Tomatoes are a high-value crop for many farms and having tomato varieties that match the farm site and the local markets is of critical importance to many of us who farm. Fortunately, there are many varieties we can choose from today, and trialing new varieties that become available is an important way to find varieties with good production and high value. Additionally, we can also fine-tune what we grow on our farms using simple breeding practices. Here I provide some simple advice on how to select and create varieties suited to your farm.


The basics of tomato breeding

Tomato breeding can be characterized as two major endeavors – making crosses and selecting for desired traits in variable tomato populations.

Once you know how to recognize basic stages of flower development, and learn how to emasculate tomato flowers, making crosses is simple and pretty straightforward. A genetic cross requires taking pollen from one plant and physically depositing it on the flower of another plant. Thus, the fruit that develops has two different parents.

Stating the fact that the resulting cross has two different parents may sound obvious, however tomatoes and many other crops naturally tend to inbreed. This means that if you take the seeds out of any random tomato, chances are that the seeds are the result of self-pollination from the plant it came from, and that those seeds did NOT have two different parents.

If you want to make crosses with two different parents, and (especially important for breeding) know what the parents were, you’ve got to make the crosses yourself. The female parent is the one with the flower where the pollen was deposited. The male parent is the one which provided the pollen.

The seeds produced in this way are “hybrid” and have one copy of genetic material from each parent. In breeding terminology, the first generation after a cross of two genetically different parents is called “F1,” or “first filial.” Subsequent generations grown out from the same cross are referred to as F2, F3, etc.

Selections from the initial cross (F1) typically stabilize around the seventh or eighth generations (F7 or F8), but if you are trying to develop open pollinated varieties, you can’t simply wait for the seventh or eighth generation. You have to grow the variety out until it is visibly stable i.e. all the progeny from a particular generation are similar.

In subsequent generations after the F1, the genetic material from the two parents is scrambled, and starting with the next generation after the hybrid generation (the F2 generation) the traits of both parents are assembled differently in each plant in the population. In other words, after the hybrid generation (which is uniform) subsequent generations contain a diverse array of trait combinations.




With each subsequent generation, if seed is saved from individual plants, the variation is reduced, and the next generation becomes more uniform. While most people focus closely on variation in the F2 generation, it is critically important to continue to select from populations with multiple plants (as many as are practical) in the F3, F4, F5 and F6 generations as well. By approximately the F7 or F8 generation, if lines in each generation represent seeds saved from individual plants, the lines become uniform and “stable.” At this point a new true-breeding (often called “open-pollinated”) variety has been created.


breeding-tomatoes-farm-practical-selection-adviceSome of the author’s variety trials under “challenging” outdoor conditions in Todos Santos, Baja California Sur, Mexico.


Selection is breeding

I think it is likely that most amateurs who want to breed new tomato varieties focus too much on making crosses and too little on carefully selecting from the variation in their gardens or fields. For gardeners, an inadequate focus on selection is understandable and can limit selection opportunity. They often don’t have the space to grow the larger populations of F2, F3 and F4 plants needed to find the most valuable combinations of traits. Farmers, however, have the power to grow larger populations of tomatoes, making it is easier to do better selections in the creation of new varieties.

Importantly, you don’t have to make a single cross to breed tomatoes on your farm. In fact, many exceptional new, and improved, varieties are created simply through the careful selection of variants by gardeners and farmers. The variety Speckled Roman is one example of the power of selection.

It was created by Seed Savers Exchange member John Swenson, from an inadvertent cross between Antique Roman and Banana Legs that occurred in his garden. Although an insect probably made the cross for him, he did the equally important work of careful selection for multiple generations to create Speckled Roman.

Furthermore, numerous farmers worldwide know the value of saving the seed from the best plant of each variety. Although there is no way of knowing whether a slightly different (and better) plant, in a row of plants expected to be identical, is the result of subtle environmental or genetic differences, savvy farmers know that genetic mutations are normal, and that sometimes they result in heritable improvement.


Selecting Maglia Rosa, our first release

We have been breeding tomatoes as a family since 2002. Initially we worked primarily in a large community garden plot, but since 2006 our primary breeding site has been our small farm in Sunol, California.

The first tomato we generally released was Maglia Rosa – a pink elongated cherry tomato with yellow stripes. Maglia Rosa was developed from a Black Cherry X Speckled Roman cross. The sizes of the F2, F3 and F4 populations we selected from for Maglia Rosa were much too small to capture all of the diversity of trait mixing possible in those generations.

Though at the time, we were limited in planting space, as almost all of the Maglia Rosa selections occurred prior to farming. We were focused primarily on flavor and fruit aesthetic, and Maglia Rosa represents perhaps the best/luckiest result of our early breeding. We made many other crosses at that time, and very few of those crosses have contributed anything of value to our ongoing efforts.

One other note about Maglia Rosa selection: Maglia Rosa has a clean/bright flavor, and I think this is partly because the taste panel at the time was myself and my kids (4 and 5 years old at the time). Kids have very sensitive taste buds, and so it seems to me they can be very useful in screening against “off” or bitter flavors. Just watch out for child labor laws.


breeding-tomatoes-farm-practical-selection-adviceBlush, Fred’s breakthrough tomato, a striped cherry with interior marbling (final stabilized version). Blush flavor can be traced to one selectee plant in 500, in the F3 generation. Most of Fred’s varieties reach peak flavor before they are at peak color. When they are at their deepest colors, they are already too soft. For growers who pick their tomatoes a few days before selling them, this may mean picking them with green stripes still showing, and the final ripe color only partially developed. Don’t worry! They will reach their full flavor potential as long as they are picked with partial ripe coloration. When you factor in a few days after picking, and a few days after being sold before being eaten, the flavor will be peaking as people are eating them.


Selecting Blush, our breakthrough

Blush was our breakthrough in flavor. It Is a yellow marbled cherry tomato with yellow stripes that appear as the marbling appears late in fruit development. Blush’s parents (Maglia Rosa and Yellow Fire) were chosen by my son Alex, who was eight years old. He chose them simply because they were his two favorite tomatoes at the time.

Unlike with Maglia Rosa, all of the selections for Blush occurred after we started farming. We originally saved seeds from a few F2 plants, from a population that was much larger than the one Maglia Rosa was originally selected from. In the F3 generation, we grew over 500 plants from seed saved from the one F2 plant that Alex had selected that looked and tasted pretty much like the finished Blush.

While the F2 plant had fruits that were flavorful, there was one F3 plant (again, out of 500) that was better than all the rest. In other words, through brute-force selection we managed to improve the flavor between the F2 and F3 generations. This lesson was not lost on us, and we have continued to grow F2, F3 and F4 generations that are as large as practical (at least 20 plants when we can), although we may never grow an F3 population of 500 plants again.

That very large number, while quite useful, was in part due to the fact that the Blush F2 had flavor that made it our highest priority project at the time. And also due to the fact that our breeding was in its infancy, and there were not many competing projects. Furthermore, we had a pretty good idea that all of those F3 fruits could be easily sold to excited customers at our farmers market stand, and they were.


breeding-tomatoes-farm-practical-selection-adviceAbove are tomato selection (breeding) rows in late June 2010 at Green Bee Farm in Sunol, CA. Selection is typically done under the same conditions used for tomato production, and saleable fruits from breeding rows are sold in mixed boxes. All photos courtesy of the author.


Population size conclusions

It is hard to give recommendations for population sizes, but here are some rules of thumb that we have developed over time for use when selecting for new varieties:

Don’t neglect the F3, F4 and F5 generations. Variation is not “captured” at the F2 generation, and there is still plenty of trait segregation that occurs in following generations.

Grow at least 10-15 plants for each F2 population, and also for each F3, F4 and F5 population. Grow more if you can, or if the F2 populations contains very interesting trait combinations.

These rules apply to dehybridizing hybrids. If you find a hybrid you like in the store, or you have a hybrid that you like to grow, you can collect seeds and select for combinations of traits you like in subsequent generations. You won’t get something exactly like the hybrid, but you might get something that suits you, and your farm, well.

These “rules” apply to inadvertent crosses too. Treat unexpected variants as accidental crosses, until they behave otherwise. If you see a variant that is potentially valuable, and potentially the result of an inadvertent cross, treat it as such. Plant 10-15 plants the next generation and see if multiple traits segregate, like they would if you had done a specific cross by your own hand.


breeding-tomatoes-farm-practical-selection-adviceAt left is one of Fred’s striped cherry tomato varieties, Pink Bumblebee. Photo by farmer Davide Adamo outside of Naples Italy


The importance of trialing

If you are spending precious time and space for breeding projects, you will want to know you are not wasting your time working on a specific type of tomato, when there are better varieties like that already out there. This is particularly important for farmers. While gardeners may have the resources to breed simply for curiosity, most of us farmers do not. That is why it is important to trial available varieties similar to the ones you are developing. You need to know how your breeding project tomatoes are comparing to the best of what is out there, because novelty only goes so far in selling tomatoes.

Furthermore, if your variety compares favorably to similar varieties, even if it is not useful to you in your garden or field, it may be useful to others. It may also be worthwhile investigating if there are opportunities for selling seeds, or plants. At that point, if you have trialed similar varieties, you will know better how to market your new variety.


Final thoughts on breeding with partners

When developing your unique breeding vision, and while you go through the initial steps of selecting some new varieties – from your crosses, from hybrids or from inadvertent crosses – I think it is best to work alone, and to develop a unique vision, before setting up collaborations with others. While enticing others to grow out populations for can be a good idea, I have found that it muddles things in the initial stages of breeding.

However, once you have developed your own vision, and you have created some lines that you know are potentially valuable, expanding your breeding efforts by including others is often desirable. It may be particularly valuable to develop relationships with folks who can test your new varieties in other climates or under other growth conditions.


Fred Hempel is a first generation farmer. He grows produce and breeds tomatoes on a small farm (Green Bee Farm) in Sunol, CA. Tomatoes he develops are sold via Artisan Seeds ( and Johnny’s Selected Seeds.