Woohoo! My potatoes germinated

Potato seedball harvested last July. I'm now in the process of growing its seeds.

It may not seem like much to get excited about, but it is for me. It's the first time I've tried growing potatoes from seed. That's right, not the usual tubers left sprouting in eggboxes on windowsills (though I have plenty of those too) but real, true seeds.

Just to confuse things, the term 'seed potatoes' or 'potato seed' usually refers to tubers which are supplied for planting. But they're not seeds really, they're vegetative cuttings. Potatoes do produce real actual seeds though, and to differentiate them from seed tubers they are called 'true potato seed', or TPS.

If you've been looking at the blog for any length of time you may remember the small green 'apples' I collected from my potato plants last year. Not all potatoes produce them, because many these days don't have fertile pollen in their flowers and so have lost the ability to set top fruit (or seedballs, as they're properly called). But every now and then they appear.

They look very like miniature green tomatoes, for the very good reason that potatoes and tomatoes are closely related (hence their susceptibility to blight). When you open up a potato fruit it has juicy white flesh full of small tomato-like seeds. I saved some last year, not knowing whether they would be viable or not, and they have taken a couple of weeks to germinate ... but there they are, little pale green noses poking up out of the compost just like tomato seedlings.

You may be wondering why I'm bothering to grow potatoes from seed when it's easier to grow them from tubers. In fact, it will take two seasons to get a decent harvest from them this way. But the appeal (for me) is that each seed is unique, and effectively a brand new variety. Some will be worth keeping and some won't, but they'll all be unique.

When I do my pea breeding projects I know which genes control the traits I want, and I can predict to some extent what I'm going to get if I cross Plant X with Plant Y, even down to the likely ratio of particular characteristics in each batch of offspring. But that's because peas are efficient inbreeders, genetically stable and easy to control. Potatoes are the opposite. They are a gene lottery. I don't even need to do any crosses ... I can take self-pollinated seeds from one plant and the resulting offspring will all be different from the parent plant and from each other.

The reasons for this are a bit complicated, and if you're not interested in plant genetics you may want to hum a pleasant tune to yourself while you read the next bit. It's basically down to potatoes having twice as much genetic material as normal.

Here goes then. Genes are carried on chromosomes, which are chunks of DNA. In most plants chromosomes are arranged in matched pairs, which means there are two copies of each gene. Each gene pair may consist of two dominants, two recessives, or one of each, and these combinations determine which traits are expressed in the plant. The technical name for this gene arrangement is diploid. When a diploid plant is fertilised and produces seed, the chromosome pairs are torn apart down the middle and only one half is passed on by each parent. The seed (and the plant that grows from it) inherits one set of chromosomes from the ovule of the flower which produces it and one from the pollen which fertilised it. The two sets match up into their correct pairs, so the seed ends up with a new diploid (paired) arrangement made up of one set of chromosomes from each parent.

Potatoes are not diploid though, they're tetraploid. Instead of having two sets of chromosomes they have four. So when they reproduce by setting fruit, each seed inherits two complete sets of genes from each parent instead of one. Every chromosome carries many thousands of genes, and as both parents have four sets of chromosomes each, and the chromosomes break up into parts which randomly recombine, it becomes quite a major reshuffle ... more or less any combination of genes could end up being passed on to the offspring (or not). Which makes for a right mixed bag.

And suddenly the system of one gene being dominant over its neighbour becomes a lot more complicated. Instead of a simple 'either-or' situation you get four copies of each gene vying for dominance. You may get one dominant and three recessives, two of each, four recessives, four dominants, etc. Recessive traits which have been hidden for generations suddenly emerge. The likelihood of a seed-raised potato having the same genetic makeup as its parent(s) is very small indeed.

One way of looking at it is to imagine a diploid plant is like a one-armed-bandit fruit machine with only two wheels spinning inside it. The number of combinations possible with two wheels is relatively small. Increase the number of wheels to four though, and suddenly you have a massive increase in possible combinations and much less chance of predicting what you might get. That's basically what happens with a tetraploid.


Even if my potato seeds come from a flower which pollinated itself, the number of possible different combinations of genes should ensure that every plant I grow is a new variety. Some will be crap and may not even survive. Most will be good, but probably not very interesting. But with any luck a few will be excitingly different. At the end of the first season I will harvest the tubers from each plant and select the ones I like best. They will only be mini-tubers, not full sized potatoes. To evaluate them properly I will have to save them over next winter and replant them in 2008, when they should reward me with a decent crop. And once I've done that, I can select any I want to keep long-term, and that's how I get my new variety.

OK ... so how the hell do you maintain a new variety, when there's so much unavoidable variability every time you reproduce it? Well actually that bit is easy. Once you've grown a potato variety you like from seed, you just save and replant its tubers. The great random genetic lucky dip is only an issue when plants reproduce sexually by producing seed. Because tubers reproduce by vegetative propagation they are actually root stem cuttings (clones) from the parent plant. So each new plant grown from a tuber is genetically identical to the plant which produced it, and remains so in successive generations. This is true of all commercial potatoes too of course. If I plant a tuber of King Edward, for example, I'm essentially growing a cutting from the original King Edward plant which flourished in a Northumbrian field back in 1902.

So that's the genetic basis for wanting to grow potatoes from seed. But there are other benefits. Tubers are genetically identical to the parent plant but they also very efficiently store and pass on all the viruses and nematodes the parent plant may have had. These can build up over successive generations and become a serious problem, which is why most countries have laws requiring commercially available seed tubers to be tested and certified disease free (usually involving hefty blasts of chemicals). However, potatoes raised from seed are generally virus-free anyway, because there are very very few seed-borne potato diseases. And growing from seed represents a good opportunity to select new varieties with better disease resistance, specially adapted for your own garden, because most modern commercial varieties are pretty rubbish in that department.

If you're interested in developing your own potatoes I highly recommend a fantastic eBook, Amateur Potato Breeder's Manual by Raoul A. Robinson, which the author has generously made available as a free download. It's angled towards breeding for sustainable disease resistance, but the information is so thorough and clearly explained it's a gift to anyone who wants to grow spuds from seed. He even explains how to graft a potato scion onto a tomato rootstock so that it has to put all its energy into forming flowers and seedballs instead of tubers. Now that sounds like fun!