I’ll take a chance to try to interpret what Pierre said. I think he is referring to things like insertion sequences and other DNA pieces that move round the genome in very sneaky ways. They are a sort of virus, but not necessarily easy to transmit with a simple vector like aphids.
What we do know for sure in the case of Arabidopsis, and for all dicots in general based on that, is that formation of flower parts is controlled by a few master genes. The number of cycles of making sepals, petals, stamens and pistils is determined by their timing of turning on and off selected activities. So doubling, converting stamens to petaloids, and reverting back to stem and leaf growth (phyllody) is determined by the duration of signals in the apical meristem.
I believe that even the number of leaves produced before a flower is initiated is genetically very precisely determined, but by a counting function. A Rainbow Knock Out has only 2-3 leaves per flower, but Peachy Creeper in my yard this year had 42, while New Dawn had 14. In maize too this is a definite genetic trait. Lots of plants behave the same.
Sometimes a change in one of the master genes, by some sort of insertion/deletion upstream of the place where RNA starts to be produced, (or even within the RNA sequence itself),leads to putting a new switch into the circuit, or taking one out. Maybe the controlling protein, or RNA is temperature sensitive, or a different regulator can bind better or worse at different temperatures. Temperature control is an ancient regulatory function, which is critical to seed germination in very obvious ways, in temperate climates, along with daylength and presence of light. It may control some other growth patterns in ways other than simple speeding up and slowing down total biomass production. But daylength effects are more dramatic and easier to study.
The kind of definite phyllody that Kim sees, sounds like the function that is supposed to form pistils is overtaken by the leaf forming function. This may be the analog of Lfy, the leafy trait, in Arabidopsis.
Elliot Meyerowitz, when at Cal Tech did a lot with the control of flower pattern formation. I think you can find lots about it on a wiki.
Queen of the Desert wears a really striking crown, or is it a turban like in a turban squash?