The disappointments and joys of breeding tetraploid Siberians

Late August and early September is harvest season for Siberian seed pods. After a couple of weeks to dry out, it’s then time to open up all those promising capsules and find out what’s inside. I just finished doing this for the 2011 crop and came away rather disappointed, particularly with the yield of tetraploid seeds.

Breeding tetraploid Siberians can be a test of resolve for several reasons. First, with a few exceptions, they are reluctant to set pods – just less fertile than diploids which generally set seed with abandon. Whereas maybe 70% of diploid crosses end up producing seeds, probably only 20% of tetraploid crosses do so – and that in a good year. Not only that, the tetraploid pods typically contain fewer seeds. Finally, it’s not unusual for these seeds to germinate more slowly and less completely than diploids. So the overall result is much effort for few seedlings. Maybe this relative infertility explains why tetraploid Siberians are not found in the wild – not only would their production be a rare accidental genetic event, but they are less likely to produce seed to compete with diploids in the next generation. For every diploid cross that produces a hundred seedlings you might get only ten from a tetraploid cross. You might think this relative infertility would improve with time as the more fertile tetraploids produce later generations that breed more freely. If so, I’m not sure I have seen it yet.

There are a few tetraploids that are quite fertile: 'Moonlight Fair' and 'Simon Says' show up repeatedly in successful crosses along with a handful of so-far unnamed tet seedlings. Most are more reluctant, but with persistence may yield some seeds, and a final frustrating few just refuse to ever breed. Perhaps the most poignant result is to produce a remarkable new variety and find it is resolutely sterile. Two of the loveliest seedlings we ever produced are 97B1B25 and 97A2B52. Dozens of attempts to get them to be parents have failed utterly and neither were introduced for this (and other) reasons. Because of this differential fertility, after a while, the fertile ones begin to dominate the seedling gene pool in the garden and there is a certain sameness about the new seedlings, so it takes a continual infusion of new genes to maintain the variety that every hybridizer prizes. The main frustration though is that often you can visualize a good cross that should produce new and wonderful things, but you can never get there. Many lines of development are either closed off or peter out prematurely as fertility is lost.

I mentioned 20% success “in a good year”. What does that mean? Well, the chances of getting seed to set is clearly temperature-dependent. In my experience there is no point in making tetraploid crosses when the daytime temperatures rise into the higher 80s, as they often do. Diploids too are less likely to set seed then, but not to the extent of tetraploids. Even in a cooler year like this year, many pods develop but contain few, if any, seeds – I call these false pods. Strangely, this year virtually every cross made in the later part of this season, whether diploid or tetraploid, gave mainly these false pods. I have no idea why, but that’s the reason I’m not entirely happy. As if that were not enough, there is yet another challenge with breeding tetraploids – it is harder to get recessive traits to reveal themselves. This arise from the fact that tetraploids have four sets of chromosomes not just two, and therefore have four copies of each gene. For a recessive trait to show up in the progeny, all four genes for that trait have to be in the recessive form, not just the two as with diploids. In the end it is roughly four times as difficult to bring out any hidden recessive trait (red or white/yellow colors for example). So, you would need many more seedlings to get the same expression of recessives as with diploids, and for the reasons above, you actually get many fewer in most cases. This again makes certain breeding strategies involving the subsequent recovery of recessive traits in later generations much more difficult.


So, there’s a quite obvious question, if breeding tetraploids is so much more difficult and frustrating than diploids, why bother? Partly because of the challenge and sheer stubbornness, but more significantly because when things do go right you can get some wonderful flowers. By no means everyone prefers tetraploids and they certainly haven’t displaced diploids from the Society for Siberian Irises' Favorite Fifteen list, but they are appreciated by many (seven of the top fifteen vote-getters are tetraploids in the most recent 2011 Siberian popularity poll).

Large flowers, strong ruffling, vibrant colors, and even new patterns show up, generally with excellent vigor. Here for examples are four recent tetraploid seedlings still under number showing further development of the plicata pattern that first appeared with 'Emily Anne,' and a couple of others in which the ruffling and form are superb. This is what makes it all worthwhile.