Plant Breeding for Resilience & Nutrition

I’ve been saving my own seed to some degree for over twenty years, and messing around with breeding projects for eighteen. On our farm at this point we save our own seed from a number of crops, have several varieties that are unique to our farm, and always have at least half a dozen breeding projects going on in any given year. I don’t see any real line between seed saving and breeding. Seed saving, when done properly, should involve selection and improvement. Over time, any variety being well stewarded by a conscientious seed keeper ought to become more well adapted to its location, as well as its flavor, color, and texture preferences. An attentive reader will have noticed the conditional hedging; there is lots of room for a variety to run down instead of improving. This goes a long way to explaining the low regard many commercial growers have for heirlooms and older open pollinated varieties.

Breeding for Goodness Sake
Our seed saving and breeding work is based on a mixture of socio-economic political issues and philosophical beliefs. On the socio-economic political front, most folks in the organic community are aware at this point that the seed industry has undergone waves of consolidation, so that the vast majority of our food supply is controlled by a handful of corporations. Four of the largest seed companies in the world (Bayer/Monsanto, Corteva/Dow/Dupont, ChemChina/Syngenta/Novartis, and BASF) are primarily chemical companies – selling seeds is a way for them to sell more pesticides. Several major vegetable seed companies (such as Limagrain/Vilmorin and Sakata) have maintained their independence from the chemical giants, but engaged in the same economic strategy of buying out smaller companies to acquire greater world-wide “market share.” As an independent small-scale market farmer, I want my customers to support small-scale, local production, and I feel it is hypocritical to base that production on seeds from international corporate conglomerates. Luckily, there are a lot of small scale farm-based seed companies these days Our seed ordering is a lot more complicated now that we have decided to let ethics be a stronger influence on our purchases than ease and cost (though they aren’t always in conflict). We have almost entirely weaned ourselves off of hybrids and the international seed corporations.

On the philosophical/religious side, I believe that domestication is a covenant between humans and other species, wherein they support and feed us and we care for them. The care we owe other species is not just at the individual level but also at the population level. I feel good about growing seeds from small seed farmers who I trust to be caring for their seeds with respect for the sovereignty of that seed – I know many seed folk who view themselves as servants of their seeds, helping them spread. I also feel an obligation to enter into that relationship directly, to care for a population in partnership on this farm where we both now live. I am descended from immigrants and colonists, but I wish to move past that history and, as Wendell Berry put it, become native to this place. Likewise, I feel that my crops wish to dwell as natives where they are, not merely as seasonal migrants imported from the far corners of the globe annually. They are not wild species, they require cultivation to live, just as I am not a member of a wild species, and require culture for my support. A garden is our native home, and part of being native to a place is being a member of a community that lives there.

Breeding for Fun
We rarely follow recommended strict isolation distances when we are growing a seed crop for our own use. We have found that, practically speaking, if you grow a decent size patch of a variety, it will mostly cross within itself and not a neighbor variety planted several beds away. There are times when strict isolation is appropriate, but it’s not necessary if you don’t mind getting the odd off-type in the following year. It is impossible to ensure 100% purity in seed – there is always a slight chance that pollen will travel. Commercial seed should be 99% pure or better. I’m usually happy with 90%-95%, and that is a lot easier to get. Because we are commercial growers, we have enough plants in our fields that pretty much every year we get some interesting off-types of something. Even commercial seed will regularly give you off types if you grow enough plants and pay close attention. It’s usually not that hard to save seed from an individual that stands out, though there are some crops that we aren’t willing to take on the commitment of growing out for seed (like cabbage and carrots). Before professional breeders took over about a hundred years ago, that is how all breeding occurred – a farmer or gardener would notice an off type that appealed to them, and decide to save seed from it and see what happened. That is how most of our breeding projects begin.

Tomato breeding

photo courtesy Tevis Robertson-Goldberg
Tomato breeding in the hoophouse. Flagging tape on each plant with date of first harvest and tally of fruit picked so earliness and yield can be compared.

Tomatoes are a great “gateway drug” to seed saving and breeding. They are so easy, they are so tasty, and there are just so gosh darn many of them. Like any habit, they can get out of control – when we noticed that we were growing over a hundred varieties, we knew we had to cut back and let a few go. Because tomatoes are a self-pollinating crop, typically they don’t need or get isolated (unless being grown for a commercial seed crop). I have been growing and saving seed from some varieties for two decades, with no isolation, and haven’t seen an off-type yet. Other varieties, however, probably due to flower structure, seem to cross much more readily. If we were “good” seed savers, we might rogue out those pesky off-types so that they don’t contaminate our variety. Instead, being drawn to diversity, we tend to grow them out, sample them, and if we like them, save seed. Just a warning – breeding work can quickly take over a garden. Part of why our tomato collection got out of control was that we were saving every tasty new thing that came along. We had a Zebra rainbow descended from a green Zebra cross, with red, orange, yellow, white, brown, bronze, and black Zebras, as well as a solid yellow and a solid orange. Plus a pink Zebra from different parentage. Then there was a green Pineapple cross which parented another half dozen varieties. Eventually, we had to engage in the flip side of selection, and reject some of them.

Another fun project the bees got us into is breeding dry beans. Beans are, like tomatoes, an inbreeding crop that crosses just enough to keep life interesting. Several years back we started to grow dry beans for market, and it didn’t take long before we noticed some off-types showing up. Beans can be lovely little jewels, and their colors and patterns fascinate us. The genetics regulating pattern and color of bean seed coats are, to put it mildly, complex. There appear to be genes that regulate the production of the anthocyanins and carotenoids that give them their color, as well as genes regulating their patterns.

With some crops, it can be unclear whether the variation you see is genetic or environmental. Seed colors of beans are so strikingly different that it is often obvious when a cross has occurred.

Several of the interesting crosses that we have seen in our fields have happened despite isolation. Some crop genera have multiple species in cultivation, and so we usually will save our own seed from species that we are only growing one variety of. We grow too many Cucurbita pepo varieties to save seed from without hand pollination (which we haven’t gotten into yet), but we do save seed from Cucurbita moschata and Cucurbita maxima varieties. Being from different species, we don’t need to isolate those varieties from each other. Several years back, we noticed our first off-type. It stood out in the field at harvest time because it was the only squash plant in a quarter acre patch with healthy green leaves. Everybody else had succumbed to mildew and general end-of-season malaise. Unfortunately, when we cut open the fruit, the seed inside was not properly formed, just husks. Only six seeds seemed like they were filled, and none germinated the next year. We’ve seen that same C. moschata x maxima cross happen, independently, in our fields at least six times since, and gotten viable seeds that sprouted from some of them, but haven’t yet had success getting seed from the next generation.

While we haven’t gotten past the F1 generation of our interspecies squash, some of the other interspecies hybrids we’ve seen have had better viability. Brassica species (oleracea, napus, rapa, and juncea), in my experience, can make fertile hybrids fairly readily. B. napus is itself considered to be an interspecies cross between B. oleracea and B. rapa, and produces great confusion in nomenclature with varieties like Gilfeather Turnip and Macomber Turnip, both of which are technically considered Rutabagas despite the Turnip in their name, but technically technically they both originated as interspecies crosses between Rutabagas (napus) and Turnips (rapa) that fall into the B. napus range of the Brassica spectrum.

The interspecies crosses that we are currently playing with are Basils. A couple years ago we had a Tulsi plant that looked like it had crossed with Genovese Basil. We liked the flavor, especially as a tea, and when Basil Downy Mildew rolled through that plant was completely unaffected. We tried to backcross it to Genovese to create more of the classic Basil look and flavor, with limited success. Some professional breeders have now released a couple of BDM resistant Genovese-type Basils, so now our basil breeding has been freed from commercial considerations and we are able to just play with what our basil patch is throwing out these days. We’ve now got compound crosses involving Tulsi, Purple, Genovese, African, Lemon, and Lime basils. None are going to contend with Genovese for market share, but we sure do enjoy them. I’m not going to attempt to list scientific names here, as taxonomy within the Ocimum genus is unclear at best, and once you get outside of a few common varieties of Ocimum basilicum, sources often disagree on the proper nomenclature. I have seen the Tulsi we grow listed under at least four different scientific names. It likely falls outside official taxonomy, as it can (or at least has in our garden) cross with at least three different closely-related species. Taxonomy is a human construct, and nature doesn’t always fit neatly within the lines.

One advantage to breeding basil is that it can be evaluated as a seedling in the greenhouse. We do an initial visual and aromatic comparison to cull the first round, then plant them out arranged according to type. While Basil plants are very popular with the bees (especially the Tulsi crosses), they don’t cross as readily as we had supposed: the bulk of the crossing seems to happen within a pretty small radius (about six feet, though of course they can cross at distances up to two miles). Like with the Cucurbitas, many of these interspecies Ocimum crosses don’t produce much seed, but we have succeeded in getting past the first generation. With each passing generation, the genetics ought to settle out until they find a stable point where they are viable and either fertile with one of the parent species (functionally rejoining that species, while carrying in some DNA from the other species) or fertile within their own population but not with either parent species (functionally, a new species).

Breeding for Market Appeal
While playing around with interesting off-types is fun, and we enjoy following breeding paths laid out by bees and the plants themselves, sometimes we let commercial considerations guide breeding projects. Seventeen years ago, I decided that I wanted an open-pollinated version of the popular Sungold F1 cherry tomato. I was growing Sunsugar F1 at the time, which is very similar but with better split resistance. So I saved seed from it, and grew it out the next year to see what would happen. Over the next couple years, we decided that we weren’t interested in mimicking Sungold or Sunsugar, what we wanted was a really good cherry tomato with tropical fruit sweetness and good production characteristics. Eventually, we created a variety that we named Honeydrop, which we offered to Fedco Seeds for trialing and they commercially released. Some of the traits that we selected for were flavor, early ripening, split resistance, and disease resistance. We purposefully never engaged in single plant selection – at every generation we made sure that we had several plants in the seed lot. Our goal was to keep enough diversity in the strain that further selection could be done in different conditions. Our disease pressure changes from year to year, and we wanted broad resistance potential, not just really good resistance to a single disease.

During the early years of that breeding project, I kept rejecting “off types” that weren’t the orange cherry tomato I was going for. One recurring off type carried some recessive gene that resulted in pink fruit, and tended to have even better flavor. Eventually, I was convinced that this wasn’t an “off type” but another variety that wanted to exist in it’s own right, just as good or better than Honeydrop. My daughter, who helped convince me, named it Pink Princess, and as soon as we had it stabilized we offered it to Fedco, who released it commercially.

Both varieties have a higher tendency to cross than is typical for tomatoes. As a result, we’ve had a few new crosses happen over the years, and have a couple new varieties in the family in progress. These also have great commercial potential, but I’m going to wait until they are “finished” before releasing them to the world. Our market customers are our beta testers, and while we typically sell our cherry tomatoes as mixed boxes some of our regulars ask for single variety boxes. We pay attention to which varieties they ask for! While it seems to be a bit more susceptible to disease, my daughter’s selection Anona’s Sunset is also the favorite of several customers. Similarly, we pay attention to which of our larger tomatoes our customers select from our table. Orange Zebra is a variety we bred that tends to go into almost every bag of “mixed heirlooms” that a customer picks out. It is visually appealing, which makes the first sale, and it is delicious enough that regulars buy it again. I have been a bit surprised at the popularity of Zombie (it is small, green and fuzzy with a purplish bruise at the blossom end) as it is no beauty, but interesting can make the first sale also, and the flavor has won over a couple customers who fill their bag with only that, if we have enough that week.

We were once given a local strain of Buttercup that had been being saved by a local farmer, Harry Guyette, since the 1930s. The first year that we grew it out, we saw that the fruit of one plant had red streaks. In subsequent years, we determined that it must have been a mutation, not a cross, because in all other ways that strain is the same as the original green buttercup strain from Harry. We have kept both the red and the original green strains going since, because the red streaking on a dark green background sure does catch the eye at market.

Breeding for Resilience

Bean crosses all from Arikara Yellow

photo courtesy Tevis Robertson-Goldberg
Bean crosses all from Arikara Yellow

Several years back we started to breed our own strain of Butternut. Our farm conditions are marginal enough that it seemed likely that we could breed ourselves a more well adapted variety than is available “off the shelf” commercially. We decided to create our own Butternut Grex. “Grex” is a term that means “flock” and was brought into vegetable breeding circles by Alan Kapular. A grex is a diverse population from a cross or several crosses. Typically breeders will select a grex for certain production traits or local adaptability while leaving as much genetic variation as possible, to allow for continued adaptation. We got six different varieties, in a range of sizes, with a range of disease resistances, and grew them side by side. Then we saved seed from the best fruit. Each year at harvest we select the nicest looking field-ripened fruit and set aside a bushel or two for seed. We then usually forget about those bins until January or so, at which point in our storage conditions some rot has usually started.

Waiting this late to save seed allows us to select for storage ability – which is correlated to healthy, disease-free plants in the field and nutrient-dense fruits. The final test is to nibble a slice of each squash as it gets cut open. Some of the seed squash come back to the house and get cooked, and so we can assess flavor then, but sometimes we chop up a big batch in the barn. The raw flavor is different when cooked, but it’s a lot easier to compare a lot of different fruit by nibbling on raw slices than it is to cook them all separately at the same time.

Our butternut grex started with both hybrids and heirlooms. We included the hybrids because there are some valuable genes for disease resistance in them. But modern breeding of hybrids specifically selects against a crops ability to mutate or adapt. The first stage of creating a modern hybrid is to create parent lines. Because a key part of the sales appeal of hybrids is uniformity, breeders want their parent lines to be as genetically uniform as possible. Hybrids are not inherently uniform – they are only as uniform as their parent lines were. So in the breeding of parent lines, breeders inbreed as much as possible, and discard any lines that refuse to meet the breeder’s standards of uniformity. Any line prone to throwing “sports” is unacceptable. As a result, the genetic basis for adaptable, resilient crops is discarded.

We have used commercial hybrids as parent material on several occasions. Well-funded breeding programs can create some great varieties, and they are good at breeding in specific resistance to new diseases (like Late Blight in tomatoes). But I have also come to appreciate that the current dominance of hybrid varieties, and the tendency for breeders to build off of current varieties, means that many vegetable species may be losing some of their innate ability to mutate and adapt. Biotech scientists seem to be trying to overcome the narrowing of genetic diversity created by modern breeding practices by using genetic engineering and editing. I’d rather have crops that are naturally adaptable, rather than relying on scientists to adapt crops for me. So when I can, I’d rather throw an heirloom or two into any mix and not base a breeding project entirely on commercial hybrids. Ideally, use the heirloom as the mother line, since mitochondria and other non-nuclear DNA is only passed down maternally.

Breeding for Flavor and Nutrition
Flavor and nutrition are closely linked. Nutrients in food affect its flavor, and flavor compounds have nutritional benefits. Modern crops can be somewhat lacking on the flavor and nutrition fronts. There is a long history of breeding vegetables to be less flavorful. Mild and sweet is high praise for most vegetables and is what breeders have been working towards for hundreds if not thousands of years. In recent years, they have just gotten too good at breeding the flavor out of vegetables, and have gone too far. While few of us want our vegetables to taste like their wild cousins, there is definitely room to bring some of the flavor back into our food.

There are two main issues leading to the poor nutritional qualities of crops – soil fertility and the genetics of the crop itself. For maximum nutrient density, you need to address not just the soil fertility component, but also the genetic component. Most modern hybrid varieties have been bred to thrive in an industrial farming environment. Soluble fertility is added to meet the crop needs, microbial life in the soil is low, and flavor is rarely a main breeding objective. Most professional breeding is primarily focused on yield, disease resistance, appearance, storage and shipping qualities. Flavor is considered primarily in terms of “does this meet market standards” rather than “does this exceed market expectations.” If flavor is considered, generally sweetness is what is being looked for, not the subtler flavors that are associated with vitamin and mineral content.

Our fertility regime on our farm is based on cover crops and compost produced on farm. While we import some nutrients in livestock supplements, we’re mostly working with the minerals available in the dirt under our feet. We’re relying on the microbial life in that dirt to make minerals available for plant growth. Our crops often need to work with that microbial life in order to thrive. It’s my belief that in our system, a crop plant that is more generous with offering root exudates to its mycorrhizal neighbors is more likely to thrive than a more “selfish” plant that keeps all it’s sugars to itself, because it will have a stronger rhizobial neighborhood that will return it the minerals it needs and help protect it from disease (we also don’t use any fungicides). The selfish plant is at an advantage in a typical breeders system, because it keeps all it’s sugars and so grows faster and bigger and sweeter, and has the soluble nutrients it needs (often in excess) as well as, often enough, pesticides of various sorts to protect it. That selfish plant is likely to do poorly in a system like mine without a full complement of inputs. The further your garden conditions get from the industrial norm, the less likely it is that the varieties being bred by the seed industry will be a good fit for your garden.

I don’t have the means to directly assess which plants are good rhizobial neighbors. We rely on our senses for assessing the plants in our breeding projects. Our eyes can give us a good assessment of plant health – both looking for obvious signs of disease and also looking for subtle things that we can’t clearly identify, but that tell us that one plant is thriving while it’s neighbor lacks luster. We use our nose to assess aroma, and our tongues to assess flavor, and hence nutrient levels. We may not be able to assess every chemical compound that a mass spectrometer could, and certainly not with such precision, but our senses have evolved over millennia to tell us what we need to know about what is nutritious and what isn’t. Our tongues are great at detecting brix without needing a refractometer. Sure, the refractometer gives you an objective number that you can brag about, and feels all science-y, but if you’re comparing two plants in your garden, it’s pretty basic – the one that tastes better is going to give you more of the nutrients you need. Note that that is a subjective assessment. One person’s medicine is another’s poison. My wife likes the bitter compounds in Chicory, and I’m not a big fan.

This part of the breeding process ought to involve children, if possible. I’ve heard that kids have more tastebuds than adults, that the tongue loses some sensitivity as it ages, and that is part of the reason why kids tend to be picky eaters and reject strong flavors. I think that kids are generally better at honest assessments without letting preconceptions get in the way. Whatever the truth to that, when a kid determines that this plant right here has the best tasting cherry tomatoes in the whole patch, don’t argue. Save the seed. That’s the one.

Tevis Robertson-Goldberg farms with his wife and family at Crabapple Farm in Chesterfield, MA. They raise plants, vegetables, beef, lamb, and eggs that they sell year round at Farmers Markets in Northampton, MA.