I see no reason that the size of the seed should matter in the use of colchicine for introducing somaclonal variation via unreduced gametes. Remember that meiosis occurs in the developing inflorescence, and not in the seeds themselves, in which fertilization has already occurred. (This doesn't seem to stop folks from soaking their seeds in colchicine, however, and I'm not real sure how that works. Perhaps just for creating polyploidy.)
But before you spend a lot of time and effort in trying to improve your Begonias, it's worth noting that commercial varieties already have much variation in ploidy, and reports vary from 7 to 56 chromosomes (Marasek-Ciolakowska et al, EUPHYTICA, October 2009). Furthermore, some studies in begonia suggest that this may not be a very useful route for variety improvement. Check out the 1995 article "Measurement of somaclonal variation in begonia" by Bouman et al in ISHS Acta Horticulturae 420:98.
On the other hand, maybe you're not planning to work on begonia, or maybe you've got nothing else to do with your time. In any event, the size of the reproductive organs shouldn't matter.
Linear Mode
