The emergence of agriculture about 10,000 years ago was not planned and was completely the consequences of species interaction. The hunter-gatherers who became the first farmers were going around collecting wild seeds, especially of what have become the cereal grains­wheats, barleys. These people were collecting these seeds for a long time but at some point their technology developed to where they had the capacity to store food and store the seeds and somebody got the brilliant idea that they could plant these seeds closer to where they lived and then they wouldn't have to wander so far to collect them. Right when they planted stored seeds things began to change. Prior to that they weren't harvesting seed from what they ate, they were harvesting from seed that dropped on the ground and escaped the harvest. The seeds that were resprouting naturally in the wild were doing what they had always done, dropping to the ground under the selective pressure of the ecology. Once people stored the seed and then planted it everything started changing because now they were collecting seeds from plants that did not shatter. Suddenly there was a selective pressure in the surviving plants for non-shattering. People also began to select for clustering of the seed at the ends of the branches. Quinoa is an example of what lambs quarter (Chenopodium) turned into.

In North America the Indians were using a different Chenopodium, Chenopodium berlandearia. They were using it as a food crop in Central America and southern North America, making ground meal out of it. In lambs quarters the seeds are very evenly distributed along the stems throughout the plants. On a still-existing wild variety of Chenopodium berlandearia the seeds are distributed throughout the plant. If you go to Central America and see a version of Chenopodium berlandearia there called huauzontle, all the seeds are clustered at the tips of the branches just like quinoa. Non-shattering and clustering seed heads are the two things that began to emerge first in what eventually evolved as our agriculture.

Most of the plants that we use agriculturally are pioneer species, quick movers that get to the flood plain, landslide, fireburn site first and grow very quickly. Seeds are usually wind-blown or hitch rides on animals or birds, germinate quickly, grow fast and grow a lot of biomass. They take advantage of nutrients that are not being claimed by any perennial plants so they don't have extensive root systems. In some cases they get taproots. The species we rely on for food are herbaceous, grow quickly, produce a lot of biomass that is easily digestible, not a lot of lignin. Pioneer plant communities are characterized by annual plants, a lack of shelter, a population very transitory and forming a monoculture (eg wind-blown dandelions or fireweed). All of the photosynthetic energy goes into making this fast-growing plant body and nutritious seeds, not into a perennial root or woody tissues. Over time we developed these agricultural crops and they diversified. A certain diversification happened in nature then we took those characteristics and isolated them within garden systems so that their genetics were isolated from the wild plants and so we created changes in genetics that weren't being swamped by the wild pollen. All this happened completely unconsciously. It was a self-organizing system.

Every time that you save seeds you are affecting the genetics of the seed. Those seeds are a reflection of the growing season that produced them. If it's a droughty year and you saved seed from a mixed population you're going to tend to get genetics that are good for drought resistance. If it's a wet year and you collect seed from a mixed population the ones that do best under wet conditions are going to be disproportionate in the population. Just as the plants change themselves to fit any situation in a natural environment, they also change in response to our artificial environment. Adaptation and evolution happens in every generation; there are no time outs.

Q: If you have a plot and one end is clay soil, another is sand, with a variety of soils between, then you took all the seed from that area, bulked it for a progeny test, would that work?
A: In principle, yes, it would work to maintain genetic diversity for the plant's ability to grow in those soil types. But as a member of the system you would make it work a lot faster if you went into those plots and chose the best plants from each. If you choose all the plants then you will have some representatives that are so-so. You create fast changes by controlling and isolating the genetics by selection. If you are doing it all en masse you are effecting changes but they are much more gradual.
Q: If you select the best plants in those groups separately will that work?
A: If you are trying to save them for a particular site, yes. If you are trying to preserve diversity for the sake of diversity, if you're trying to preserve the entire genetic spectrum of a variety, then you want to save seed from as many plants growing on as many soil types that are doing well as possible. In order to maintain the quality of the genetics you need to remove the weak ones constantly. John calls this throwing out the uglies. They are not really that ugly, they are just not performing that well. Its like the bell curve. I whack it off here and take 40% off the bad end. If you do that and bulk it together you're making progress. But if you do the same sort of thing in a progeny testing method you have so much more control. If you do mass selection in a field you are doing maternal selection, you are selecting only on the basis of the mama's characteristics. But if you're not doing anything to control the male parentage and you are letting weak ones cross back in you're not making progress and you may be breaking off your best genetic linkages. So, cut off 20-40% of the worst performing populations regardless of what you're selecting for.

What is a variety? It is easier to say in a self-pollinating plant such as a lettuce than in a crosser such as kale. In a kale the genes that are together in any one plant is an ephemeral event. Pollen's going to come in and whatever combinations produced the most beautiful lacinato kale you ever saw, you can plant that the next time and it doesn't look like that at all. If you take the seed from that beautiful plant you're going to get a range of qualities. Some may look beautiful like that and some a little bit different and some a lot different. In a cross-pollinated species I've begun to think the concept of variety is a passing thing. As Lacinato gets moved from Italy to Maine the genetic combinations that survive are going to be different. How different the combinations would have to be to be a striver in Italy vs. one in Maine. It might look the same but what's underneath that look, the frost tolerance, the disease tolerance, flavor, are all changing.

Q: I used to grow Winter Density from Johnny's, then the rule came out and I wanted organic seed so I got it from Seeds of Change, grew it the same way, and I got a different lettuce with the same name.
A: That's not necessarily a genetic switch.
Q: Is it because the seed was grown in different locations?
A: If it really doesn't look the same, then there was a cross. It's not the environment that changes the lettuce.
Q: Looks the same, doesn't taste the same, doesn't have the same cold tolerance.
A; I've been growing Winter Density, too, from Johnny's and you caught me right before I wanted to disavow my statement. Self-pollinated crops do change but much more slowly. I grew Winter Density from Abundant Life and it wasn't the Winter Density that I remembered from the 80s. It wasn't as good as I remembered, but I was growing it under pressure in my trials where the environment was vastly different. If you grow out side-by-side the variety from 3 different sources you may find that they look the same or that they look different, because selfers do change, albeit more slowly. If you had a pure-line Winter Density bred to be stable, the rate of change is relatively slow because you can get only 5% crossing max. After 4 or 5 generations you could see some pretty big changes in adaptation to its environment if you were selecting, but not if you let everything reproduce. The genetics are changing themselves because those that are better fitted make more seeds, those worse fitted make fewer seeds. There may still be some of the same combinations in 2003 as there were in 1980 of the traits we can't see such as disease resistance. That could have drifted if one batch were maintained without fungicides for 23 years and the other were maintained with fungicides.
Q: Is it possible for amateurs to experiment with crossing lettuces?
A; Sure.

Cross-pollinated plants can change rapidly. In three generations you could take something that was beautiful and have something that's awful if you allow off-types to reproduce. But if you did the same thing with lettuce it would be relatively much more stable. Over time, varieties are constantly changing because the environment is constantly changing. Our world is different from Vilmorin's in 1885. Varieties are not something the aliens delivered to us that we have to do everything possible to keep just as they are.

You need to be watching the variations that are coming into the varieties, whether you're inducing them on purpose or whether they are happening accidentally. I pay attention to all of them because every variation is an opportunity, maybe to have something better than it was before. Uniformity is not my God, diversity is. I want elasticity, I want the plants to be able to change in response to changing environments. I don't have a template for what Lacinato ought to be. I'm working with Lacinato intensively because I think it's an interesting plant and people like to eat it. It has been inbred, I have been outbreeding. In 2000 I let a little cabbage cross into Lacinato and then I began trying to get back to Lacinato as fast as possible. I didn't want those changes to be dramatic, I didn't want the glaucousness, there are a lot of things I did not want to become part of the population, so I allowed one generation of open-pollination and then immediately began to clean up the population again. Now 2003 Lacinato looks just like it did when I first received it, but more vigorous with a lot more leaves, so I have changed something in the Lacinato. I haven't changed the Lacinato look but I have made this Lacinato more elastic and better for my environment. Phenotype is the physical characteristics that you can observe. The genotype is the actual genes on the chromosomes. I've changed the genotype but not the phenotype. The genotype in a crosser changes in every generation but not necessarily the phenotype.
Q: You said diversity is your guideline but I remember some times looking at crops with you where you said, "See, I can make something that's uniform."
A: I'm always dismissing bowing down to uniformity, but I love to show somebody who appreciates uniformity that I can do it, too.
Q: There's certainly a level of uniformity that is desirable for certain situations commercial quality.
A: Definitely.

We want it elastic, flexible so when the wind blows it doesn't fall over, when bad things happen it doesn't get killed. When I'm breeding I like to find the potential in every situation. Rob Johnston once told me, "When you look at something, you don't see what it is, you see what it could do. He then drew a contrast, interestingly, "When I look at something I look at what it is, I judge." That's fine, that's what we need, it takes diversity in humans, too. He's looking at something he's gonna sell to somebody, he's making a judgment on what I've been working on. I'm looking at the potential and he's judging whether I've met it.

It's really good to walk into a diseased field and look for the potential. Before you harvest, you need to go through the patch to see what's happening. You're gonna want to throw out sick stuff, but you've got to be looking closely. When I go through my lettuce patch and I see a plant with bad downy mildew here and here and here and right in the middle is a plant with some mildew but looks a whole lot better than the rest, that is seeing good potential in a bad situation. That plant is far more auspicious in its potential than if I go further down the patch where there's no downy mildew, and take seed from all those beautiful unchallenged plants. You don't know what they would look like if a little downy got on them. You need to look for the good potential in a bad situation. That will move you ahead faster than anything else.

Q; Do you rogue out the bad ones or do you leave them there?
A: For what? When I'm breeding or when I'm selling seeds? This is actually a business decision. As a breeder I will take out everything from 4 down and keep the rest. But when I'm selling seed, I throw out the worst ones but I don't throw out 40% of the patch or I wouldn't make any money. I assume the seed I'm selling is good enough or people wouldn't be buying it and in that patch improvement is not most important. If I'm producing 40# of seed, I don't want to throw out 40%.
Comment: "You already did the improving."
A: Yes, hopefully you've done the improving before you get to the production spot. But I do these things together oftentimes. In kale last year I threw out 60% of something that I'm actually selling because I could see the stuff wasn't good enough to sell. That's big. You just knocked out your profit margin. Really you should be doing all that intensive rogueing in the 40-50% range in your breeding patch. 10% is my usual rogueing target for seed production.

Novelty is always arising. That's great. You don't always have to generate new forms. They just come up. But if you want to generate new forms you cross two varieties. How do you decide which varieties that you want to cross? To induce variation you do crossing but you don't just want to cross anything to anything. You want to think about your end goal and you want to look for a couple of parents (or possibly three or four) that have qualities. You need to be able to look at those plants and know what the traits are. In tatsoi a characteristic is dark green, another is round leaves, another is broad midrib. In mizuna the characteristics are dissected leaf, light green, round thin stems, all the opposite traits of tatsoi. So when I was thinking of what would be interesting, cross tatsoi and mizuna because its like having a really diverse deck of cards and shuffling it and every time you shuffle it is a new generation.

If you're looking for really new stuff to be adapted to really new conditions such as organic agriculture, you probably want to choose parents that have good characteristics but are very divergent because taking divergent points and putting them together is interesting. You'll get the biggest spread of possibilities, you'll have the most to choose from because you've put in the traits curliness, roundness, disectness all together. What else is missing from this in terms of being cool?
Audience: color!
A: Yah, man, you've got it, it's all green! Wouldn't it be good to have red and purple in your life? What if you had a purple tatsoi? A purple mizuna?
How 'bout stem color? What about purple stems? Those two both have white stems.
If you want a purple pac choi, that has purple midrib, but that purple pigment never goes into the leaf blade. You can cross stuff to purple pac choi all day and you will never get the blade of the leaf to be anything but green. But you can get a dark purple stem and I've worked on that.

Now I'm going to tell you a real secret, something plant breeders never do. There is a Brassica rapa that does put color into the blade of the leaf. This is a gift because it is going to produce some of the most dramatic salad greens you ever saw. Fedco helped me find the germplasm a few years ago for Scarlet Ohno Turnip from Japan. You could buy these in the mid-eighties. They're traditionally used as a pickling turnip, sort of like a flat onion. A few years ago Tim Peters told me you can cross the Scarlet Ohno, the pigment expression is in the blade of the leaf, and it will even express without light. That blew my mind because the purple in a purple pac choi stem only becomes purple in response to strong sunlight. In a population of Scarlet Ohno about 25% of the plants infuse the bright scarlet red pigment in the young leaves.

Now I'm proposing a three-variety cross. In addition to the color all the turnip characteristics have been thrown into the mix including long strapped leaves, with blade all the way to face, and a food storage root. Tatsoi with a turnip underneath it? Would that work? I don't know. There's some limit because plants only make so much photosynthate and its gonna put that photosynthate someplace, maybe in a broad petiole or maybe in the root. If you try to develop the root on the tatsoi you may be left with leaves that have no succulence. What's this thing going to look like? What traits? Hairiness could pop out, but unlike purple top turnips Scarlet Ohno has virtually naked leaves. Hairiness may nevertheless jump out because it is a dominant trait, it will really happen fast and if you're growing salad greens you don't want hairs on them. We begin selecting against that very early.

We proposed that we take a project we were discussing yesterday, crossing tatsoi and mizuna, and put in Scarlet Ohno and make it a 3-way cross. It makes things a little more complicated and you might not get where you want to go as fast because all these linkages are being broken up and only two plants can cross at any one time. You can't get the genes of three plants into the seed in one generation.

You can make them flower at the same time. You just have to plan. Scarlet Ohno is not necessarily biennial. If you start it in January or February it'll vernalize enough to bolt the first year. Everybody's going to have to figure out on your own how to make all these things nick, in the business, to nick is to bloom simultaneously so you can get crossing. That's not always easy.

Q: Does the pollen have to be fresh?
A: I assume the pollen has to be fresh.
Tom: 200 meters away from the Brassica patch on the back hip of the bee the pollen is dead.
A: Pollen tends to dry out.
We're proposing doing this 3-way cross. Eli would like me to mentor to some extent. I don't have a lot of time and I'm not gonna hold anybody's hand. I've already given you a bunch of interesting information and I guarantee that the plants that come out of this cross will be dramatic. You're gonna be motivated to keep doing this because you just can't wait to see what happens next time.

If it were me, I'm gonna grow all three in a population. I'm gonna start with at least 40 of each plant. These are all small and you can do them in a small space. (Serge suggested a way to do it which I couldn't follow.) Frank: You could do that if you were organized and had the time. He has the mentality for growing towards progeny testing. Somebody else in the room may not be patient enough. There are other ways to do this. You could take one plant of each, put them together, let them pollinate, save the seed of each separately. You end up with the same information because you take all the seed off the turnip and you'd see this is the turnip crossed with the tatsoi and that is the turnip crossed with the mizuna. You'll see it, it's obvious.
Tom: They are self-mcompatible enough not to self?
A: You will see some selfers. You'll know. If you do them in seed flats you can sort them all out. Your genetics is limited to three plants and if any one of them contains recessive deleterious genes then you've got trouble. Another way to do it is plant three rows side-by-side, collect the seed from each row separately, put those in their own bags.
There's a lot of genetic variation within the one variety of turnip.
I've done the tatsoi x mizuna cross. It equals Mitzpoona Salad Select that you can buy from Fedco. It's a genetic mix that's been selected for ten years to be a diverse population. Everything in it is good but no two plants are exactly the same. When you do this cross you're going to see that there are some dominant phenotypes (a lot of the plants are going to look very similar. Those are groupings of dominant traits). Every now and then you're going to see a rare type, something that looks like a tatsoi with a very savoyed leaf, but you might see only 1 out of 1,000 or more likely 20 out of 1,000.

I am not precisely sure how we are to organize this working group. This is part of the team-building concept. All you folks are in the same eco-zone and I assume you're gonna self-organize. I'm not going to organize it. I will mentor it, I will provide seed, I will talk to you on the phone if you call me up and describe to me what you're seeing in your experiment.

Q: What about isolations?
A: In my breeding garden I have Brassica rapa patches or groups of 50 or 100 plants. You can minimize distances dramatically by having a row of corn in between, by doing it on a different side of your house, putting it right up against the house so it is only approachable from one side. If you start thinking like a honeybee Honeybees are very organized in how they do their work. A particular worker keeps going back to the same patch until that patch is done or until he's told there are better pickings somewhere else in which case he/she will change his way. You'd be amazed at how limited the crossing is.
F: I think the wave of the future, the next frontier in plant breeding, is going to be breeding for symbiosis. Symbiosis is a highly evolved stage and highly efficient and the cracks the symbionts fit into are amazing.