The Killer Tomato

Flavor: The Science of Our Most Neglected Sense - Bob Holmes 2017

The Killer Tomato

Near the southwest fringe of the University of Florida’s campus in Gainesville sits a nondescript single-story building of whitewashed brick. A long walk from the massive football stadium and the modern, glass-and-steel high rises of the medical center, the little building looks like it could house a maintenance shop for the campus groundskeepers, or perhaps a storage area for recyclable trash. But for anyone who loves the flavor of a good tomato—and who doesn’t?—this could be the most important building on campus.

The supermarket tomato is the poster child for the failure of modern agriculture to produce food with decent flavor. Picked too green, shipped and gassed, the pale pink, styrofoamlike spheres are a faint echo of the sweet, juicy, luscious fruit they could be. Just ask anyone with a sunny backyard tomato patch or access to a good farmer’s market. Everybody likes to complain about what’s happened to commercial tomatoes. But in that little building in Gainesville, Harry Klee is actually doing something about it. Klee, a horticultural scientist who’s spent the past decade and a half trying to uncover the secrets of tomato flavor, knows exactly what’s wrong with the supermarket tomato—and he knows how to fix it. Someday in the not too distant future, thanks to Klee, all of us could be enjoying much tastier tomatoes, even from the supermarket, without having to pay a fortune for them.

In his office off the main lab room, Klee—a tall, graying man with a long, thin face, flyaway eyebrows, and a slight cast in his left eye—explains that tomato flavor has been sacrificed to tomato breeders’ success in boosting yields, because growers get paid for yield, not for flavor. “Breeders have developed modern varieties that basically yield too much,” he says in his pleasant tenor voice. “Think of the leaves as factories that produce sugars, and think of the fruits as consumers. Since 1970 versus today, the modern variety yields 300% more. That’s a lot. What the breeders have done, they’ve made plants that are producing so many fruits that the leaves can’t keep up with the fruit.” As a result, he says, modern commercial tomatoes are starved of the ingredients that make a tomato tasty: sugars and the volatile odor compounds that deliver a rich tomato flavor. “These modern varieties are literally sucking all the nutrients out of the leaves, and they still can’t get enough. So the modern varieties have less volatiles, less sugars, less acid, everything. What’s in the modern fruit? Water. So the varieties that the modern consumer is given just don’t have the capacity to taste as good as an heirloom grown in your backyard.” At first sight, the problem looks intractable. The only way a plant can afford to put more sugar and volatiles—that is, more flavor—into each tomato is to make fewer tomatoes. Flavor and yield, it seems, are on opposite sides of the seesaw: if one goes up, the other must come down. Or must it?

Tomatoes are just one of the many crops that don’t taste like they used to, at least in popular belief. But where Klee and a few other researchers have clearly shown that modern commercial varieties of tomatoes lack the flavor of older heirlooms, we know much less about other crops. In fact, there’s precious little hard evidence to show that most fruits and vegetables actually did taste better in the past.

If anyone should know, it’s probably Alyson Mitchell, a food chemist at the University of California, Davis. UC Davis is smack in the middle of California’s Central Valley, where a huge share of America’s fruits and vegetables are grown, and it’s been a mecca for agricultural research for a century. Yet long-term studies of flavor just aren’t available. “There’s a lot of speculation—and probably rational speculation—that we are not growing foods with the same flavor,” says Mitchell. “It doesn’t take a rocket scientist to understand this. When I was a little girl, here in California, we used to go out into the field and pick peaches. And those peaches tasted so delicious. As time goes on, and we buy peaches in the grocery store, the flavor and aroma impact is just not the same. But if you ask my daughter ’What does a peach taste like?’ she doesn’t have the same historical memory of what a peach tastes like. We don’t have that library. The data’s just not available to make those kind of comparisons.”

Even so, we know that breeders of many crops have focused for decades on traits like disease resistance; yield; appearance; uniform size; and ease of packing, shipping, and processing—all the traits that make the crops easier to grow and deliver to distant markets. Their focus hasn’t been on flavor. As one horticulturist told me, kiwi fruit are considered “good quality” if they’re the right size and free of blemishes. Flavor doesn’t even enter into the equation.

Despite the lack of good, scientific studies that measure crop flavor directly, there might be a backdoor way to document its decline. Fruits and vegetables that are more nutritious are also likely to be more flavorful, because at least some of the molecules that make them more nutritious—the antioxidants in leafy greens, for example—are volatile, or break down into flavor volatiles. Comparisons of the nutrient content of foods over time are easier to find, and they do indeed show that nutrient levels in modern crops, by and large, are as much as 40 percent lower than they used to be. Not every nutrient shows the same decline, and not every nutrient has a direct effect on flavor. However, the overall trend is hard to ignore.

The industrialization of agriculture must also share some of the blame for the tasteless stuff in grocery stores. The peach or cantaloupe in my grocery store in wintry Canadian February has traveled thousands of miles to get there, and to survive the trip, it was almost certainly picked before it was perfectly ripe. This premature harvest cost it the chance to get the full load of sugars and volatiles that a full-term fruit could have acquired. For most fruits, which don’t continue to make sugars after harvest, there’s no way of making up for the loss. Even in August, when supply chains are shortest, many large-scale producers can’t afford to handle fruits carefully enough to let them ripen fully on the tree or vine.

But scientists like Klee are finding ways to put the flavor back into our fruits and vegetables. You’ll recall that some aromas—vanilla, for example, or strawberry—can make a sugar solution taste sweeter. If some of the volatiles in tomatoes can pull off the same trick, Klee thought, then maybe growers don’t have to sacrifice flavor for yield. He gathered up a wide variety of tomatoes—152 in all, mostly heirloom varieties but including commercial ones as well—and measured the amount of sugars and flavor volatiles that were present in each. The varieties differed enormously, with some volatiles varying by as much as three thousandfold from one variety to another.

Klee chose sixty-six varieties with very different sugar and volatile profiles and, working with Linda Bartoshuk, fed them to a taste-testing panel made up of ordinary people from the Gainesville area. For each tomato, the tasters rated its sweetness, its aroma, its tomato intensity (which Klee defines as “that concentrated ’Wow, that’s a tomato!’”), and a few other attributes. They also rated how well they liked that tomato on a scale from negative one hundred to positive one hundred, with the endpoints being the worst and best thing they’d ever experienced. “Effectively, the tomatoes go between 0 and 35,” says Klee. “Thirty-five would be a fabulous tomato. Zero is absolutely neutral.” (I think I’ve eaten a few zeros on fast-food burgers or in February salads.)

The panelists generally liked the sweetest tomatoes best. But when Klee looked closer, he saw something more interesting: The level of sweetness that tasters perceived sometimes didn’t have much to do with the amount of sugar actually present. Tasters thought a variety called Matina, for example, was about twice as sweet as the Yellow Jelly Bean variety, but the analysis showed that Yellow Jelly Bean actually contained more sugar. Matina tastes so sweet, despite its low sugar content, because it’s rich in volatile odor compounds such as geranial that make our brains think “sweet.” (Geranial, by the way, is derived from lycopene, the molecule responsible for a tomato’s red color. Orange- or yellow-colored tomato varieties make less lycopene and hence less geranial—so they taste about 25 percent less sweet than redder varieties. Something to keep in mind when you’re buying tomatoes.)

It’s worth taking a moment here to examine why plants have these volatile aroma molecules in the first place. The volatiles that account for the flavors of the plants we eat are what botanists call “secondary metabolites.” The term refers to the fact that, for the most part, they’re not absolutely essential to the life of the plant, as molecules like chlorophyll, sugars, proteins, or DNA are. Instead, these secondary metabolites serve more subtle functions, often in defense or signaling, or are merely by-products, molecular garbage left behind as the plant performs some other biochemical task.

“Usually, I can best explain what a secondary metabolite is by comparing with humans,” says Kirsten Brandt, a plant scientist at Newcastle University in the United Kingdom. “In humans, the main secondary metabolite we have is melanin, a brown pigment. Most people have it in their hair—if you don’t, you’re blonde—and most of us can make it in the skin. It protects the skin from UV light. Plants make lots of chemicals that they could actually survive without, but they need them to interact with the world around them.”

Often, those secondary compounds are there to defend the plant against predators. The bitter taste of broccoli and mustard greens comes from molecules called glucosinolates that are poisonous to many animals, particularly insects, that might otherwise munch on the plants. They’re not especially toxic to humans—we dodged that one—but even cattle are more sensitive to the chemicals, which is why canola breeders have made low-glucosinolate varieties for growers who want to feed them to cattle. Similarly, most of the pungent flavors we know from culinary herbs are actually pretty effective deterrents to feeding. (When’s the last time you sat down and ate a plateful of rosemary or sage?)

Fruits, on the other hand, want to be eaten. The whole point of a tasty, sugar-filled fruit is to tempt some animal into eating it, carrying the seeds away to be dropped somewhere they won’t compete with the parent plant. To help achieve that end, plants endow their fruits with a suite of volatile chemicals that shout out, “Good stuff here! Come and get it!” As Klee notes, many of the flavor compounds in a fruit like the tomato are closely related to essential human nutrients such as particular fatty acids and amino acids that our bodies can’t make on their own. That makes these flavor compounds a cheat-proof signal of the fruit’s nutritional quality: the plant can’t make the flavor compounds without having the nutrients, as well.

The fact that fruits want to be eaten, but only once their seeds are ripe, also explains why “ripeness” only applies to fruits, not to vegetables. Immature fruits contain sour acids and astringent polyphenols—think of an unripe apple, or an immature persimmon—that discourage their consumption; as the seeds mature, the chemical content of the fruits changes from discouraging to encouraging. Vegetables, on the other hand, are always trying to dissuade you from eating them, so ripeness isn’t an issue.

But both fruits and vegetables have a common interest in having a recognizable flavor. Remember how we use flavor to learn which foods we want to eat and which to avoid? This is the other side of the exact same coin. Vegetables want us to remember them like Dana Small remembers Malibu and 7UP: that was awful, and I never want to ingest that again. Fruits want us to remember them for their good consequences—except, maybe, for the seeds themselves. The seeds of the coffee plant, for example, pack a potent neurotoxin, caffeine, that most of us are familiar with. In nature, far from espresso machines, this poison teaches an important lesson. “We can learn that we should not eat that plant, because it makes us giddy,” says Brandt. “But we need to be able to recognize that plant. It’s important to us, and to the plant, that we should recognize the taste.” So the coffee plant has evolved distinctive-tasting seeds, and we (that is, mammals) have evolved the taste and odor receptors to recognize those distinctive flavors. After millions of years of this coevolution, says Brandt, “you’re not in doubt, when you’re eating something, whether it’s pea or potato or broccoli. All those past defences are now sitting around being useful for us—and the plants—as labels, to point us in the right direction.” Even flavor compounds that today have no toxic effect at all may well have arisen as toxins sometime in the distant past, she notes. And the pressure to evolve new compounds—new flavors—is ongoing. “Anything the plant has been using for a while, their enemies will have evolved countermeasures. So you have an arms race.”

Thanks in part to that arms race, tomatoes have at least four hundred volatiles in their fruits. However, only about two dozen are important to the flavor of the fruit, Klee finds—and the important ones aren’t necessarily those that are easiest to smell. Until recently, tomato scientists had always winnowed the hundreds of volatiles by comparing their concentration to people’s measured detection thresholds. Those compounds whose concentration soared well above threshold, they assumed, must be the most important, while those that fell below the detection threshold could be discarded as unimportant. But when Klee actually tested what made for a tasty tomato, he found that that obvious assumption didn’t hold. Some of the most prominent volatile odorants, such as the classic “tomato stem” smell you get when you brush against a growing tomato bush—and which always brings me back to happy memories of the backyard garden—make no difference to whether people like the tomato. On the other hand, some volatiles that turn out to be really important contributors to flavor are present at below-threshold concentrations. Several below-threshold volatiles, it turns out, can work together to alert the brain to their presence—just like Paul Breslin’s rose-sweet chewing gum.

Those volatiles are the secret to sweeter-tasting tomatoes, says Klee. It takes a lot of sugar to sweeten a tomato, so growers can’t do it without crippling the yield. That’s why today you can buy excellent tomatoes in the store only if you’re willing to pay a lot more for them. But volatiles don’t cost a tomato plant much—they’re present in such small quantities anyway that tomato breeders could crank up volatile levels manyfold while barely denting the yield at all. “All of a sudden, you’re doubling the perception of sweetness,” says Klee. That should make sweeter, richer-tasting tomatoes possible at a price everyone can afford.

Volatiles, incidentally, are the reason why you should never, ever put a tomato in the refrigerator. A tomato is constantly leaking volatiles into the air (which you can easily verify by sniffing a good, ripe one) and replenishing the loss by making new ones. Chilling turns off the enzymes that make the volatiles—and one of the peculiarities of the tomato, a tropical plant, is that the enzymes stay off, even after you take the fruit out of the fridge. Volatile content goes down as molecules leak out into the air and aren’t replaced, so a tomato that’s spent time in the fridge tastes less sweet and has less tomato flavor. (And, by the way, most of the volatiles leak out the stem scar at the top of the tomato—so, all else being equal, a tomato sold “on the vine” with a bit of stem attached ought to be a little more flavorful than one without.)

Klee has already taken the first steps down the road to the flavorful tomato of the future. In 2014, his team released its first two new varieties, Garden Gem and Garden Treasure, that were created by crossing high-volatile heirlooms with modern, high-yielding varieties. The hybrids yield nearly as much as the commercial varieties but keep almost all the flavor of the heirlooms, he says. As I talked tomatoes with Klee, five golf ball—sized Garden Gem tomatoes sat between us on his desk. After a couple of hours, he offered to cut them up so I could taste what we were talking about. It wasn’t an ideal test—these were April tomatoes, after all, ripened when days were short and temperatures relatively low, so there was little chance they’d reach the glorious heights of a midsummer heirloom. They didn’t—but they certainly tasted sweeter and tomatoier than anything I was likely to get in the grocery store at the time. Grown under better conditions, Klee’s two new varieties have definitely gotten people excited. As I write this, the varieties are not yet commercially available, but Klee has sent seeds to more than thirty-two hundred people who donated money to the tomato research program, and the feedback is enthusiastic. “We’ve had several people write to tell us they’re the best tomatoes they’ve had in their lives, which makes us feel good,” he says. “People really want these things. It just shows how big the pent-up demand is for good tomatoes.”

Just an hour down the road from Klee’s lab, in the sandy soils of central Florida, a plant breeder named Vance Whitaker is trying to solve the flavor problem for another fruit that’s often disappointing in the grocery store: strawberries. The big problem with strawberries is that they’re what’s called a “nonclimacteric fruit,” meaning that they don’t ripen any further after harvest. You can’t treat a strawberry like a banana or an apple or a pear—or a tomato, for that matter—and prod it into ripeness in the warehouse with ethylene gas. All you can do is let it ripen on the bush as long as you dare: once you pick the berry, it’s all downhill. It’ll never get any better. And because strawberries are so fragile, growers can’t risk letting them ripen fully before picking, because they’d never survive the rigors of shipping and handling. The result is that a grocery-store strawberry will rarely be as ripe as one you’d pick yourself—and the clearest sign of that is the white “shoulders” you’ll see at the stem end of most grocery-store berries.

What to do? Scientists could try to find a way to help the strawberries last longer, so that growers can afford to pick them when they’re riper. Or they could look for a way to boost the flavor directly. Whitaker chose the second route, and started digging deep into the chemistry of strawberry flavor, using the same techniques that Klee used for tomatoes. (In fact, the two research teams share several scientists in common, including Klee himself and Linda Bartoshuk.) Strawberries are a lot like tomatoes, he found: People like sweeter ones better, and they also prefer a more intense strawberry flavor, which depends on the volatile chemicals. And, just like tomatoes, if the plants make too many berries, they can’t afford to stock them with enough sugar. That puts breeders like Whitaker in a bind. “We could increase yield by a pretty sizeable percentage in just a couple of generations,” he explains, “but we would drastically reduce sugar content.”

One solution might be to breed a more vigorous strawberry plant that photosynthesizes more energy, so that it can afford a bigger sugar budget. A more likely option, though, might be to take a page from Klee’s playbook and tinker with the volatiles. Sure enough, when Whitaker and his colleagues looked at the volatiles in strawberries, they found several that made the berries taste sweeter, independent of the amount of sugar that was actually present. (Curiously, even though many of the same volatiles turn up in strawberries as in tomatoes, different ones affect sweetness in the two fruits. It’s all in the context.)

Strawberry flavor intensity, too, depends strongly on the mix of volatiles that are present in the berry. Whitaker is looking closely at a molecule called gamma-decalactone—the very same peachy-smelling molecule that bridged top notes and bottom notes in the artificial strawberry flavor Brian Mullin designed for me at Givaudan. Some strawberry varieties have it, some don’t. Whitaker’s team sorted through the genotypes of the haves and the have-nots—a harder task than it sounds, because strawberries have not two but eight copies of each gene—to find a single gene variant that accounted for the difference. With a clear target identified, breeders will have an easier time ensuring that any new varieties have the good gene for this important flavor compound. They can use the same technique—and all of Whitaker’s genotyping work—to find other flavor genes more quickly.

There are a few other, nongenetic secrets to growing tasty strawberries, says Whitaker. Cool temperatures, especially at night, help the plant store more sugar in the berries. As a result, Florida strawberries always taste best early in their growing season, in December and early January; quality declines as the weather heats up into February and March. (It seems counterintuitive to put strawberries at the top of your shopping list during the darkest days of winter, but that’s exactly what you should do, at least if your grocer gets berries from Florida. Berries that come from California or Mexico have different seasons for peak flavor.) And just a little bit of water stress, or just a little bit of fertilizer limitation, also tends to improve flavor by slowing growth and giving the plants plenty of time to stock the berries with sugar and volatiles. In contrast, good soils don’t seem to make any difference. Whitaker’s soils in Florida are nothing but coarse sand, and many growers in Asia and the Netherlands produce delicious berries hydroponically, with no soil at all.

Tomatoes and strawberries are unusual in having crop scientists pay much attention to flavor. Most other fruits, and almost all vegetables, exist in a vast, undifferentiated sea of produce where one head of broccoli is interchangeable with the next. “The guy who’s buying the stuff for the supermarket, he wants it to taste the same as last time,” says Brandt. “What most of the supply chain wants is predictability and low price. There are no consumers wanting special broccolis. They’re just not there.” In such a milieu, it’s not surprising that the science of flavor on the farm is almost nonexistent.

We don’t know much, for example, about how a farm’s soil affects the flavor of the crops. (Not much at all, in the case of hydroponic strawberries!) Here’s Alyson Mitchell on one of the crops she studies, spinach: “I don’t think there has been a single sensory study done on spinach looking at the effect of growing environment on flavor. I would be shocked to find it.” And most other crops are in pretty much the same boat. Once again, we can get a little more information by asking about nutritional quality instead of flavor, since that attracts a little more research money, but even there, big lessons aren’t easily found.

There is one crop, however, where flavor matters above anything else, even yield: wine grapes. The whole point of growing wine grapes, of course, is to make a wine with a distinctive, appealing flavor. If anyone knows how soils and farming methods affect the flavor of a crop, it’s going to be viticulturists. For an example that’s been worked out in great detail—an example, moreover, that you can taste for yourself tonight—let’s journey halfway around the world from Harry Klee’s tomato lab to New Zealand.

Mike Trought loves to talk wine—and there’s probably no one in the world who knows more about the acclaimed white wines made from the sauvignon blanc grape, particularly in the Marlborough region of New Zealand’s South Island. A cheerful, balding fellow who’s spent more than three decades in Marlborough as a researcher, university lecturer, winery consultant, and viticulturist, Trought recalls the time, in the midnineties, when he made a pilgrimage to the world-renowned oenology department of the University of California, Davis. At the time, New Zealand wines were just beginning to emerge onto the world stage, and Trought poured two bottles of Marlborough sauvignon blanc for a group of Davis researchers. Everyone dismissed them, calling them too acidic, too herbaceous, unsubtle—in short, unripe and inferior. Two decades later, the joke’s on the folks from Davis. “New Zealand sauvignon blanc has now become a benchmark for sauvignon blanc around the world,” says Trought. “We can’t produce enough.” One of the wines Trought poured at Davis that year, Cloudy Bay, quickly became so popular that wine shops couldn’t keep it in stock, especially in Britain. (Trought wonders, a bit impishly, whether wine experts might actually hinder progress in the wine world.)

If you’re at all fond of wine, you have probably encountered the distinctive flavor of New Zealand sauvignon blanc. Sip one of those wines, especially from Marlborough, and you’ll be struck by the pronounced aromas of passion fruit, green pepper, and what’s often described as boxwood, or “cat’s pee on a gooseberry bush”—the latter being an actual, if unconventionally named, commercial wine. The intense, distinctive flavors make New Zealand sauvignon blanc an ideal test case for understanding where a wine’s flavor comes from and how growers and winemakers can influence the outcome. As an added bonus, New Zealand’s wine industry is relatively new, so tradition doesn’t get in the way of science.

What gives Marlborough wines their distinctive flavor? It’s certainly not the grape variety alone. Virtually all of New Zealand’s sauvignon blanc vines are descended from a single clone originating in the vineyard of France’s fabled Chateau d’Yquem, where they yield a vastly different wine. Instead, a lot depends on the vineyard soils. Not, perhaps, in the way you’d think. The notion that you can somehow “taste the soil” in a wine is completely false. Grape vines take up only water and simple nutrients like nitrogen, potassium, and calcium from the soil. They make all their more complex biomolecules—including the flavor volatiles—in-house. To put it more bluntly, none of the volatile molecules that determine a wine’s flavor come directly from the soil. (This calls into question one of wine writers’ favorite buzzwords these days: “minerality.” You won’t find the term much in wine writing before the 1980s, but it’s become a term of high praise today. Whatever the term means—and the experts don’t exactly agree—it’s not the flavor of the vineyard. One study suggests that “minerality” is a description that emerges only when a wine lacks any other distinctive flavor.)

Instead, a vineyard’s soil affects flavor indirectly, by altering how the vine grows and, especially, how quickly the grapes ripen. In Marlborough’s Wairau Valley, the vineyards sit on an old river flood plain, where the soil is a jumble of sand, gravel, and cobble deposited by the river channel as it meandered across the plain. Soil quality can change rapidly as you walk through the vineyard, so that grape vines only a few yards apart experience very different soils. Where the soil is shallower, vines tend to be less vigorous, and their fruit ripens earlier. (Trought isn’t sure why the smaller vines on stony soils ripen earlier, but he suspects it might be because the vines put more of their energy into the grapes when growing conditions are poorer.) When harvesters go through the vineyard, the patchwork of soils means that some bunches of grapes will be harvested at a riper stage than others. The volatiles responsible for the green pepper flavor, known as methoxypyrazines, form early in grape development, so they are more prominent in less ripe grapes. Meanwhile, the thiols that account for the passion fruit notes dominate in riper grapes. This mix of ripeness, and the different flavors it delivers, helps give the Marlborough wines their complexity. “To some extent, that’s the characteristic of Marlborough sauvignon blanc,” says Trought.

There’s more to the story, though, as Trought and his colleagues discovered. “When we started our sauvignon blanc program, we did it because we thought it was going to be easy,” he says ruefully. “As we got into it more and more, we realized it was much more complex. It’s not just what’s in the vineyard that matters.” If you pluck a grape off the vine and chew it, you won’t notice much passion fruit flavor, because the thiol molecules haven’t formed yet—only their odorless precursors are present. The thiols themselves form during fermentation, as the yeast attack the precursors and split off thiol molecules. Rough handling of the grapes causes them to accumulate more of the precursors, so machine-harvested grapes yield wines with about ten times as much thiol as handpicked ones. This, incidentally, may be part of the reason that New Zealand sauvignon blanc, which is generally mechanically harvested, tends to have a much more pronounced passion fruit flavor than French sauvignon blanc, which is usually hand harvested. Even trucking the grapes from vineyard to winery leads to more thiols in the finished wine.

The biggest effect on the final flavor of a wine comes from fermentation, as wine yeasts and other microbes attack the sugars, proteins, and other molecules in the grape juice and convert them to alcohol and flavor volatiles. Each strain of yeast approaches this task with its own unique tool kit of genes and enzymes, and as a result different yeasts can yield very different wines from the same juice. Winemakers are very aware of this, and put a great deal of thought into their choice of yeast. Here, too, regional differences matter, because every winegrowing region—and, quite possibly, every vineyard—harbors its own unique microbial ecosystem. Winemakers rarely sterilize their grapes before fermentation, so these microbes end up in the fermentation tank. In fact, many winemakers rely exclusively on natural microbes for fermentation. So it makes sense that part of the regional character of a wine—its “terroir,” to use the term beloved of wine critics—might be the result of different microbial actors taking the stage during fermentation.

Plausible, but until recently, untested. A few years ago, geneticist Sarah Knight and her colleagues at the University of Auckland, New Zealand, set out to see whether it really was true. To rule out any differences in the grapes themselves, Knight started with a single batch of Marlborough sauvignon blanc grapes and sterilized them to kill any resident microbes. Then she divided the juice into a series of tiny fermentation tanks, and sowed each one with a different wine-yeast variant gathered from one of New Zealand’s six main winegrowing regions. Same juice, same fermentation conditions—the only difference was the yeast itself. In the end, the yeast variants from each region produced a wine with a detectably different aroma profile. Theory confirmed! Not only that, but Knight’s study probably underestimated the effect of microbial differences, because she used only wine yeasts, not the whole microbial flora.

For other crops, too, any effect of the soil on flavor is likely to be indirect. The soil a plant grows in determines how much water and nutrients it has access to, and therefore its energy and materials budget for the sugars and volatiles that determine flavor. You’d think that more would always be better, but it’s more complicated than that.

To explain, I turned to Carol Wagstaff, a crop scientist at the University of Reading in England—just a few minutes’ drive, actually, from chef Heston Blumenthal’s Fat Duck restaurant in Bray. Reading’s research group is one of the few to actually study how growing conditions, shipping, and storage affect the nutritional value and flavor of crop plants. Wagstaff has unruly, long brown hair and a large, strong face that lights up when she talks about her work. If conditions are too easy, she says, plants have little need for secondary compounds and put all their energy into growing as fast as they can. Only when they begin to feel a budget crunch do they invest in defending what they’ve already got. “A bit of controlled stress doesn’t go amiss. When a plant is stressed, you’ll get more secondary compounds, and that means more flavor and more nutrients,” she says. That’s likely why Whitaker’s strawberries also profit from a little water stress. Exactly what that stress response means in flavor terms is likely to depend on what Wagstaff calls the “metabolic bureaucracy” of the plant—that is, its particular genetic endowment of enzymes and the particular balance of secondary chemicals they favor. Much of Wagstaff’s research centers on arugula, which the British call rocket, and that’s exactly what she sees there. “You can see quite clearly that some genotypes of rocket will preferentially shunt production in one direction, and other genotypes head down another route when they’re stressed,” she says.

Soil microbes, too, could play some role in determining the flavor of the crops they grow with. For example, baby corn—a popular vegetable in some Asian cuisines—contains a volatile flavor molecule called geosmin, the same molecule that gives red beets their earthy flavor. Researchers think the young corn plants don’t make the geosmin themselves, because corn grown in an English greenhouse lacks the compound. Instead, they think microscopic fungi living with the roots of the corn plants make the geosmin. The plants take up the fungi through their roots, and the geosmin comes along for the ride. It’s possible that soil microbes affect flavor in other ways, too, but so far there’s little actual evidence.

So far, we’ve been talking as though more flavor was always better—but for many vegetables, particularly members of the sharp-tasting mustard family, like arugula and brussels sprouts, that’s not necessarily true. Many people—especially those who carry the bitter-sensitive version of the T2R38 taste receptor—find the bitterness of their secondary compounds off-putting and would prefer that their brussels sprouts have less flavor, not more. “Horticulture is essentially messy,” says Wagstaff. “You’ve got the variable genotypes of your plants, you’ve got the environment you’re growing them in, and you’ve got the varying genotypes of the consumer.”

Once a fruit or vegetable has been picked, its flavor continues to change during storage and en route to your grocery store. Partly, that’s because volatile flavor molecules leak out into the air, as we’ve seen happens with tomatoes. At the same time, though, enzyme activity in the tissues can produce new flavor molecules or alter old ones. Occasionally, this can mean that a fruit or vegetable actually improves with storage. Arugula, for example, continues to produce glucosinolate molecules during cold storage after harvest. When you chew a leaf, they turn into flavorful isothiocyanates. That’s good news for your salad: The arugula you buy in the grocery store—if it’s relatively fresh—may actually be more flavorful than if you’d picked it in your own garden this afternoon. After a few more days in the refrigerator case, though, that advantage goes away, as the “fresh” set of flavor compounds gives way to nastier products that result from fat breakdown. This happens at different rates depending on the variety of arugula—some store better than others, Wagstaff has found.

Some other vegetables last a long time with little or no change in quality. An onion or a potato, for example, is meant to just sit there like an inert lump—that’s its job, as a storage organ for next year’s growth. So it makes sense that we don’t notice much of a decline in flavor. Others, such as corn and carrots, are sweetest just after picking, because enzymes convert their sugar into starch, and no new sugars arrive after picking. They’ll last, but their flavor will be disappointing. But a head of broccoli or an asparagus spear hasn’t evolved to be long lasting. Far from it—both are rapidly growing shoots, and as soon as you pick them their flavor starts to degrade. One Spanish study, for example, found that more than 70 percent of the glucosinolates in a freshly cut head of broccoli have vanished after a week in cold storage, and another 10 percent disappear after another three days on a grocer’s display. That’s a lot of lost flavor.

Many people think that another way to ensure tastier fruits and vegetables is to buy organic when possible. It makes sense, in theory: If a little stress is good for flavor, then you’d expect that organic crops ought to benefit, flavor-wise, from the extra insect damage and weed competition they experience. Hundreds of scientific studies have compared the flavor—or, more often, the nutritional content—of organic and conventional crops. The results, unfortunately, are a mess. Some studies show that organic crops are indeed better, while others find no difference. Even the so-called meta-analyses—in which researchers scour the library for every comparison they can find, then add up the results to get a majority opinion—haven’t reached consensus on whether organic is better.

A big part of the problem is that the answer you get depends on how you ask the question. You could go to the grocery store, buy a head of conventional and a head of organic broccoli, and measure—or taste—the difference in secondary compounds. But if the conventional broccoli was harvested two weeks ago in Mexico, and the organic was picked yesterday just down the road, that difference in freshness might have much more impact on the flavor than any organic versus conventional effect. It’s hard to generalize, though. The Mexican broccoli could have gone straight into a refrigerated warehouse and stayed under refrigeration right until the time you put it in your shopping cart, while the local one could have spent a hot summer’s afternoon in the back of a pickup truck, followed by hours in the sun at the farmer’s market. In that case, local might not mean fresher.

Ideally, you’d like to compare the flavor of identical crops grown side by side with organic or conventional methods, because that cuts out a lot of the potential sources of confusion. Researchers at Kansas State University did exactly that a few years ago, planting onions, tomatoes, cucumbers, and several leafy greens in greenhouses in identical soils. When the crops were harvested, about one hundred volunteers tasted organic and conventional samples of the same vegetable—without knowing which was which—and rated how much they liked them and how intense the flavors were. The results? It didn’t matter one bit whether the vegetables grew organically or conventionally. The taste testers liked them all equally (or, in the case of mustard greens and arugula, disliked them equally—evidently Manhattan, Kansas, is not the place to get rich from an arugula greenhouse). The only difference was that people thought the conventional tomatoes had a little more flavor, probably because they were also a little riper.

That’s not to say that you won’t find organic produce more flavorful. As we have seen, our expectations play a big role in flavor perception—for example, wine tastes better when we think it’s expensive. That bias probably comes into play here as well: If you think organic produce will taste better, then it probably will, to you. Consider what happened when Swedish researchers gave unwitting university students two identical cups of coffee, telling them that one was “eco-friendly” and the other conventionally grown. Sure enough, most volunteers thought the eco-friendly coffee tasted better—and the effect was strongest for people with the strongest environmental consciousness.

Even if organic farming or other differences in growing conditions do make a difference to crops’ flavor, it’s likely to be less important than flavor differences among varieties. If so, then breeders, not farmers, may be the critical link in producing tastier fruits and vegetables. In upstate New York, for example, Cornell University plant breeder Michael Mazourek has been working on breeding a more flavorful squash. Squash and other vegetables are the poor cousins of the agricultural world, Mazourek says. You can go into any grocery store and find perhaps a dozen different apple varieties, each offering its own recognizable flavor profile. And we know them all by name: a Granny Smith will be tart and firm, a Spartan sweet and softer, Golden Delicious rich in estery fruit flavors. No doubt you have your favorites. But can you name your favorite variety of broccoli, or your favorite butternut squash? I’ll bet not.

“Vegetables are still part of a commodity system, where sameness is one of the overarching goals,” says Mazourek. “There’s not value in people being able to tell that the bell pepper in the grocery store is different from the one last month. It’s an antivalue.” With all the commercial pressure working in the direction of sameness, there’s little incentive for anyone to develop a tastier version.

Mazourek is trying to change that. He starts by seeking out heirloom squashes noted for their good or unusual flavor and crossing them with commercial varieties, then planting the progeny out in his field. When the fruits ripen, he gathers them and selects the most promising ones for taste testing. “I can’t possibly eat some of every squash and stay sane, so we have some proxies that narrow the pool that we do the taste tests on,” he explains. First, he picks fruits with the highest level of dissolved solids—that is, sugar and other molecules that might contribute to flavor. Then from those he selects the ones with the deepest yellow flesh. Those have the highest levels of carotenoid pigments, key precursors of many flavor compounds. Unlike Harry Klee’s work with tomatoes, Mazourek doesn’t yet know which of the many flavor molecules are most important in a good-tasting squash, so he can’t measure them directly with a gas chromatograph. Mazourek has to do his flavor analysis the old-fashioned way: he roasts several varieties of squash and sees which ones he likes best. (By the way, here’s a squash specialist’s advice for the most delicious way to roast a butternut squash: halve it and scoop out the seeds, cover and roast in a four-hundred-degree oven for forty-five minutes. Then uncover the squash, baste it with butter or oil, and continue roasting until tender. “It’s not what Betty Crocker says,” Mazourek admits. “Roasting them for longer and hotter really is a way to bring out a lot of the savory flavors layered on top of the sweetness.”)

After many generations of breeding, Mazourek has ended up with what he thinks is the best butternut squash on the planet. Sometimes called the Barber squash—after Dan Barber, a New York City chef who encouraged his efforts and now serves the squash in his restaurant—Mazourek’s squash has more dissolved solids and a higher carotenoid content than any other squash. “Everything is amped up,” says Mazourek. Better yet, the fruits have a built-in ripeness indicator, changing color dramatically from deep green to a rich caramel brown when they’re perfectly ripe, so that pickers can be sure they spend long enough on the vine. “There’s about a fourfold boost to the carotenoid content in the Barber squash,” Mazourek says proudly. “Half of that is in the squash itself. The other half is that it’s ripe.”

For most consumers, though, the biggest immediate payoff is likely to come from Klee and his tomatoes. Since my visit, Klee has been digging deeper into the genetics of tomato flavor. In collaboration with a research group in China, he has now fully sequenced the genomes of more than four hundred tomato varieties, and fully mapped out their chemical content. In the same way that human geneticists search the genome for gene variants, or alleles, that contribute to diseases, Klee has scoured these tomato genomes for alleles that are important for production of sugars and volatiles. And he can compare modern commercial varieties to heirlooms to see exactly where breeders went wrong.

Back in the 1920s, breeders latched on to a chance mutation that eliminated the dark green “shoulders” on unripe tomatoes. The new, uniformly colored fruit helped growers decide the best time to harvest, and consumers preferred the solid red color in the grocery store. (“People do buy with their eyes,” says one tomato grower.) The mutant looked like a big winner—and, in fact, almost all commercial tomatoes grown today carry this mutation. But there was a downside: To allow the green shoulders to redden fully, the mutation interfered with the production of chlorophyll in the fruit. Less chlorophyll means less photosynthesis. The new tomatoes lost out on a sugar boost that green-shouldered tomatoes enjoy—and, as a result, uniformly ripening tomatoes have about 20 percent less sugar.

For volatiles, the losses were even worse. Over the course of decades of breeding for high yield, the high-producing alleles for flavor volatiles have simply fallen by the wayside, because breeders didn’t know they were important and didn’t test for flavor. “For the volatiles, I’d say at least half of them are the wrong alleles,” says Klee. Fortunately, the good alleles are still there in the heirloom varieties—and now that Klee knows which genes are important, it should be straightforward to breed the good alleles back into higher-yielding varieties. “The road map is very clear. We know exactly what we need to do,” says Klee. “It just takes time.”

It shouldn’t be long before everyone should have access to tastier tomatoes, and perhaps other crops as well. Even before that, though, cooks will still strive to draw as much flavor as they can from the raw materials in their kitchens.