As anyone can see by looking at the Wild Garden variety list, I have an oversized relationship with lettuce. Lactuca sativa was not always my most important crop. In the earliest days of our “neurotic salad” business (as Alan Kapular once referred to it), lettuce just seemed too common and unremarkable to include in the exotic mix of 20 greens that no restaurant-goer had ever seen before. This was around 1984, and Mark Musick of Pragtree Farm (60 miles north of Seattle) had given me the keys to my career by sharing his amazing revelation that chefs wanted wild gathered foods, especially salad greens and edible flowers. And they wanted “European greens” that most of us had never seen in a garden, like arugula, chervil, and mâché. It was a different time, when so many things we now grow commonly were completely exotic.
After we put our mark on it, salad went from lettuce, carrots, and tomato to something that waiters could no longer identify. To help them out, we included “identification sheets” with salad items taped to stationary, little sprigs and leaflets of chickweed, garden cress, hot and mild mustard shoots, dandelion, lambs quarters, endive, chicory, radish pods, tatsoi, mizuna, pungent mustards, red Russian kale, kale shoots (now called kale raab), sheep sorrel, sacred basil, lemon thyme tips…and somehow, lettuce struck me as too tame to include in this assemblage. Strangely, some chefs insisted on calling the whole mélange “Lettuce,” which always made me completely crazy.
But then lettuce handed me the keys to my other career. A single outcrossed lettuce seedling, red in a seedling flat of otherwise green ‘Salad Bowl,’ caught my attention and engaged my imagination. This will be something new, I thought. I didn’t understand how new until I had collected the seeds of that plant and grew them out. The seedlings were all unique. Sixty-five plants, each a little or a lot different from the next, a genetic reshuffling of all the traits of the original parents—a green oakleaf crossed by a red romaine. I was looking at an F2 generation for the first time. This is where new varieties come from, I realized. This was how to create salad diversity like no chef had ever seen, and I could do this myself, on an organic farm. This would be lettuce exotic enough to use in Wild Garden Salad.
Lettuce soon became my spirit plant. I had a knack with it. Though it is strongly self-pollinating, I could make crosses without the clumsy methods described in text books that involved careful timing, trimming away the pollen parts, washing off the receptive parts with water, collecting and applying pollen from the other parent, marking the individual crossed flower heads with little strings and catching the ripened seed before it was dropped or eaten by goldfinches. Rather, I just “married” the plants, and expected nature to take it from there.
In any new breeding project, this crossing of individuals is the first step in creating genetic variation from which new kinds may be selected. For plant breeders, the world is divided between the “selfers” and the “crossers.” Self-pollinating plants require no pollinators to make seed. Their flower structure is designed to assure that pollen will easily pass from the stamens to the sticky sigma of the ovary pistil, often before the flowers have opened themselves to insects. Selfers may also be cross-pollinated with the help of pushy insects like bumblebees that force themselves into unopened bean flowers, or by tiny insects like thrips that are ever about the flower head and shimmy themselves between the unopened petals, dragging pollen with them. Crossers may be capable of self-pollinating (or not) but they are really designed to encourage insect visitation with offers of nectar and pollen for food. They also offset time schedules for pollen ripeness and stigma receptiveness. All of this serves to assure that outcrossing will occur. These are two different evolutionary strategies, one depending on pollinator partners and the other, not. Crossers may need an ecosystem of support but selfers could live happily forever in a greenhouse.
When making lettuce crosses between varieties, I grow them closely together in rows, removing any plants that are subpar or afflicted before they start to bolt. I note my favorite plants and mark them as such. Once their flower stems have elongated and flowers are about to open, I gently bend them toward one another and press their heads together so they are touching which will allow flower thrips to meander between them, carrying pollen into places we can’t even see. This is what I call “marrying” the heads. With some luck, my favorite selections of each variety will be close enough to bring their heads together. After three weeks of flowering embrace, before the seed caps have ripened, the heads are pulled apart so that ripe seeds may be collected separately from each plant. Seed from the best plants of each variety will go into individual envelopes with names and descriptive notes, seeds of the non-select plants are bulked together by variety, essentially as insurance, lest the selections fail.
As a salad growing plant breeder, I would use these seeds as salad planting stock early the following year. By planting 3 seeds per cell in trays, one could easily find the “off types” (the F1 crosses) in each variety, perhaps 5 or 10 per 100 seedlings. It’s best, of course, is to find these from the preferred parents. Should it happen that the best parents don’t produce the desired cross, I’d resort to the larger amount of bulked seed. So why not just cross two good plants in the first place? Why the redundancy? Because favorite plants are known to die before making seed, more often than most can imagine. Time is one thing you can’t get back, so you want this one mating scenario to be successful at all costs. By some reckoning, you really only need one successful cross to create the next generation, but I will never count on a single plant to keep progress ongoing. Diversity and redundancy are insurance against failure.
How does one “easily” spot the F1 (“off types”) at a young age? The parent varieties always have some dominant and recessive traits that distinguish them. If one parent is red and the other green, red F1s will show up among the green selfs. If one parent is oakleaf and the other romaine, the oaky types in the romaine flat are the F1 crosses. Red is dominant to green. Oak is dominant to entire. Spotted leaves are dominant to green. Dark green is dominant to light green. Sometimes the traits are not clearly dominant but in any case, the F1 cross will in some way be different from the parent’s self-pollinated progeny, and a breeder who knows the material will see the difference.
The F1 generation is boring, looking somewhat averaged between the original parents. In any case, it is nothing to get attached to, because the F2 generation is going to make you forget all about it. The F1 generation just needs a safe place to pass its time. If you have multiple F1s, let them all flower together. They may or may not be identical depending on the uniformity of the parent varieties (heirloom diversity can result in F1 diversity). If the look or performance of some plants is lacking, remove them. Collect seed of each F1 plant into separate envelopes when ripe. This is the F2 seed generation and it’s where the widest range of choices are found.
The F2 generation represents the newly shuffled deck of traits, each plant like a new hand in a card game. Because lettuce is self-pollinating, each of these new F2 individuals is the beginning of a line of descendants. If the individual is green and smooth edged, you may be sure that none of its (selfed) descendants will be red or oakleaf. Those traits are not in the cards for those individuals unless, by chance (less than 5%), they cross with a sibling. But if an F2 individual is red and oakleaf, there is a good chance that it contains momentarily hidden recessive traits (green, entire margins) that will appear in the next (F3) generation and generations to come. This natural process of genetic recombination toward stable (homozygous) lines is called segregation, and it takes 6 or 8 generations or more. It is important to remember that genes for traits we cannot see (disease resistance, flavor, nutrient profiles, drought and cold tolerance) are also going through the same resorting process as the ones we can see (color, shape).
Selection is tempting in the F2 generation with all those beautiful forms displayed so individually, each one of a kind. The problem with doing so, picking winners so early in the process, is that every one of them is an unstable genotype. They might be winners this year, losers the next, or (possibly) vice versa. For this reason, it is important to preserve as many F2s as possible so you can see the F3 families that spring from them.
Each F2 individual gives rise to thousands of F3 seeds, thought of as a “family” with traits that are dependent on the genetic potential provided by the F2 parent, and the various recombinations that are possible from that potential. In this generation one can see that some families have pronounced strengths, others have weaknesses, but in every family some individuals do better, shine brighter, taste finer, than others. This is the beginning of selection. We would like to remove from our load the least promising genetics so that we can focus on the most promising. One way to do this is by first removing whole families that seem weak, then removing the weakest individuals from the best remaining families. What’s left are the best individuals from the best families. These are F3 family selections and each family bears some familial resemblance. Some may be green romaine types, others red or blushed romaine. Some may be oakleaf in all the various colors, others with oakleaf shape but upright stature, like a romaine. The distinctions are endless so it’s best to think about the similarities. Maybe that’s a lesson for life.
Time and space are limited. I usually look at 27 or 54 individuals from each F3 family that interest me at the time. I’m looking at vigor, downy mildew resistance, other disease issues, size and shape, color with its intensity and distribution, leaf gloss, internal color, tip burn, and flavor. Flavor is last on that list for a reason. If the plant is seriously flawed for any of the important criteria mentioned, its flavor will never come to the table because no one will want to grow or purchase it. Flavor is also the most difficult trait to nail down because it varies with the weather and time of day, age of the plant, which leaf you taste, and whether this is your first or 40th taste of the day. There is also the problem of personal preferences and we all know there’s no sense in arguing taste. Some of my favorite varieties have great taste and fatal flaws (tipburn in ‘Red to the Heart’ comes to mind), and I don’t sell much of them to commercial growers or seed companies. Of course, it’s possible to improve a flawed individual that has good flavor, but you may lose the flavor while doing so. The preference is to find something faultless with flavor and hold onto that, but it’s elusive.
The final F3 selection happens in the seed-bearing stage. Under wet conditions, 50% of a lettuce seed plot will die as it bolts, primarily from botrytis and sclerotinia. Selection at this stage is for resistance to these diseases. But even with drip irrigation and a dry climate, any number of late stage pathogens might remove a favorite plant from the game. I harbor a supernatural hypothesis that placing a flag near a plant invites death. For this reason (and others) it is best to save several (or dozens) from the F3 families. My practice is to save at least 6 “best” plants as individuals in their own packets, and all the “good” ones together in one bag as an F4 population. As a seed grower, selection for a quality seed plant is as important as selection for a beautiful rosette or heavy yielding head.
The following February, these F4 family selections and the population are replanted in flats in the greenhouse, 3 seeds per cell. Every generation begins this way. I am looking for the most vigorous seedlings at the beginning of each generation, but I can also begin to steer the families toward some internal uniformity by removing red seedlings from mostly green families, or removing dark green stragglers from light greens, or oakleaf from romaine, etc. I can also see color intensity and distribution at the young seedling stage. Seedling selection can advance a project toward uniformity in fewer generations, I think, but there is always something lost when greater diversity is stymied by overselection in the early generations. I think. But time and space are limited.
The F4 families are represented by 27 or 54 individuals from the selection packets, and as many from the population bag as there is room for, perhaps a whole flat. The same selection process is repeated in the field. The breeder mind begins to see what it wants. The most healthy, interesting, unique, “best” kinds begin to get all the space and attention, but packets of the more average types enter the seed bank, or as one of my workers once observed, “the seed morgue,” where the unseen or “unoutstanding” go to expire. I have trouble throwing things away. But as my old friend John Navazio likes to quote from an older alfalfa breeder, “It isn’t how much you keep. It’s how much you throw away that really counts.” This is a comforting mantra every time I feed the expired to the compost pile.
Over the next 3 or 4 generations, the best plants of each generation are saved and evaluated, their numbers increasing. Family selections become more uniform, then are bulked together and grown as a population, a proto-variety that has usually earned a working name. After this, ongoing mass selection removes rogues and weaklings until, as my friend Navazio likes to say, “We can wave the chicken bone over it,” and call it done.
Sort of. Like creation, selection is never done. All varieties require maintenance, which means rogueing is a part of every seed production cycle. In truth, I often release early versions of new varieties as F5 populations. These are uniform enough to be commercially useful and provide others an opportunity for final selection in a region of intended use. In this way I am actually sharing breeding opportunity with others, not something normally done by commercial breeders. The OSSI-pledged variety ‘Emerald Fan’ (available from Uprising, Siskiyou, and Quail Seeds) was selected by Jonathan Spero of Lupine Knoll Farm from an early release of my variety ‘Chartreuse Butter Tongue.’ His selection has enjoyed more attention than mine, so I consider this an example of cooperative breeding success and a kind of affirmation for my unorthodox approach to varietal release. As we say in the open source seed community, These are free seeds—free as in speech, not free as in beer.
May the tribe increase.
Originally published in the 2020 Wild Garden Seed Catalog.