Great question. It seems to be all over the board and depends on the rose and other conditions. Some triploids tend to produce more functional 1x, 2x, or 3x gametes than others. The aneuploid ones (incomplete sets) tend to abort, but not always. In addition, pollen competition and how thick one places the pollen on the stigmas can play a role. More pollen applied seems to result in the 2x gametes outcompeting the 1x.
Theoretically a triploid should produce 50% 1x and 50% 2x gametes, so a triploid x triploid cross theoretically should have 25% diploid, 50% triploid and 25% tetraploid offspring. Also a cross of a tetraploid x triploid should theoretically have 50% triploid and 50% tetraploid offspring. In David’s paper on pollen diameter and guard cell length, the overall number of tetraploid and triploid offspring from a number of tetraploid x triploid crosses was close to 50%-50%. But some of the crosses had more triploid offspring and some of them had more tetraploid offspring. Perhaps meaning that for some triploids the 2x pollen out performed the 1x and in other triploids the 1x pollen performed better.
I’ve used a triploid as both the pollen parent and as the seed parent but I’ve only checked the ploidy of one of the offspring from those crosses. It was a tetraploid x triploid cross and the offspring had huge pollen grains, the size one would expect from an octoploid. Probably because the chromosomes weren’t pairing up properly and thus weren’t reducing in numbers when they formed the pollen grains. So sometimes you don’t get what you’re expecting.
Can anyone speculate what one might wind up with (percentages) when crossing a diploid with a triploid? I’m thinking of ‘Softee’ specifically as the triploid.
There is much variation between different triploids. Wulff (1959) discussed a few cases.
A lot depends on the types of chromosomes that are present. If all three sets come from a single species, or from two closely allied species, the chromosomes are likely to get tangled up in trivalents, along with pairs and the occasional unpaired chromosome. However, if two sets came from the same diploid species, while the third set came from a distantly allied relative, you might get a bit more fertility.
For example, Rosa macrophylla var. Korolkowii is an autotetraploid. If you cross it with the diploid form of the species, the resulting triploid would present a tangle of chromosomes at meiosis.
On the other hand, cross Korolkowii with a Tea or China. This time the two sets of macrophylla chromosomes pair up, leaving the Tea/China chromosomes pretty much to themselves. There’s no guarantee that this triploid will be highly fertile, but at least there shouldn’t be so many tangles of three chromosomes.
Finally, cross ‘Basye’s Amphidiploid’ with Rosa xanthina. In this case we cannot expect normal pairing of any of the three sets of chromosomes. It might give the occasional viable pollen grain or ovum, but I won’t be holding my breath.