest invisible load in Kerala eating, and nobody counts it because nobody calls it a meal. This chapter has not abolished four o'clock. It has moved the frying into a steamer or an oven, kept the shapes and the flavours that make the hour what it is, and shown the numbers both ways so that when you do fry a pazham pori — and you should, sometimes — you know what you are choosing. The habit to build is not abstinence; it is knowing the difference between the Tuesday version and the Sunday one. The chapter that follows deals with what goes into the glass, which is the other half of the same problem.