How to breed your own potatoes

Breeding new potato varieties is easy. You can hand-pollinate potato flowers in far less time than it'll take you to read this article, but I'm going to attempt a reasonably thorough explanation, so I hope you find it helpful.

Potato breeding is done through sexual reproduction, i.e. pollinating flowers to produce berries which contain true seeds (TPS).  Normally when you plant potatoes you propagate them from tubers, confusingly called seed potatoes but which are not actually seeds, but root cuttings. You can't cross tubers. They can only reproduce themselves as they are. Occasionally a plant may produce a spontaneous mutation but it doesn't happen often enough to be useful as a breeding method. Flowers are the way to go, because they give you the option to combine and reshuffle genes from the parent varieties of your choice.

There's a lot to be grateful for in the anatomy of a potato flower. Hand-pollinating them is very easy. The flowers are large and easy to work with, and the individual parts are easy to manipulate.

What's not so easy is making careful plans and predictions for what you might get out of it, and that's because potatoes are tetraploid. If you've no idea what I'm talking about then have a look at my previous post about TPS for a simplified explanation. To give a one-sentence summary: a tetraploid has double the amount of genetic material that a normal (diploid) organism has, which is a bit like inheriting traits from four parents rather than two. Tetraploids are a quirk of nature but in potatoes they are a very successful one, and the vast majority of cultivated potatoes in Europe and North America are tetraploid.

You may still come across the occasional diploid. Mayan Gold and its associated varieties are diploid, and those who are growing TPS from Tom Wagner may have a few diploid lines from him. Diploid potatoes can be recognised by a tendency to have smaller and less fleshy leaves, but the most distinctive feature is the berry. A diploid potato berry has a distinctively pointed end, kind of strawberry shaped, while tetraploid berries are more rounded and tomato-like. If you're feeling experimental you can try crossing a diploid with a tetraploid. At best you will only get a few viable seeds out of it, but it's a brilliant way of introducing new diversity into potatoes. At some point soon I will give it a whole article of its own, as it's too elaborate a subject to go into here.


Potato berries (these are tetraploid ones) in development.

However, assuming the potatoes you want to cross are tetraploid, since most of them are, it's very difficult to predict what the resulting offspring will be like because of the genetic variability involved. With tetraploids, the convenient order of the Mendelian ratio is thrown out the window and replaced by something more akin to a gene tombola. F1 hybrids are not uniform as they are in most other types of breeding project. If you grew a lot of offspring from your cross you'd find that many traits show continuous degrees of variation through the population, rather than segregating into Mendel's either/or groups … which happens because there are so many different ways the alleles can arrange themselves. To quote a research paper by Scotland's premier spudmeisters, Meyer et al (1998) "[Tetraploid] inheritance implies the random pairing of four homologous chromosomes at meiosis, and in a highly heterozygous outbreeding species results in a large number of possible allelic combinations at a single locus. In the most extreme case, eight different alleles could segregate independently in a population, resulting in 36 possible genotypic classes in the progeny." In other words, potatoes naturally have a mixed up genepool (from outbreeding) and when they pollinate and set seed those alleles can arrange themselves in any order - with each different combination having a unique effect on how that trait is expressed. And we're just talking about an individual locus here … the same is happening at every other locus throughout the whole genome. Yowza!

In short, tetraploids are complex and contain a lot of genetic material which can be immensely variable. Scientists doing genetic research on potatoes often choose to work with diploid lines instead, because tetraploids make such a muddle of their data it's hard to interpret anything.

So where does that leave you as a home gardener or small-scale farmer wanting to develop your own potato varieties? It leaves you in a position where you may as well have fun, experiment, use your imagination, be creative. As the results can't easily be predicted, you don't actually need to know anything about genetics. Think more along the lines of what you might get if you cross this colour with that colour, or this flavour with that shape - and then be prepared to be surprised!

One piece of misinformation I see spread all over the internet is a belief that you won't get anything worthwhile out of a home-made potato hybrid because producing just one good variety takes thousands of plants and many many years. This myth has arisen from under the slow-grinding wheels of the potato industry, which does work like that. Sure, if you want to breed a variety which will be listed in all the catalogues and sold in Tesco's and will make you rich from the royalties, your chances are very slim. The selection criteria for commercial potato varieties are immensely restrictive - and largely at odds with what most gardeners would want. Commercial breeders may well churn through (and reject) 200,000 seedlings to find one with commercial potential, then spend the next eight years doing field trials with it before it's ready for release. But don't let that put you off. You can breed a good variety within two years - easily. The majority of your home-made potatoes will be worthwhile, at least decent enough to eat and enjoy and feel proud of. A few will be exciting and wonderful. Even if you're only working with a very small patch of garden, you will almost certainly get some tubers that are worth saving and growing on next year.

The beautiful thing about potatoes is that it only takes two seasons to get a completely stable new variety. So it's actually quicker than most other vegetables. You make a pollination the first year and produce the F1 seeds, which are all unique individuals because of the genetic diversity. The second year you grow plants from those seeds and they make tubers. If you like the tubers, you simply propagate them by saving and replanting them. As the tubers are basically root cuttings of the parent plant (clones) they are genetically identical. There's no arsing about trying to make F2 and F3 hybrids (unless you want to) or years spent roguing out unwanted recessives. Once you've got something interesting, it's instantly a new variety.


Hybrid potato grown from TPS. This is one of Tom Wagner's hybrids, an F3 of Pirampo x Khuchi Akita. The parent varieties are Bolivian landraces, and are diploid.

If you've dipped a toe into plant breeding before, you'll know that plants tend to be either inbreeders or outbreeders - though that's more of a sliding scale than a polar absolute. Potatoes are really a bit of both. The natural status of potato is outbreeder. The various (diploid) landrace species from which cultivated potatoes are derived have a self-incompatibility mechanism which prevents them from pollinating themselves. The majority of diploid varieties are self-incompatible, although there are exceptions. This forces them to hybridise and mix their genes up in every generation, hence the wondrous diversity found among diploid landraces. However, when potatoes went tetraploid the compatibility barrier got screwed up somewhat. Many tetraploid potatoes have sterile pollen which can't fertilise anything at all, but others can fertilise themselves as well as each other. So they're designed to be outbreeders, but in practice a lot of flowers simply get knocked up by their own pollen.

Which gives you a choice: you can make hybrid seeds by crossing two different varieties, or you can make self-pollinated seeds which are the product of just that one variety.


Choosing parents: hybrid or OP?
Both are worth experimenting with, but for different reasons.

When a potato plant sets berries naturally without your intervention, it's most likely that it self-pollinated, but it may also have made hybrids with other potatoes flowering nearby, and you may have a mixture of selfed and hybrid seeds in the same berry. This is called open pollination (OP) … and the results are basically pot luck.

Making a deliberate hybrid is the most usual way to breed a new variety, as it introduces a lot more diversity. The basic method is to emasculate the flower to stop it from pollinating itself (which you don't even need to do if it's one of the many varieties with sterile pollen) and fertilise the female part of the flower with some pollen from a different variety. The offspring will be very varied, but that's the fun part and you should also get some hybrid vigour which makes for healthy and abundant plants. The only problem with this is that so many varieties of potato are poor berry setters, so not all varieties can be hybridised.

If you want to grow seeds from one specific variety it can be as simple as saving naturally pollinated berries from it, but if you want to be sure of getting self-pollinated seeds, it's easy enough to do (as long as it's one of the fertile varieties). Just dab a flower with pollen scraped from its own anthers, or other flowers on the same plant, or from other plants of the same variety. Bear in mind though that you will not get a true-breeding offspring of the parent variety by doing so. As potatoes are very heterozygous and have four lots of genetic material to throw around with cheerful abandon, even when they're self-pollinated they segregate into many different phenotypes. If you grow self-pollinated seed from Salad Blue, for example, you will not end up with a lot of spuds which look like Salad Blue. You will get varying shades of blue flesh, some much lighter than the original, some darker, and a few with pinky skin. If you grow selfed seeds from Congo, another blue variety, you may end up with a baffling range of purples, pinks and pure snowy whites, with considerable variation in tuber shape. What's happening is that all the genetic material which has been funnelled into the variety from various ancestors is segregating. Recessive traits emerge which weren't apparent in the variety you started with. If you grow enough self-pollinated offspring, you can start to build up a picture of the variety's pedigree, as many of the ancestral characteristics magically come back to life. So it can be a really fascinating thing to do.


A seedling grown from self-pollinated seed of Salad Blue. It's the only plant in the batch which has this striking black tinged foliage and black stems. I'm hoping it'll produce some dark tubers to go with it.

It's important to remember though that nature designed potatoes to be outbreeders, and if they self-pollinate they may show some degree of inbreeding depression. Only a bit though. As most spuds have such a rich and diverse genepool they can get away with a certain amount of inbreeding, but you may find self-pollinated seeds grow less vigorously than hybrids. That's not a problem and shouldn't put you off trying self-pollinated seed … but it's better to sow a few more than you need and then select the seedlings which show the most vigour and whoomph.


Variety differences
Differences in the fertility of individual varieties will most likely dictate what crosses you make, and how you do them. Over many years of being propagated by tubers, cultivated potatoes have moved away from the idea of flowering and producing seeds, and many of them can't be bothered to do it any more. Ironically for such a naturally variable and heterozygous plant, a historic lack of genetic diversity is thought to be the cause of the potato's fertility issues. The vast majority of modern cultivated potatoes are descended from one single Chilean spud, which had what is known as T-type cytoplasm, a genetic predisposition to making offspring with infertile pollen. Over the years many of these semi-infertile lines have been selected deliberately, as the male-sterility makes the process of hybridising them much easier. Consequently an awful lot of modern spuds have infertile pollen, and some are female-sterile too. Some can't be arsed to flower at all, and just dump their buds as soon as they appear. There are things you can do to force a reluctant variety to produce flowers, but it's a lot of hassle which I won't go into here, and from the point of view of future breeding work it makes more sense to choose varieties that at least show some willingness to come up with the goods.

I wish I could give a simple list of which variety does what, but I only know about the ones I've grown and observed myself, and it can vary from garden to garden anyway. Different countries have different varieties - import restrictions have affected exchange of material - so the ones I work with in the UK may not be available to people in the USA (just as most popular US varieties are strangers to me). So you will have to experiment with whatever you have available. As far as I can see, varieties fall roughly into four categories.

Some potatoes are very fertile and make excellent male or female parent varieties. Salad Blue is the Cassanova of the potato world - it only has to look at another potato and a berry starts swelling. You can usually tell a fertile variety because it naturally sets its own berries in profusion. Desirée is another very fertile one, and so is Mayan Gold, although the latter is a diploid so it needs to find the right kind of partner, or get lucky mating with a tetraploid.

Most cultivars fall into the male-sterile or almost-male-sterile category - these are the ones which flower happily enough but don't tend to set berries. Highland Burgundy Red is a good example of this, as is British Queen. It gamely produces a mass of dainty little flowers but in years of growing it I've never had a single berry. Give it a dab of pollen from a fertile variety though, and it sets berries very readily. So it makes an extremely good female parent. The advantage of male-sterile varieties is that you don't have to emasculate them, which makes it much quicker and easier to hand-pollinate them. Some varieties which appear to be male-sterile may actually be female-sterile. So it's worth trying the pollen on another variety to see if it will take. The disadvantage of using these partially sterile varieties is that it perpetuates the poor fertility of potatoes. If you want to do Solanum tuberosum a real favour in your breeding projects, select the progeny for good berry production. Because good berry production is what will keep its genetic heritage alive, as well as enabling some much needed new diversity to come in.

Then you have what you might call the awkward buggers category. These include Pink Fir Apple (syn. Rose Finn Apple) which not only has male sterility issues, it often can't be bothered to set a berry even when it's given fertile pollen. What usually happens is that the flower opens happily enough and you carefully pollinate it two or three times and on the third day the whole bloody thing drops off. Or worse, it starts to set a berry and then it falls off before it's mature. It pays to try again though, because there's a good chance that one of the pollinations will take eventually, when the plant is in the right mood and the planets are in the right alignment or there's an 'r' in the month. It's a pain in the backside to have to keep pollinating more flowers, but bearing in mind that each berry can easily produce 100 seeds or more, it only takes one successful pollination to give you loads of future breeding material - so it's worth persevering. Again, with a variety like this you don't need to waste time emasculating. I just go through the whole crop each day dabbing fertile pollen on every stigma I can find and saying "come on, set a bloody berry you sod!"

And finally you have the total refuseniks. There is a wonderful Victorian potato called Witch Hill which is reputed to be one of the best flavoured potatoes around - it is truly delicious. I would love to use it in breeding work. But every year the flower buds appear, and just as they're starting to look promising they drop off. All of them. Little dessicated posies cast to the ground. Now, unless it changes its mind, I cannot breed from it. If a variety won't flower, there is no breeding possibility, it's as simple as that. I could grow fields of the stuff and hope for a spontaneous somatic mutation, but that may never happen. Witch Hill is a genetic dead end. This is why breeders like Tom Wagner select breeding lines from varieties which are good berry setters. If a variety won't flower or won't set berries, it has no future.

The annoying thing is, Witch Hill did flower for me a couple of times when I first got it (it came to me as a laboratory-grown microplant) but I hadn't got into potato breeding at that time so I didn't think to make any crosses with it. D'oh!


Hybridising potatoes: the practical bit
Let's be grateful for small mercies: potato flowers are nice and simple and easy to work with. They are 'perfect' flowers which contain both male and female parts.


Anatomy of a potato flower. Each of the anthers is a double sac, both halves containing pollen. When the anthers mature they develop little holes in the ends (like a salt cellar) and the pollen falls out onto the stigma.

The Mayan Gold blossom shown above is fairly typical, but there are variety differences in the exact shape of the flower. Some have a long style where the stigma protrudes some way out of the flower (be aware that a sticky-outy stigma has more chance of being cross-pollinated by passing insects than one where the stigma is hidden away). Some produce a neat little fused anther cone, others produce a rather grotty collection of misshapen anthers which don't hold together properly. Some (like Pink Fir Apple) do weird things where anthers and petals morph into one another. None of this matters - the principle is the same. You'll get to know the individual character of the flowers in your own garden as you work with them.

In order to control what pollen fertilises the flower, you have to stop the flower from fertilising itself, so that means removing the male parts of the flower before they mature. As I explained above, with some varieties you don't need to do this - if the variety produces sterile pollen or none at all, you can save yourself the trouble. The instructions shown here are for if you have a fertile variety or want to be sure of getting hybrid rather than selfed seed.

Potato flowers are produced in cymes - bunches of flowers which open consecutively, 2 or 3 at a time. The flowers last two to four days but tend to close up in late afternoon. The anthers develop holes in their tips when they're ready to dehisce, though they're not very glamorous - in fact they look more like some insect has had a go at them. Potato pollen is white, powdery and very fine. The stigma is receptive for about 2 days and the period of pollen shedding also lasts about 2 days. Fortunately for the garden dabbler, the female part tends to become receptive just before the pollen starts to shed, so you have a window of opportunity to intervene.

The best time for hand-pollination is in the morning when pollen is most abundant, and when the temperature is fairly cool. But I wouldn't worry too much about this, it works at other times too.

If you're using a variety with infertile pollen, or you aren't bothered about the chance of a few self-pollinated seeds, you can skip steps 2 to 4.


Step 1: Having chosen the variety you want to use as the female parent, find a blossom at the right stage. Potato pollen can be shed quite early, before the flower opens, so emasculation has to be done while it's still at the bud stage. What you're looking for is a nearly-ready bud where the calyx (outer green bit) has started to open but the petals are still shut. This is a variety with a sticky-outy stigma, but with many varieties it will still be hidden inside the petals. Doesn't matter either way, although a sticky-outy like this inevitably carries a small risk of picking up stray pollen from elsewhere.

You may notice a strand of mauve wool poking out underneath. I tied this around the stem of the flower (or in this instance the whole cyme, as I'm going to hand-pollinate all of them) as a marker, so I can be sure I know which ones I've hand-pollinated. I use a different colour of wool to indicate different pollen fathers.



Step 2: Peel back the petals and you'll find the anthers inside. They are still immature at this stage - with no holes in the ends. If they do have holes and are shedding pollen, try a slightly younger bud instead!



Step 3: Using a blunt scalpel blade, tweezers or similar, pull/scrape the anthers off, being very careful not to damage the style - the central stalk with the stigma at the end.



Step 4: After removing all the anthers you're left with a denuded female part, ready to be pollinated with the pollen of your choice.



Step 5: Next find the flower you want to use as the male parent. Choose a blossom which is newly opened, as those are the ones most likely to have a good pollen stash (the ends of the anthers should be open at this stage). Pull off a single anther using tweezers/scalpel/fingers.



Step 6: When you turn the anther over you'll see it has a seam down the back, separating the two pollen sacs. Additionally, each individual sac has a little slit down its centre. Carefully slip the tip of a blunt scalpel blade through the slit and slide it along. Note that the slit should be open so you can insert the blade freely ... you want to avoid cutting into the anther if you can. 

If there is pollen inside, you will see it on the tip of the blade. It's a very fine white powder. If you don't see any white powder, try another anther from a different flower. You don't have to collect all the pollen at once ... just scrape out enough to dab on the female flower, and use the rest for more pollinations. (When the first sac is empty you can do the same with the other side. You can often pollinate ten or a dozen flowers from the pollen in a ripe anther like this.)



Step 7: Armed with your pollen-tipped scalpel, go back to the bud you just emasculated and dab the pollen powder onto the stigma - which is the knobbly-bobbly thing at the end. The stigma is mildly sticky when it's receptive, so you should find the pollen grains sticking to it quite readily. No need to make a song and dance with it - just a gentle dabbing so as not to risk damaging the stigma.



Step 8: The next day, go back to the same flower and pollinate it again with pollen from another fresh anther. The stigma remains receptive for around two days in total but you don't know exactly when that is, so for best results give it a pollen dab on three consecutive days. You'll notice that the petals have opened on this flower now, although it looks a bit weird as it has no anthers. Once the petals have closed and wilted a bit, you can assume it's no longer receptive.



Step 9: The berry starts to form. Yay!


From this point on, patience is the order of the day. Potato berries seem to mature painfully slowly. Try to resist the temptation to prod and poke them, you don't want them to fall off as they dangle clumsily on their alarmingly scrawny stalks. After about four weeks you need to watch for them dropping off naturally. Ideally, tie a little cloth or paper bag over them at this stage so that they are caught safely if they drop. Alternatively, be very vigilant, and ready to rummage about on the ground if you notice them suddenly go AWOL. Fortunately they don't taste nice enough for animals to be interested in them, at least not in the way of UK garden wildlife, so they're unlikely to be carried off, but you don't want to chance it.

One key factor about potato berries, significantly unlike tomatoes (to which they're closely related), is that the seeds carry on developing even after they've detached from the plant, and the berries will stay firm for months. This has several advantages. For one thing it takes the pressure off you to extract the seed from them ... you can leave them until you've got the time and inclination, even weeks or months down the line. Secondly it means that all is not lost if the berry is dropped too early or the plant dies prematurely. This is very significant in the light of the blight problems we are all besieged with. The plants can be struck down by blight, wither and rot - and the berries will still survive. Save the berries and let them mature, and as long as you clean them thoroughly they will yield perfectly healthy seed.

The topic of cleaning and processing TPS from the berries is to be the subject of a separate post. In the mean time, go out and make some berries!

With thanks to Tom Wagner and friends at the TaterMater forum for advice and suggestions (any errors are entirely my own responsibility).