Breeding New Fruit Tree Landraces

Creating this topic to house the discussion of landraces started on the topic sharafuga-peach-x-plum-x-apricot-hybrid

@Justin, allow me a few minutes to respond to your points.


Our jumping off point.

1 Like

This is extraordinarily unlikely for several reasons. Like, beyond extraordinarily unlikely. (1) lemons are one of the least cold hardy citrus, any freezing will severely damage them and more than a few degrees below freezing kills them dead. They have no genes for cold resistance. Breeding lemons for cold hardiness is like breeding cacti to be aquatic plants–the genes just aren’t there, no amount of rearraigning or small chance mutation will fix that. (2) lemons, like many citrus, are highly true-to-seed. In this case, by true-to-seed, I mean literally clones. Many citrus, including virtually all edible citrus, produce a majority of seeds that are not the result of sexual reproduction, but of asexual cloning via apomixis. Even when cross-pollinated, most edible citrus will produce seeds that are mostly 100% genetic clones of the mother plant. Planting thousands of lemon seeds will get you thousands of the same lemon, and maybe a dozen actually sexually reproduced child plants. (3) there are several hobbyists breeding cold hardy citrus–ilya, walt, and kumin on the TropicalFruitForum being good examples of some who are currently active. None of them are simply relying on large populations of OP seed “adapting” to gradually colder and colder temperatures. All of them are relying on using hybrids of cold hardy but inedible wild citrus that were developed originally as rootstocks and are doing controlled backcrossing and very painstaking selection of the seedlings. The cold hardiness comes from the wild, inedible citrus (mostly poncirus, but also sour mandarins and some kumquat and ichang papeda, and some folks have talked about using Australian desert lime as well).


So, one thing that’s I need to get off my chest is, he could plant pretty much any decent cultivar of plum with the right chill requirements and they’d be just as well adapted as the ones from his seed grown population. Aside from some summer drought, which arguably is good for most stone fruit anyway, it sounds like his island location is a pretty nice place to grow plums. There isn’t much benefit to having a lot of adaptability in a cultivated plum population there, because there’s not much to adapt to. And, more importantly, a population that’s inbred long enough to become a landrace is going to be pretty darn genetically uniform. So if he were to breed his plums and plant their seed until he got a stable landrace, because there was so little disease, environmental or other negative pressure on his population, that resulting landrace would not have much genetic “room” for adapting to a new negative pressure were it to show up.

On the flip side, I live in the coastal plains in the South, and here virtually all Euro and Japanese plums will die after a few years unless drenched in chemicals. The only trees that survive well enough that you could maintain a population of are the native plums and their hybrids, and even a lot of the hybrids don’t do well. But, here’s the thing, fruit quality, compared to the plums that have been domesticated for a thousand years, sucks. So if I were to just OP and plant seeds for a few generations, I’d end up with plums that will survive here, sure, but they wouldn’t be worth growing because the fruit aren’t really worth eating. Here, the disease and environmental pressure on the population is too high. Anything that survives being bred with itself enough to create a landrace here would be all but guaranteed to have lost any genes for fruit quality along the way.

It is already extremely difficult just to breed hybrid plums with the combination of disease resistance and acceptable fruit quality to be worth growing in the South (without tons of spraying), so why make life so much harder by adding an even steeper requirement?

Again, it’s the process of pushing the population into enough uniformity that it becomes a landrace that’s causing the loss of genetic variation. So why push for a landrace, when we have techniques that are thousands of years old that work way better?

Landraces make sense for annuals, sure, but for fruit trees is seems a vast waste of resources. Better to curate populations that have a lot of wide crosses in their parentage as well as careful backcrossing, that way you can actually pull off those magic and rare combinations of traits that make a fruit tree worth growing. Which, again, is what professional and most competent hobbyist breeders do. And when they do get that one-in-a-thousand plant with all the right traits (a plant that would have been one-in-a-million in a natural OP population), they save it, clone it, and spread it around for all to benefit from.


And if I’m being honest, the native ones don’t do that well either half the time. They get by, well enough to maintain their wild populations, but most of them look pretty terrible out in the wild, with black knot and other disease issues being very common. The wild ones are just good enough at getting spread around by birds and other critters and at forming thickets of root suckers that they can persist in the wild. But those are hardly qualities you’d select for for something you’d want to grow and eat yourself.

I know “native” has great connotations. It’s mother nature’s doing, it’s natural, it’s biodiverse or whatever. But native also means unimproved. Which is to say, barely edible at best in most cases. I like my plums artificial, and I’d be willing to bet money you do too.


Kind of busy right now (grape harvest has started) so I can’t respond to everything yet, but I just wanted to address this point.

The initial total genetic variation within a hundred unrelated individuals is the same as exists in an interbreeding population generated from those 100 individuals. It doesn’t matter how big that population is, there will never be more genetic variation than originally existed in the parents, outside of rare chance mutations. Let’s say you let that population interbreed for 100 years, selecting for drought tolerance. After 100 years of selection, unless you maintain the original parents somewhere (which, given you can clone trees virtually forever isn’t hard to do), you end up losing genetic diversity, since invariably, some combinations of alleles will prove to be less beneficial under selection conditions and some will be lost by chance.

Adapting a population by definition means you are selecting against traits that aren’t useful or desirable for local conditions; i.e. you lose genetic diversity. I’m all for creating landraces, but I’m arguing for preserving whatever cultivars we can. Many of them are the result of thousands of years of selection followed by decades or centuries of vegetative propagation. Cultivars, especially heirloom ones, represent a part of our cultural heritage that we won’t ever get back if lost, even if some of their genes live on in a landrace.


Thank you.

1 Like

To paraphrase Burbank, he said he wasn’t creating any new traits, he was only rearranging already existing ones…


Well bear in mind I cannot remember what species it actually was. I thought it was some kind of citrus, maybe lemons or oranges or something similar, but I don’t remember. My apologies if I got it wrong. I spent ages just now trying to search for you but with no luck. If I find out I’ll let you know.

Anyway, it’s the principle which is important. Planting many genetically varied seeds or seedlings in a new environment, is an excellent way of extending the range of a species. And the greater the variation, the greater the chances of success. (That is why I am creating hybrid swarms indoors for speed before planting the resulting diverse seed outside, for example).

Yes, I have some awareness of such projects and it sounds great! I do also wonder though, depending on how far out of the range one wanted to stretch it, perhaps diverse seed from already delicious fruit such as yuzu, from as far North in Japan as one could find seeds, might also be a good method to try. Though yes if one wanted to maximise how far to stretch the range, it would make sense to cross with even cold hardier varieties such as some of the disgusting ones.

That doesn’t detract from the value of his report, which was that he had equal and superior fruits and tree structure from his seed-grown trees compared to their parents. Neither does it detract from the idea of a landrace adapting to a place and/or cultivation method to which no available commercial variety is pleasingly suited, nor to a landrace which can continue to adapt to changing conditions.

I believe this has already been discussed. But if we are to deepen into it, then I will ask you, what is the minimum population size to avoid becoming “pretty darn genetically uniform” such that they would no longer be adaptable? As I mentioned before, for SI tomatoes, the number is seen to be 75 for long term sustainability of an accession.

Sounds like you’d benefit from someone starting a landrace project in your area!

That seems to me a rather big assumption. Since you cannot know what would result from several generations of growing a hybrid swarm, what makes you so sure none of them would be worth eating?

I feel as if this question has been asked of me quite a few times already here and I feel I have answered it as many times. This time I will say: for yourself if you live long enough, and more importantly, for other people in future. But anyway I am in no way trying to convince you nor anyone else to start a landrace project. So if you don’t want to, then don’t - no problem!

And I will say again as I have so many times already, the only ‘uniformity’ I am talking about is that the population is relatively reliably delicious, and relatively fit. There could still be plenty of phenotypic diversity outside of those parameters. I was quite deliberate in qualifying that. And by the way, SI tomato species can have very uniform populations in terms of phenotype while still having more genetic diversity in a single accession than the entire planet’s heirloom tomato population.

Please see my many answers to this question in my previous comments.

And yet people have been doing it for thousands of years and are still doing it today.

If you think that is better then that would be the method for you! I have consistently acknowledged the benefits of that method. I am curious that you on the other hand seem to think that it should be the only method.

Yes, sure. That doesn’t have to be a problem. Especially if you start with a great excess of diversity, which is anyway the fastest way of creating a new landrace. And by limiting the number of traits you chose to stabilise, you can maintain a good degree of diversity even after the initial period of settling to the new conditions.

That sounds good to me too! Though my point was that in an OP population, the diversity is dynamic, whereas in a diverse clonal population, the diversity is static, hence non-adaptive.

Indeed! For example there are thousands of rice landraces in India (even more in Laos!) but so many have I think already been lost and so many more in danger of being lost, because of colonisation by foreign seed corporations, selling seeds that are nutritionally inferior and whose seeds can’t be saved for replanting even. Even in Europe we have lost so many landraces. It is a global problem. One of the (well, in most cases the) best ways to keep the diversity, is for people to simply grow them! Now in the case of fruit trees, your cloning method sounds very useful for that too. There is even a man in South India who has saved many local mango varieties by grafting the last remaining branch in some cases, onto his own large mango tree. I think he has … hundreds of varieties on his one tree! Ah yes, here it is - over 300 varieties on one tree!

So I agree that is valuable. But that does not lessen the value of growing the actual landraces. Nor the value of creating new landraces. I did give an example of plum landraces with excellent results, in a previous comment.


Landraces, by horticultural definition, are established after the fact. Your use of the word is incorrect.


I welcome you to elaborate, since I do not understand what you’re trying to communicate.


Apricots and citrus have both been planted by thousands in hopes of adapting to a colder climate. Citrus via crosses with cold hardy citrus species, and apricots by planting thousands of diverse seed. Both required letting mother nature select the survivors.

If you want to discuss a plant modified by the hand of man, look at maize. Original teosinte still exists in a few species and races. It is a multi-branching grass. Modern maize produces ears on “branches” and tassels on the main stem. Ears are enclosed by a whorl of husks. A hard shell surrounds each seed in teosinte yet was bred out of modern maize. Maize is incapable of living and reproducing in the wild. It relies on man to collect, distribute, and plant seed. Maize is a monster. It took @9000 years to do this with maize.

1 Like

The answer would be species dependent, and I don’t think anyone really knows the answer for the vast majority of perennial crops. Inbreeding depression would kick in long before you approach genetic uniformity. It took 9 generations of selfing to bring a grape up to 93% homozygosity (this was for genome sequencing before we learned to deal with heterozygous genomes). If we lowball the time to flowering to 3 years, that’s 27 years to get something that barely survives outside of a test tube.

Since you mentioned introgressing traits from wild relatives, you may find this paper interesting. Much of it went way over my head, but the concluding paragraph seems relevant:

On the one hand, evidence suggests that introgression has contributed to agronomic traits but introduced deleterious mutations. However, this combination likely contributes to a short-term gain of fitness via heterozygosity, adding to previous observations that clonal lineages may hide deleterious mutations in the heterozygous state (72). In this study, we have inferred this similar dynamic process, finding that clonal propagation can quickly select favorable and detrimental introgressed alleles in a much shorter time (< 100 generations) than for outcrossing population (>2,000 generations) Interestingly, most beneficial introgressed alleles are fixed in an outcrossing population eventually but not in clonal population. We conclude that introgression has contributed to the complement of the beneficial and deleterious variants in the heterozygous state within grapes.

My point was that in a clonal population, the diversity may be static but it’s persistent. Unless you keep every single individual in an OP population forever, you lose diversity over time.

A major advantage of grafting is that you can fit many times more diversity on a single tree by grafting on multiple cultivars than you will get in an entire orchard of seedlings.


That’s very considerate of you, no need to go to extra lengths. But thank you for your engagement on the topic.

I mean, yes, but what you need first is a source of that genetic diversity–diversity isn’t really the right word even, since it’s not a variety of genes you are looking for, it is a specific gene or set of genes–some parent that already has the trait you’re targeting.

Because, again, when you are breeding for something, you are reducing the genetic diversity. On purpose.

They’re trying both. Kumin, for example, is mostly working with poncirus, since he’s in a very cold area where even pure ichang papeda (the parent that yuzu gets its cold hardiness from) isn’t hardy. But more to the point, it does not matter where you get yuzu seeds from–all yuzu are clones. Yuzu is a cultivar, there’s only one, and it is among the citrus I mentioned earlier that are highly nucellar, so virtually all yuzu seeds will produce an exact clone of yuzu regardless of the pollen parent. As for being delicious, well, after some processing and added ingredients, sure, but yuzu straight off the tree is pretty close to unpalatable.

They have been crossing with the hardy disgusting ones (meaning poncirus cultivars and seedlings) However, (1) it is a hard cross to make that does not happen naturally (2) it’s been found, through lots of failures, that the resulting hybrids generally retain the terrible flavor and fruit quality of poncirus but lose the cold hardiness very quickly.

Which has implications for disorderly attempts at breeding edible cold hardy citrus. If you just breed a bunch of them and plant endless seedlings, all you’ll be doing is breeding terrible flavor into more and more generations, generations which very quickly approach the poor cold hardiness of the edible citrus parents. Only a careful, intentional, managed breeding program, where seedlings with off-traits are culled mercilessly, smart backcrosses are made, recessive and dominant traits mapped out, could hope to breed (some) cold hardiness into (most likely not very) palatable citrus. Random crosses have almost zero chance of ever working.

And I think you’ll find that for most fruits, if there is some combination of traits that seems like it would be good, but for some reason no such variety exists, that’s why. Breeding, even with very limited goals, is hard.

This is the point we keep trying to hammer home: once you lock in some traits in a landrace you have bred out the genetic variability or as you call it, adaptability. A landrace, by definition, is not adaptable. Think of short haired cats. When two short haired cats mate, their offspring are short-haired cats. No matter how many times you cross short-haired cats, you’ll never get a Maine Coon cat in the litter, even if the environment you were breeding them in was really cold. Short-haired cats had the long hair gene(s) bred out of them by definition. Sure, after thousands of litters, you might if you are incredibly lucky, get a chance mutation for a thicker coat, or longer hair, or more body fat, or bigger body size, something that improves their cold tolerance, but it’s very unlikely. As Fusion_power mentioned, it took 9000 years’ of mutations to develop corn, and even 200 years ago corn was much less productive.

Oh, Richard is nothing if not concise to a fault, but if you like, I can try to fill in the details on his behalf.

Basically, the issue is this–you’re using landrace to mean:

Ok, but in that case, we’ve pushed the definition of landrace so far it’s basically meaningless. As Richard pointed out, landrace has a specific, and pretty narrow, meaning. We can be a bit looser with the term if need be and, for example, drop the “historical” aspect (and there is credence enough to do so, since, for example, the Oxford definition is “a local cultivar or animal breed that has been improved by traditional agricultural methods” which kinda implies but does not require that the cultivar or breed have been developed a long time ago). But if we go so far as to say, “phenotypic variation doesn’t matter, it just needs to mostly taste good and usually grow easily” that’s going way outside the definition of landrace, or even cultivar or breed. Even at my most generous of moods, I couldn’t countenance that sloppy a use of the word, and those who prefer more precision in language would be straight up hostile.

At least when I have asked, aside from the first time, it has been rhetorical.

Uhh, no, that is the opposite. The more variability there is in the population, the longer it will take to winnow that out until you have a stable expression of traits–especially if you only use traditional breeding methods.

Open pollination, on average, reduces genetic variation within a population. When everything breeds with everything else, the population homogenizes. This is why speciation is so much more likely in mountainous areas. And this is why clonal propagation is such a powerful breeding tool. You can isolate unique genetics permanently, even for hundreds of years, and use them in thousands or tens of thousands of different crosses and backcrosses over centuries.

No, I’d benefit from if NC State University had started an extensive modern breeding program in roughly the 1920s to map out the diversity of Chickasaw plums, select the most promising parents, then crossed them with select varieties of domesticated plums, and used the progeny for a massive, multi-generational, professional breeding program over the course of decades. Why such a long program? Because I know it would need to be. Auburn University has had a similar hybridization program but for “just” a few decades, and even after numerous generations, and who knows how many seedlings and burning how much research money, the plums they ended up releasing were just ok.

Breeding is hard. Letting chaos do the hard work for you very rarely is the path to success.

I’ll be the first to admit that most if not all of us here criticizing your model actually share your goals. We want great tasting fruits of all kinds, that grow well and easily, and for there to be many different and interesting kinds. We’re all keenly interested in breeding, fascinated by it, committed to it in our own ways. We disagree, several of us quite strongly, that the method you propose would actually stand much chance of achieving those goals. Just “tasty and easy to grow” is a tall order in most places, a Sisyphean task in others. Adding requirements like open pollination or true-to-seed makes thing so, so much harder. And for what gain? You lose genetic diversity. You lose adaptability. You lose the chances of preserving unique but rare traits.


How would that work? Is it assumed that the genetics already exist in the seed to allow it to grow in the new environment or epigenetics change gene function, allowing to grow in the new environment?

1 Like

Danzeb, read about apricots. IIRC, only 4 trees out of several thousand were able to survive and thrive in the new colder climate. But those few were able to extend apricot’s range of cultivation signficantly.


Ha like this guy.

1 Like

@Justin one last thing to consider about OP is that it cannot readily be used if the selective pressure is too strong. A little bit of selective pressure and OP can, if slowly and haphazardly, shift a population towards expressing the trait being selected for, but too much pressure and OP is likely to fail.

This is because (1) recessive traits, (2) the difference between F1s and F2s. Unlike controlled breeding, where so long as you have the parents characterized correctly you can get recessive traits to express every time, OP is just a probabilities game as to if you’ll get expression–probabilities that can get pretty bad in higher ploidy plants. Similarly, while the F1 generation is often a kind of midpoint between parents, the F2 generation is highly variable and can have individual traits much closer to the original parents.

Both of these are critical if the selection pressure is high. I’ll use citrus as the example. For me in zone 8, the vast majority of citrus varieties will be killed by winter freezes. So any citrus I plant are going to face very strong selective pressure for cold hardiness.

The point about recessive should be easy enough. For example, there’s some evidence that deciduousness is a recessive trait in citrus. Obviously, being deciduous makes a huge difference in cold hardiness. If a large percent of your non-expressing carriers of deciduousness are going to get killed each winter, you’d really, really want to know which ones they are so you can protect them long enough to backcross and get your deciduous trait to actually express. You simply can’t do that with OP seedlings.

Now, for the business with F1s and F2s, consider this. Were I to just open pollinate cold hardy (CC) citrus and edible citrus (EC), I’d lose all the seedlings every year from ECxEC, naturally, but also ECxCC, since the midpoint in cold hardiness between CC and EC isn’t low enough for them to survive my winters, and only my CCxCC would survive. But CCxCC is inedible, so I’ve gained nothing, I’ve introduced no new genetics into my population of CC. The only thing I could do is put all the seedlings into a huge greenhouse until they are big enough to fruit, and then plant the second generation. But, that second generation of OP seedlings would be a huge mess of crosses, and the ones that would then survive the following winter would be dominated by (CCxCC) x (CCxCC), which, again, isn’t what I want. I don’t even really want much (CCxCC) x (CCxEC), for various reasons (I’ll miss any recessive traits in EC, and also it only takes a little CC ancestry to make fruit quality terrible, 3/4 ancestry is guaranteed to be inedible), I want (CCxEC) x (CCxEC), but very, very few of my seedlings will be that F2 generation, and I won’t even know which ones they are until that second generation itself matures and I can taste the fruit and make guesses as to which ones might be carrying some genes from EC. With controlled crosses, I’d know that every seedling was an CCxEC, I’d only need a greenhouse a quarter the size, and my F2 generation would be pure (CCxEC) x (CCxEC), which is the only F2 combination that has any hope of having the cold hardiness of CC and the fruit quality of EC. With OP, I have to protect four times as many seedlings, at least, and the second generation would have, at best, one sixteenth of it being the cross I actually want–and I’d have no idea which plants they are. After waiting for that second generation to mature, and then a few more years so the fruit quality actually presents (citrus fruit quality is much lower the first few years), I could then cull all the citrus that were not edible, and then wait another year for the next batch of OP seeds, since the previous year’s seeds would have been mostly pollinated by the far more numerous culls. What’s more, since I’ve spent so much time growing out seedlings that eventually got culled, or seedlings whose children I grew out to cull, I’ve wasted most of my population on plants that are dead ends. The number of actual crosses between plants with the right genetics quite small. All that work, with four times as many plants starting out, and an extra decade or so of time, and I ended up making only a handful of crosses of the right genes? That’s a recipe for failure, because I don’t need just one or two (CCxEC)x(CCxEC), I need many of them if I’m to hope to get that magical combination of the cold hardiness close to CC and the fruit quality close to EC. I need to maximize (CCxEC)x(CCxEC)s to even have a chance of success. In other words, I need to ditch OP, selectively cross CCxEC, protect those seedlings, and then make as many F2 crosses as I possibly can.

Breeding is hard, OP makes things way, way harder. And my example is just an idealized thought experiment. In the real world, even the controlled cross method is extremely unlikely to succeed. Cold hardiness isn’t just one trait, certainly not just one gene, and fruit quality is even more complex, so the actual number of seedlings and generations required could be staggering. Seedling vigor needs to be selected for as well, as does resistance to root rot (longer and wetter winters make root rot much more of an issue here than in Florida, even in well drained soil). The proportion of seedlings from OP plants that would just carry, not even express, traits for cold hardiness, seedling vigor, root rot resistance, and the numerous traits for fruit quality, without carrying traits for poncirus oils (foul tasting and acrid) or early bud bread (late frost danger), is vanishingly small. And those seedlings, being such a small part of the population, would then just cross with the wrong plants and lose the combination of genes we need, because, remember, we need the F2 generation for actually expressing all these traits. Not only do you need a unicorn to appear in a sea of horses, but you need at least two unicorns to not only appear, but somehow find each other and mate. Even with a population in the millions of seedlings, it’ll probably never happen.

And even that isn’t the whole story. Breeding is really hard. I’ve left out one crucial detail: nucellar seedlings. Remember, most citrus, including the majority of edible citrus, and an even higher proportion of cold hardy citrus, do not make crosses, even when pollinated. They produce seeds that are clones of the mother plant, with only a tiny percentage being actual children of the parents. An OP population of cold hardy citrus crossed with edible citrus, after two generations and decades of work, would just be a population of cold hardy citrus clones, with one or two actual seedlings mixed in, none of which would have any chance of expressing the right traits (if the vast majority of your F1 generation isn’t even F1, but clones of the mother, then you’ll never get an F2).

We make controlled crosses and clone out and graft the results because it is, by far, the most effective way of both preserving rare genetics and of attempting to create new and better plants. OP is a cudgel that will destroy most of your hard work. And breeding with the goal of establishing a landrace is literally breeding away your genetic material.


I like to breed plants so I found this thread interesting. Citrus sounds like a tough cookie to crack. My first thought is screw this let’s do some gene splicing.
With crispr technology it should be easy to do.
Be optimistic. I wanted to improve the flavor of primocane fruiting black raspberries and it took me one cross to do it.


Yeah, it’s not very friendly to breeders. But the payoff is definitely there–a good citrus fruit is really pretty exceptional. And at least citrus is culturally forgiving. So long as they have lots of heat, some good sun, dry feet, and you feed them amply, they do just fine. Up here, they are basically disease free, and will probably remain that way for a long time (most citrus disease are vectored by things that also don’t like the cold).

And citrus doesn’t have weird ploidy stuff like most other fruits do though.

A niffy trick with citrus is also that the fruit has so many uses, “good quality fruit” ends up being a pretty darn wide term. Extremely highly acidic and low sugar would be a bad fruit quality in a raspberry, but we absolutely love it in a lime. Yuzu is dry, bitter, sour, and has almost no flesh or sweetness, but the rind is super aromatic so it’s worth growing just for that.