Chapter 4 — The Inner Garden

"The dependence of the intestinal microbes on the food makes it possible to adopt measures to modify the flora in our bodies and to replace the harmful microbes by useful microbes." — Élie Metchnikoff, The Prolongation of Life: Optimistic Studies, 1907

A naturalist who spends long enough at the table eventually notices that the interesting part of a meal begins exactly where the fork leaves off. Chew, swallow, and the food vanishes from view as thoroughly as a hare down a burrow; what happens to it thereafter is generally left to the imagination, or, more commonly, not thought of at all. This is a pity, because the nine metres of coiled and folded tubing running from lip to seat conceal one of the more extraordinary transformations in the natural world — a journey in which bread becomes blood, in which a stranger's meal of rice or lentils or salted fish is broken down, atom by atom, and rebuilt into the specific and idiosyncratic substance of the eater. It is also, as this chapter will argue, a journey undertaken in company. Somewhere around the midpoint of that nine-metre tube, human digestion quietly stops being a solitary act and becomes a joint venture with some tens of trillions of fellow-travellers who arrived, mostly, within the first few years of the eater's life and have been working the same shift ever since.

From mouth to colon, in nine metres or so

The tour begins, unglamorously, with spit. Saliva is not simply a lubricant; it carries an enzyme, salivary amylase, that begins dismantling starch into shorter sugars before the first mouthful is even swallowed, which is the reason a piece of plain bread, chewed patiently and without any accompaniment, will start to taste faintly sweet after perhaps thirty seconds — the tongue detecting glucose that a spoonful of saliva has already liberated from the loaf. Swallowing sends the resulting bolus down the oesophagus by a wave of muscular contraction, peristalsis, that would carry food to the stomach even if the eater were doing a handstand at the time; gravity, contrary to intuition, is barely required.

It was, fittingly enough for a volume dressed in this particular period's tweed and brass, an Edwardian preoccupation of quite serious proportions. Horace Fletcher, an American businessman turned self-appointed digestive reformer, arrived in the 1900s at the conviction that nearly every ailment of modern life traced back to inadequate chewing, and built around that conviction a movement — "Fletcherism" — that asked its adherents to masticate each mouthful some thirty or more times, often to the point of liquefaction, before permitting themselves to swallow it at all; John Harvey Kellogg, no stranger to dietary crusades of his own, coined the movement's name and gave it a platform at his Battle Creek Sanitarium, and for a few years fashionable dinner tables on both sides of the Atlantic fell into what one historian has called "the tense and awful silence" of guests privately counting their chews. Fletcherism as a cure-all has not survived contact with subsequent physiology any better than most single-cause theories of health do, but the underlying observation was not foolish: salivary amylase does need contact time to work, thorough chewing does measurably ease the stomach's subsequent labour, and a modern reader mourning the demise of the family meal, eaten in front of a screen at speed, might allow Fletcher his one sound instinct even while declining to count to thirty-two.

The stomach itself is less a bag than a churn, and a notably hostile one: its lining secretes hydrochloric acid strong enough to hold a resting pH of around 1.5 to 3.5, acidic enough to dissolve a nail over time, held safely away from the stomach's own tissue only by a thick coat of bicarbonate-rich mucus that is constantly renewed. Into that acid bath the enzyme pepsin gets to work on proteins, unfolding and cleaving them into shorter chains, while the stomach's muscular walls churn the whole mixture into a grey, semi-liquid porridge the physiologists call chyme — a word worth knowing if only because it captures, more honestly than any recipe ever will, roughly what has become of last night's dinner four hours on. The stomach empties itself gradually, a few millilitres of chyme squeezed at a time through the pylorus into the small intestine, over a period that may run from two hours for a light, watery meal to six or more for a heavy, fatty one; it is one of the underappreciated reasons a rich meal sits so long and so heavily.

The small intestine — some six or seven metres of it, coiled into a space barely larger than a rugby ball — is where the real chemistry, and the real absorption, happen. Bile from the liver, stored and concentrated in the gallbladder, arrives to emulsify fats into droplets small enough for enzymes to work on; the pancreas contributes a flood of its own enzymes — amylase, lipase, proteases — together with enough bicarbonate to neutralise the acid chyme has carried in from the stomach. What remains of protein is reduced to single amino acids and short peptides, fat to fatty acids and monoglycerides, starch to simple sugars, all of it ready to cross the intestinal wall into the bloodstream through a lining folded and refolded — into circular pleats, then into the fingerlike projections called villi, then into a still finer fringe of microvilli on each cell of each villus — to pack the largest possible absorptive surface into the smallest practicable space. Two independent Swedish physiologists, revisiting the textbook figures in 2014 with better microscopy and a healthy scepticism, found that the old, much-repeated claim — that this folded surface, laid flat, would cover a tennis court, some 180 to 300 square metres — had never rested on solid measurement at all; their own careful reckoning put the true figure at 30 to 40 square metres, roughly half a badminton court, or, as one of the authors preferred to put it, a decent-sized studio flat. It remains an impressively large area to fold into a torso, even shorn of its more heroic advertising.

By the time the residue reaches the ileocaecal valve, the border post between small intestine and large, the body has already extracted the overwhelming majority of the calories, vitamins and minerals it is going to get from the meal. What crosses that border is, from the human digestive system's own point of view, waste: water, electrolytes, and a quantity of carbohydrate — starches, gums, and the various plant fibres — that no human enzyme has the equipment to break down at all. It is at exactly this point, where the body's own chemistry runs out of ideas, that the second digestive system takes over.

The organ we did not know we had

That second system is the microbiome, and it is worth stating plainly, at the outset, just how large a fact about the eater's own body has gone almost entirely unremarked until quite recently. A careful 2016 recount by the biologists Ron Sender, Shai Fuchs and Ron Milo — replacing a figure of ten bacterial cells for every human cell that had circulated, more or less unexamined, since the 1970s — put the true ratio close to parity: something on the order of 38 trillion bacterial cells sharing quarters with roughly 30 trillion human ones, the bacterial population concentrated overwhelmingly in the colon and weighing, all told, somewhere in the neighbourhood of 200 grams — not much more than a chicken breast, and yet by number of genes it dwarfs its host almost beyond comparison. The human genome carries some twenty thousand protein-coding genes; the collective genome of the gut microbiome, the microbiologists' so-called "second genome," is estimated at several hundred times that number, encoding metabolic tricks — the digestion of otherwise indigestible plant fibre chief among them — that no vertebrate genome has ever bothered to evolve on its own account, because it never had to.

It is more useful, and more accurate, to think of this community not as a passenger load but as an organ in its own right: one the anatomist never drew because it does not have fixed borders or a single shape, but which behaves functionally like a gland, manufacturing compounds — vitamins, neurotransmitter precursors, immune-signalling molecules, and above all the short-chain fatty acids at the centre of this chapter — that circulate through the body and measurably influence tissues as distant as the liver, the brain and the immune system. It is populated chiefly by two great phyla, the Firmicutes and the Bacteroidetes, whose relative proportions have attracted a great deal of speculative headline-writing over the years, alongside a long tail of hundreds of further species whose exact identities differ, often substantially, from one healthy person to the next; there is, this compendium will note honestly, no single fixed blueprint for a "correct" microbiome, and a naturalist should be wary of anyone who claims to have found one. What the healthiest guts do reliably share is not a particular species list but a particular property: diversity, richness, and a functional capacity to do the one job that matters most to the argument of this chapter — fermentation.

It is also, not incidentally, an organ in close conversation with the immune system: some seventy to eighty per cent of the body's antibody-producing immune cells reside in tissue associated with the gut wall, a concentration that makes evolutionary sense the moment one recalls that the gut lining is, in effect, the single largest and busiest border crossing between the inside of the body and the outside world, patrolled accordingly. The conversation runs further than immunity alone. Short-chain fatty acids arriving from colonic fermentation dock onto a pair of receptors called FFAR2 and FFAR3, sitting on specialised gut cells known as L-cells, and their arrival prompts those cells to release two further hormones, GLP-1 and PYY, that travel on to signal the pancreas and the brain that a meal has, in effect, been dealt with and that appetite may reasonably subside; it is a satiety signal manufactured jointly by the eater's own gut and the eater's own bacteria, and it supplies one of the more concrete mechanisms behind an observation nutritionists have made for decades on much softer evidence — that a fibre-rich meal leaves a diner feeling fuller, for longer, than a calorically identical meal built on refined starch and sugar alone.

Field Note — The Villus

Examined under even a modest microscope, a single villus of the small intestine looks less like architecture than like something alive in its own right — a finger of tissue perhaps a millimetre tall, waving gently in the intestinal current, sheathed from base to tip in a single layer of absorptive cells, each of those cells in turn bristling with several thousand microvilli too fine for anything but an electron microscope to resolve individually, together forming what histologists call, not poetically but accurately, a brush border. Threaded up the villus's core runs a fine capillary network and a single blind-ended lymphatic vessel, the lacteal, named for the milky-white fluid — fat, chiefly, bundled for transport as chylomicrons — that fills it after a rich meal and gives the whole structure, at certain hours, a faintly pearlescent cast under the lens. The traffic here runs only one way and only briefly: an enterocyte on the surface of a villus is born in a nursery of rapidly dividing stem cells tucked into the pits between villi, migrates up the flank of its villus over three to five days performing its single life's work of absorption, and is then shed whole into the gut and passed on, to be replaced by the cell behind it in an orderly, unceasing procession — the entire lining of the small intestine renewing itself roughly once a week, making the gut's inner surface among the fastest-regenerating tissue anywhere in the body. Interspersed among the enterocytes, and easy to miss at low magnification, are goblet cells, so named for their cup-like shape, quietly extruding a coat of mucus that keeps the whole absorbing surface slick, defended, and just out of direct reach of whatever happens to be passing through — a specimen, in short, that manages to be both a marvel of surface engineering and a small, ceaseless act of renewal, all in a structure the width of an eyelash.

<figure> <svg xmlns="http://www.w3.org/2000/svg" width="100%" viewBox="0 0 460 680" role="img"> <title>A Villus in Section — epithelium, lacteal, capillaries and crypt</title> <defs> <pattern id="stipple7" width="6" height="6" patternUnits="userSpaceOnUse"> <circle cx="1.2" cy="1.2" r="0.5" fill="#a9906c"/> </pattern> <marker id="arw7" viewBox="0 0 10 10" refX="8" refY="5" markerWidth="6" markerHeight="6" orient="auto-start-reverse"> <path d="M0,0 L10,5 L0,10 Z" fill="#43331f"/> </marker> </defs> <rect x="8" y="8" width="444" height="664" fill="none" stroke="#43331f" stroke-width="1.2"/> <rect x="12" y="12" width="436" height="656" fill="none" stroke="#c8b79a" stroke-width="0.6"/> <g fill="none" stroke="#43331f" stroke-width="0.8"> <path d="M12,12 L12,20 M12,12 L20,12"/> <path d="M448,12 L448,20 M448,12 L440,12"/> <path d="M12,668 L12,660 M12,668 L20,668"/> <path d="M448,668 L448,660 M448,668 L440,668"/> </g> <text x="230" y="34" text-anchor="middle" font-family="Georgia, 'Times New Roman', serif" font-size="15" letter-spacing="2" fill="#43331f">A VILLUS IN SECTION</text> <line x1="40" y1="48" x2="420" y2="48" stroke="#c8b79a" stroke-width="0.6"/>

<path d="M170,540 L170,240 Q170,180 230,175 Q290,180 290,240 L290,540" fill="none" stroke="#43331f" stroke-width="1"/>

<path d="M170,235 L158,235 M170,265 L158,265 M170,295 L158,295 M170,325 L158,325 M170,355 L158,355 M170,385 L158,385 M170,415 L158,415 M170,445 L158,445 M170,475 L158,475 M170,505 L158,505 M290,235 L302,235 M290,265 L302,265 M290,295 L302,295 M290,325 L302,325 M290,355 L302,355 M290,385 L302,385 M290,415 L302,415 M290,445 L302,445 M290,475 L302,475 M290,505 L302,505 M190,195 L182,185 M206,180 L200,169 M222,175 L221,163 M238,175 L239,163 M254,180 L260,169 M270,195 L278,185" stroke="#6f5a40" stroke-width="0.6" fill="none"/>

<g fill="none" stroke="#43331f" stroke-width="0.6"> <rect x="173" y="236" width="11" height="28"/><rect x="173" y="268" width="11" height="28"/><rect x="173" y="300" width="11" height="28"/><rect x="173" y="332" width="11" height="28"/><rect x="173" y="364" width="11" height="28"/><rect x="173" y="396" width="11" height="28"/><rect x="173" y="428" width="11" height="28"/><rect x="173" y="460" width="11" height="28"/><rect x="173" y="492" width="11" height="28"/> <rect x="276" y="236" width="11" height="28"/><rect x="276" y="268" width="11" height="28"/><rect x="276" y="300" width="11" height="28"/><rect x="276" y="332" width="11" height="28"/><rect x="276" y="364" width="11" height="28"/><rect x="276" y="396" width="11" height="28"/><rect x="276" y="428" width="11" height="28"/><rect x="276" y="460" width="11" height="28"/><rect x="276" y="492" width="11" height="28"/> </g>

<path d="M215,250 Q230,232 245,250 L245,470 Q230,486 215,470 Z" fill="url(#stipple7)" stroke="#43331f" stroke-width="0.9"/>

<g fill="none" stroke="#9c5a2c" stroke-width="0.7"> <path d="M198,470 Q182,410 195,350 Q205,300 198,255"/> <path d="M262,470 Q278,410 265,350 Q255,300 262,255"/> </g> <g fill="#9c5a2c" stroke="none"> <circle cx="194" cy="430" r="2"/><circle cx="192" cy="380" r="2"/><circle cx="198" cy="330" r="2"/><circle cx="200" cy="280" r="2"/> <circle cx="266" cy="430" r="2"/><circle cx="269" cy="380" r="2"/><circle cx="262" cy="330" r="2"/><circle cx="260" cy="280" r="2"/> </g>

<path d="M170,540 Q195,560 195,595 Q195,620 230,620 Q265,620 265,595 Q265,560 290,540" fill="none" stroke="#43331f" stroke-width="1"/> <g fill="none" stroke="#6f5a40" stroke-width="0.6"> <ellipse cx="205" cy="580" rx="7" ry="10"/><ellipse cx="222" cy="600" rx="7" ry="10"/><ellipse cx="238" cy="600" rx="7" ry="10"/><ellipse cx="255" cy="580" rx="7" ry="10"/> </g>

<g stroke="#43331f" stroke-width="0.8" fill="none"> <path d="M150,560 L150,270" marker-end="url(#arw7)"/> <path d="M310,560 L310,270" marker-end="url(#arw7)"/> </g>

<g font-family="Georgia, serif" font-size="10" fill="#43331f"> <circle cx="330" cy="185" r="7" fill="none" stroke="#43331f" stroke-width="0.7"/><text x="330" y="188" text-anchor="middle" font-size="8">1</text> <circle cx="330" cy="255" r="7" fill="none" stroke="#43331f" stroke-width="0.7"/><text x="330" y="258" text-anchor="middle" font-size="8">2</text> <circle cx="330" cy="330" r="7" fill="none" stroke="#43331f" stroke-width="0.7"/><text x="330" y="333" text-anchor="middle" font-size="8">3</text> <circle cx="330" cy="400" r="7" fill="none" stroke="#43331f" stroke-width="0.7"/><text x="330" y="403" text-anchor="middle" font-size="8">4</text> <circle cx="330" cy="590" r="7" fill="none" stroke="#43331f" stroke-width="0.7"/><text x="330" y="593" text-anchor="middle" font-size="8">5</text> <circle cx="330" cy="470" r="7" fill="none" stroke="#43331f" stroke-width="0.7"/><text x="330" y="473" text-anchor="middle" font-size="8">6</text> </g>

<g font-family="Georgia, serif" font-style="italic" font-size="10.5" fill="#43331f"> <text x="345" y="640">1 brush border, or microvilli</text> </g> <g font-family="Georgia, serif" font-style="italic" font-size="10" fill="#43331f"> <text x="40" y="645">2 columnar epithelium    3 central lacteal    4 capillary network</text> <text x="40" y="660">5 crypt of Lieberkühn    6 migrating enterocyte, crypt to tip</text> </g> </svg> <figcaption><em>Plate 7. A single intestinal villus in section: the absorptive brush border, the central lacteal, and the ceaseless procession of enterocytes migrating from crypt to tip</em></figcaption> </figure>

The gut microbiome is also, it is worth adding honestly, a good deal more fragile than the reassuring word "ecosystem" tends to suggest. A single course of broad-spectrum antibiotics — necessary, often life-saving medicine, and nothing this compendium would ever counsel a reader against taking when a physician prescribes it — can measurably reduce microbial diversity within days, and while most healthy adult guts substantially recover within weeks or months, certain species have been shown, in some studies, not to return to their prior abundance even a year later. The same modern conveniences that have lengthened lifespans and controlled infectious disease — antibiotics, caesarean birth, formula feeding, chlorinated water, the general retreat of the farmyard from daily life — appear, taken together, to have narrowed the diversity of the microbiome inherited by children born into industrialised nations, when set against the microbiomes documented in populations still living by traditional, non-industrialised subsistence, such as the Hadza of Tanzania, whose gut communities carry a breadth of species that has simply gone missing from most Western guts. None of this is an argument against modern medicine, which this compendium has no interest in relitigating; it is, rather, one further reason the deliberate, dietary rebuilding of microbial diversity — through fibre, through fermented foods, through the sheer variety urged in this chapter's closing section — has become a live and serious question for nutritional science, rather than a fashionable one.

Fibre, and the fermentation in the dark

Fibre is, chemically speaking, an unglamorous category: it is simply the collection of plant carbohydrates — cellulose, hemicellulose, pectin, resistant starch, various gums and inulin-type fructans among them — that pass the small intestine's enzymes without a second glance, for the perfectly good reason that no human digestive enzyme has ever evolved to cleave the particular chemical bonds holding them together. For the better part of the twentieth century this indigestibility was treated, not unreasonably on the available evidence, as fibre's whole story: a kind of harmless roughage, valued chiefly for adding bulk and encouraging things along. It has taken the microbiome revolution of the past two decades to reveal that this was rather like praising a symphony for keeping the orchestra pit occupied. What human enzymes cannot touch, the resident microbiota can, and does, with considerable relish.

Fibre that survives to the colon meets there an anaerobic fermentation vat rivalling anything a brewer might build, staffed by trillions of bacteria for whom a given fibre molecule is not waste at all but the single most valuable meal of the day. The principal products of that fermentation are a family of short-chain fatty acids — acetate, propionate and, most consequential of the three, butyrate — and it is worth pausing on butyrate specifically, because its story reframes the entire relationship between colon and colonist. Long before the modern microbiome was mapped in any detail, the surgeon William Roediger, working in the early 1980s, isolated cells from the human colon lining and simply asked them what they preferred to burn for fuel; the answer, confirmed many times since, was overwhelmingly butyrate, which the colon's own lining cells — the colonocytes — take up and oxidise in preference to glucose, deriving from it, by some estimates, over seventy per cent of their day-to-day energy. Put plainly: the cells lining the large intestine run, in large part, not on sugar carried in from the bloodstream but on a fatty acid manufactured, on the spot, by bacteria digesting fibre the eater could not digest alone. It is a strange and rather moving piece of physiology once it is stated outright — a wall of the body kept alive chiefly by feeding its own tenants.

Butyrate's usefulness does not stop at the colon wall. It crosses into circulation and acts, among other things, as a signal to the immune system: laboratory work published in 2013 by two independent teams, one led by Koji Furusawa and Hidehiro Fukuda's Japanese collaborators, the other by the American immunologist Nicholas Arpaia, showed that butyrate promotes the development of regulatory T cells, a class of immune cell whose particular job is calming inflammation rather than provoking it — a molecular mechanism that goes some way towards explaining a broader epidemiological pattern in which populations eating fibre-poor, heavily processed diets show a higher background rate of chronic, low-grade inflammatory disease. Nor is butyrate's protection confined to inflammation in the abstract. The gut's mucus lining — the same slick coat the goblet cells of the previous section were seen quietly secreting — turns out to be more than a passive shield; a 2016 study from the laboratory of Eric Martens and Mahesh Desai, working with mice engineered to carry a defined human-derived microbiota, found that when dietary fibre was withheld, the resident bacteria did not simply go hungry and wait: lacking their preferred fuel, several species switched to eating the mucus layer itself, thinning it measurably and leaving the underlying gut wall more exposed to infection by a pathogen deliberately introduced into the animals' guts. It is a mouse experiment, and this compendium will not pretend it proves the identical mechanism operates hour by hour in every underfed human colon; but it dovetails uncomfortably well with a large and independent body of human epidemiology linking low-fibre, industrialised diets to a raised background rate of gut inflammation, and it supplies, for the first time, a plausible mechanical story for why that might be so — the microbiota, denied its usual meal, quietly eating the house instead.

Justin and Erica Sonnenburg's laboratory at Stanford supplied the other half of this cautionary tale in 2016, by breeding mice on a low-fibre diet across several successive generations and watching, with each generation, which bacterial species disappeared from the gut and, crucially, whether restoring fibre to the diet afterward brought them back. It did not, reliably: a number of species, once lost from a lineage, could not be recovered by diet alone, and required a fresh transplant of the missing microbes to reappear at all — the mice, in effect, having permanently narrowed their own inherited gut ecosystem across a few fibre-poor generations. The finding travelled with an obvious and uncomfortable question attached, one the Sonnenburgs themselves have pressed in their public writing since: given that fibre intake has fallen sharply across industrialised populations within living memory — the average adult in Britain or the United States eats something on the order of fifteen grams of fibre a day, against a recommended twenty-five to thirty-eight — has the modern gut already lost microbial residents that a hunter-gatherer's gut, or a great-grandparent's, still carried, and lost them for good? It is an open question rather than a settled one, and this compendium will not pretend otherwise, but it is exactly the kind of open question that ought to make the fibre gap sound less like a minor dietary shortfall and rather more like a slow-motion extinction event, conducted one skipped vegetable at a time.

The fermented table

If fibre feeds the microbes already resident in the gut, a second and older culinary tradition supplies live microbes directly, by way of the fermented foods that stand among the oldest deliberate biotechnologies our species possesses, discovered independently and repeatedly, wherever people needed to preserve a harvest against the lean months and stumbled, by trial and considerable error, onto the same underlying trick: introduce the right microbe, deny the wrong ones their usual advantages — typically by salt, acid or the exclusion of air — and let fermentation do the rest.

Yoghurt is the simplest specimen and the oldest attested, its living culture built classically on two bacterial species working together, Lactobacillus delbrueckii subsp. bulgaricus and Streptococcus thermophilus, each producing lactic acid that curdles the milk's proteins and, not incidentally, lowers the pH to a point most spoilage organisms cannot tolerate. Kimchi and sauerkraut perform an almost identical trick on cabbage rather than milk, relying on the same broad family of lactic-acid bacteria — Leuconostoc and various species of what taxonomists now group under Lactiplantibacillus — that arrive not from any added starter culture at all but simply from the surface of the vegetable and the hands and utensils that prepared it, submerged in a brine strong enough to favour the salt-tolerant lactic fermenters over their less hardy rivals. Miso and soy sauce take a third route entirely, relying on a cultivated mould, koji (Aspergillus oryzae), to break soybean protein down into free amino acids over months of slow fermentation, a process this compendium's earlier volume already met, from the flavour side of the ledger, in its account of umami; here the same fermentation is worth revisiting from the physiological side, since a spoonful of a genuinely long-fermented miso or a splash of naturally brewed soy sauce carries not only glutamate but a modest, still-living population of the bacteria and yeasts that made it.

Kefir and kombucha extend the same principle to a still more crowded ferment. Milk kefir is cultured not by one or two bacterial species but by a whole small consortium bound together in visible, cauliflower-like grains — Lactobacillus, Lactococcus and Leuconostoc species working alongside fermenting yeasts such as Kluyveromyces, the yeasts contributing a faint fizz and a trace of alcohol the bacteria alone would never produce. Kombucha performs an analogous trick on sweetened tea, its SCOBY — the acronym stands, with more accuracy than romance, for symbiotic culture of bacteria and yeast — combining acetic-acid bacteria such as Acetobacter and Gluconobacter with their own yeast partners in a floating cellulose mat that any Edwardian naturalist would have been delighted to sketch. In both cases the diversity of the ferment, rather than any single starring organism, appears to be the point: a wide guild of species, cooperating and competing in the same jar, in miniature rehearsal of the diversity this chapter argues the gut itself wants.

It is worth pausing, too, on the vocabulary this whole family of foods has generated, since the terms are used loosely in the popular press and precisely in the scientific literature, and the compendium prefers precision. A probiotic, by the definition jointly adopted by the Food and Agriculture Organization and the World Health Organization, is a live microorganism which, administered in adequate amounts, confers a measurable benefit on the host — a fairly high bar, met by particular studied strains rather than automatically by every spoonful of every fermented food. A prebiotic, by contrast, is not alive at all: it is a substance, fibre chief among them, that selectively feeds beneficial microbes already resident in the gut, doing its work at one remove. A postbiotic, the newest and least settled of the three terms, refers to the beneficial compounds — butyrate among them — that a live microbe leaves behind after fermentation has finished, whether or not the microbe itself survives to be eaten at all. Kimchi, sauerkraut and a live yoghurt sit at the generous, food-based end of this vocabulary, carrying probiotic, prebiotic and postbiotic elements all at once in a single, unglamorous jar; a supplement capsule, by contrast, generally isolates one term from the trinity and sells it alone, which is one more small instance of this compendium's governing rule that the whole food, more often than not, is doing quiet work no single extracted ingredient can fully replace.

What all of these foods share, and what a bottle of vinegar or a jar of shop-bought pickle preserved by pasteurisation and heat does not, is a population of live microorganisms still present at the moment of eating — which is the detail that has made fermented foods, rather than fibre alone, the subject of one of the more interesting microbiome trials of recent years. In 2021 a team led by Hannah Wastyk, working under the joint supervision of Christopher Gardner and the Sonnenburgs, put the two strategies to a direct ten-week test in healthy adults: one group ate a diet deliberately engineered to be very high in fibre, the other a diet rich in fermented foods — yoghurt, kefir, fermented cottage cheese, kimchi, sauerkraut, vegetable brines and kombucha among them. The fermented-food group showed a measurable rise in overall microbial diversity and a broad, significant fall in nineteen separate markers of inflammation circulating in the blood; the high-fibre group, over the same ten weeks, showed neither effect reliably, and among participants who began the trial with a less diverse gut, the fibre diet was associated with a slight rise in some inflammatory markers rather than a fall. It is a genuinely surprising result set against the tidy, single-cause story fibre advocates might have hoped for, and this compendium reports it exactly as its authors did: not as evidence that fibre is unimportant — the case for fibre rests on a much longer and broader body of evidence than one ten-week trial could ever overturn — but as a reminder that fermented foods and fibre-rich foods appear to work on the gut by genuinely different mechanisms, one supplying new residents, the other feeding the residents already there, and that a diet built on either alone is likely doing only half the available work.

<figure> <svg xmlns="http://www.w3.org/2000/svg" width="100%" viewBox="0 0 640 420" role="img"> <title>The Fermenting Garden — colonic bacteria producing short-chain fatty acids</title> <defs> <pattern id="hatch8" width="5" height="5" patternUnits="userSpaceOnUse" patternTransform="rotate(45)"> <line x1="0" y1="0" x2="0" y2="5" stroke="#6f5a40" stroke-width="0.5"/> </pattern> <pattern id="stipple8" width="6" height="6" patternUnits="userSpaceOnUse"> <circle cx="1.2" cy="1.2" r="0.5" fill="#6f5a40"/> </pattern> <marker id="arw8" viewBox="0 0 10 10" refX="8" refY="5" markerWidth="6" markerHeight="6" orient="auto-start-reverse"> <path d="M0,0 L10,5 L0,10 Z" fill="#43331f"/> </marker> <marker id="arw8b" viewBox="0 0 10 10" refX="8" refY="5" markerWidth="6" markerHeight="6" orient="auto-start-reverse"> <path d="M0,0 L10,5 L0,10 Z" fill="#9c5a2c"/> </marker> </defs> <rect x="8" y="8" width="624" height="404" fill="none" stroke="#43331f" stroke-width="1.2"/> <rect x="12" y="12" width="616" height="396" fill="none" stroke="#c8b79a" stroke-width="0.6"/> <g fill="none" stroke="#43331f" stroke-width="0.8"> <path d="M12,12 L12,20 M12,12 L20,12"/> <path d="M628,12 L628,20 M628,12 L620,12"/> <path d="M12,408 L12,400 M12,408 L20,408"/> <path d="M628,408 L628,400 M628,408 L620,408"/> </g> <text x="320" y="32" text-anchor="middle" font-family="Georgia, 'Times New Roman', serif" font-size="15" letter-spacing="2" fill="#43331f">THE FERMENTING GARDEN</text> <line x1="40" y1="46" x2="600" y2="46" stroke="#c8b79a" stroke-width="0.6"/> <text x="320" y="62" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#6f5a40">a section of the colon, where bacteria complete the meal</text>

<path d="M50,120 Q320,105 590,120 L590,168 Q320,150 50,168 Z" fill="url(#hatch8)" stroke="#43331f" stroke-width="1"/>  <path d="M50,292 Q320,310 590,292 L590,340 Q320,355 50,340 Z" fill="url(#hatch8)" stroke="#43331f" stroke-width="1"/>

<g fill="none" stroke="#43331f" stroke-width="0.5"> <rect x="70" y="152" width="14" height="14"/><rect x="100" y="150" width="14" height="14"/><rect x="130" y="149" width="14" height="14"/><rect x="160" y="148" width="14" height="14"/><rect x="480" y="148" width="14" height="14"/><rect x="510" y="149" width="14" height="14"/><rect x="540" y="150" width="14" height="14"/><rect x="565" y="152" width="14" height="14"/> <rect x="70" y="292" width="14" height="14"/><rect x="100" y="293" width="14" height="14"/><rect x="130" y="294" width="14" height="14"/><rect x="160" y="295" width="14" height="14"/><rect x="480" y="295" width="14" height="14"/><rect x="510" y="294" width="14" height="14"/><rect x="540" y="293" width="14" height="14"/><rect x="565" y="292" width="14" height="14"/> </g>

<line x1="18" y1="230" x2="95" y2="230" stroke="#43331f" stroke-width="1" marker-end="url(#arw8)"/> <text x="55" y="215" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10.5" fill="#43331f">dietary fibre</text> <g fill="url(#hatch8)" stroke="#6f5a40" stroke-width="0.6"> <path d="M105,215 L120,208 L128,220 L118,232 L102,228 Z"/> <path d="M100,250 L115,244 L122,258 L108,266 Z"/> </g>

<g fill="url(#stipple8)" stroke="#6f5a40" stroke-width="0.5"> <ellipse cx="200" cy="200" rx="10" ry="4" transform="rotate(20 200 200)"/> <ellipse cx="230" cy="185" rx="10" ry="4" transform="rotate(-15 230 185)"/> <ellipse cx="190" cy="240" rx="10" ry="4" transform="rotate(-25 190 240)"/> <ellipse cx="235" cy="255" rx="10" ry="4" transform="rotate(10 235 255)"/> <ellipse cx="260" cy="215" rx="10" ry="4" transform="rotate(35 260 215)"/> <ellipse cx="215" cy="220" rx="10" ry="4" transform="rotate(-5 215 220)"/> </g> <line x1="225" y1="270" x2="225" y2="345" stroke="#a9906c" stroke-width="0.5"/> <text x="225" y="358" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10.5" fill="#43331f">gut bacteria, the microbiota</text>

<text x="400" y="95" text-anchor="middle" font-family="Georgia, serif" font-size="11" letter-spacing="1" fill="#43331f">SHORT-CHAIN FATTY ACIDS</text> <line x1="330" y1="102" x2="470" y2="102" stroke="#c8b79a" stroke-width="0.5"/>

<line x1="270" y1="205" x2="365" y2="175" stroke="#43331f" stroke-width="0.8" marker-end="url(#arw8)"/> <polygon points="365,168 378,168 385,175 378,182 365,182 358,175" fill="none" stroke="#9c5a2c" stroke-width="0.8"/> <text x="371" y="150" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#43331f">acetate</text>

<line x1="275" y1="225" x2="365" y2="228" stroke="#43331f" stroke-width="0.8" marker-end="url(#arw8)"/> <polygon points="378,213 391,213 398,228 391,243 378,243 371,228" fill="none" stroke="#9c5a2c" stroke-width="0.8"/> <text x="384" y="260" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#43331f">propionate</text>

<line x1="270" y1="248" x2="365" y2="278" stroke="#43331f" stroke-width="0.8" marker-end="url(#arw8)"/> <polygon points="365,271 378,271 385,278 378,285 365,285 358,278" fill="none" stroke="#9c5a2c" stroke-width="1"/> <text x="371" y="303" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#43331f">butyrate</text>

<line x1="385" y1="278" x2="470" y2="300" stroke="#9c5a2c" stroke-width="1.1" marker-end="url(#arw8b)"/> <text x="500" y="325" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10.5" fill="#9c5a2c">chief fuel of the colon wall</text> </svg> <figcaption><em>Plate 8. The gut as a fermenting garden: fibre entering from the small intestine, colonic bacteria at work, and the short-chain fatty acids — acetate, propionate, butyrate — that result</em></figcaption> </figure>

Metchnikoff's wager

No account of the fermented gut is complete without a visit to the Russian zoologist whose words open this chapter, and this compendium's own rule of honest history is worth restating before the visit begins: report what a pioneer got right, and just as plainly what he got wrong, without letting either fact erase the other.

Élie Metchnikoff earned his Nobel Prize in 1908 not for anything to do with yoghurt but for phagocytosis, his earlier discovery that certain white blood cells actively engulf and destroy invading microbes — a founding insight of cellular immunology, and by any measure a towering piece of nineteenth-century biology. It was only in his later years, as director of research at the Pasteur Institute and increasingly preoccupied with his own advancing age, that Metchnikoff turned his formidable attention to the colon, and arrived at conclusions considerably less sound than his earlier work. He had come to believe, on evidence that was suggestive at best, that the large intestine was — in his own memorable and rather ungenerous phrase — a "vestigial cesspool," a reservoir of little use beyond harbouring bacteria whose fermentation and putrefaction released toxic compounds, autotoxins in his terminology, into the bloodstream, there to poison tissue after tissue and hasten the degenerations of old age. His remedy, elaborated at length in The Prolongation of Life in 1907, was to crowd out these harmful "putrefactive" bacteria with a daily dose of lactic-acid bacteria, delivered as sour milk fermented with what he called the Bulgarian bacillus — a microbe first isolated in 1905 by a young Bulgarian physician, Stamen Grigorov, from the traditional yoghurt of the Balkan peasantry, whose apparent longevity had caught Metchnikoff's attention in the first place. Metchnikoff practised what he preached, drinking soured milk daily and pronouncing himself, in his own words, "very well pleased with the result"; the fashion he set off was immediate and considerable, with Parisian newspapers running headlines asking whether old age could now be cured, and advertisements inviting the public to sample the same curdled milk "the illustrious Professor Metchnikoff has recommended for suppressing the disastrous effects of old age."

The honest verdict, delivered by the better part of a century of subsequent microbiology, is mixed in exactly the way real scientific history usually is. Metchnikoff's central mechanism has not held up in its particulars: subsequent laboratory work, as early as the 1920s, established that Lactobacillus bulgaricus does not, in fact, survive passage through the human gut in any lasting way, and cannot itself take up permanent residence and displace the resident colonic flora as he had hoped; nor has "autointoxication" by putrefactive bacteria ever been established as a driver of ageing in anything like the sweeping terms Metchnikoff proposed, and he himself, notwithstanding his daily sour milk, died at seventy-one of heart failure, some way short of the hundred-year lifespan he considered humanity's natural birthright. And yet the broad intuition underneath the mechanism — that the composition of the gut's bacterial population is neither fixed nor irrelevant, that it responds to diet, and that deliberately introducing beneficial live microbes by way of fermented food can measurably shift that population and its downstream effects on the body — has been vindicated so thoroughly by modern microbiome science that Metchnikoff is now routinely, and fairly, credited as probiotic science's founding father, a full century before the word "probiotic" itself came into use. He guessed the destination without ever seeing the map; a fair judgment on a career that had already, decades earlier, given the world one of its accurate answers.

Diversity as the practical lesson

Two large and largely independent lines of modern evidence converge, rather satisfyingly, on the same practical rule of thumb, and it is worth stating that rule before this chapter closes, since it is the one piece of microbiome science sturdy enough to survive the coming decades of finer detail. The American Gut Project, a crowd-sourced study of the gut microbiomes of more than ten thousand volunteers across the United States, the United Kingdom and Australia, published its central finding in 2018: gut microbial diversity tracked far more closely with the number of distinct plant species a person ate across a given week than with any dietary label — vegan, vegetarian, omnivore — attached to their diet, with participants who reported thirty or more distinct plants a week showing meaningfully richer, more diverse gut communities, and fewer antibiotic-resistance genes among their resident bacteria, than participants eating ten or fewer. Different vegetables, grains, legumes, herbs, nuts and fruits carry different fibres in different structural forms, and it appears that a varied bacterial community wants, unsurprisingly, a varied menu, in much the same way a diverse ecosystem above ground supports a wider range of species than a monoculture ever could.

Nor need the thirty be exotic. Legumes alone — the lentils, chickpeas, black beans and split peas that appear so regularly in this compendium's earlier pages — carry a particular type of fibre, resistant starch and oligosaccharide, that colonic bacteria appear to relish disproportionately, so that a bean-rich week does a great deal of this chapter's work on its own; a scattering of different herbs counts toward the thirty as readily as a different vegetable does, and the American Gut Project's own investigators were at pains to note that variety, not exoticism or expense, was the operative variable. A naturalist keeping the kitchen larder rather than the specimen cabinet could do worse than treat the weekly shop as a small census-taking exercise — not "five a day" but something closer to thirty a week, counted generously across grains, pulses, vegetables, fruits, nuts, seeds and herbs alike.

Return, as this compendium's own governing rule insists a chapter on nutrition always must, to the plate rather than the pill. Butyrate itself can now be bought as a supplement, and glutamine, inulin and any number of isolated prebiotic fibres besides; none of them, taken in isolation, recreates what a varied week of whole vegetables, legumes, whole grains and fermented foods appears to accomplish together, because the benefit does not seem to reside in any single molecule so much as in the ecological complexity of the diet as a whole — the same lesson, differently dressed, that this volume's earlier chapters drew from the orange and the whole grain. A naturalist tending a garden does not sow a single miracle seed and call the plot complete; she plants widely, feeds the soil, and lets the resulting diversity do work no single species could manage alone. The gut, it turns out, wants tending on exactly the same principle.

The inner garden keeps its own clock, tuned less to the calendar than to the day — bacterial populations that wax and wane through the twenty-four hours much as the tides do, busiest fermenters at certain hours, quieter at others, in step with a rhythm the eater rarely notices consciously but obeys nonetheless every time hunger arrives, uninvited, at roughly the same hour it did yesterday. That the body keeps time at all, and that the timing of a meal might matter nearly as much as its contents, is where this compendium turns next.

บทถัดไป · 8 / 17Chapter 5 — The Chronology of Consumption