This is a great discussion! I am so happy that so many GC members actually give a hoot about the science behind breeding and glider care. So often I interact with people who would just rather not bother learning the scientific ins-and-outs of their field (and I teach pre-med students - yikes!). This just adds more evidence in support of my hypothesis that glider owners are awesome! <img src="/ubbthreads/images/graemlins/thumb.gif" alt="" />
Some comments: Cheers to all who recognize that you need a large sample size to make any sorts of statements about statistical frequency - that is completely correct. Punnett squares predict probabilities, and usually pretty accurately *in the big picture*. The only way we could get a big picture view of the genetics of glider color variation would be to look at all the crosses made by many
breeders over several years. For every cross, it would be ideal to know the parentage and grandparentage. Since a majority of
breeders keep track of their crosses, it would probably not be too hard to gather these data (but... interpreting them is another story).
There could be other factors affecting our interpretation of the outcomes, too. Technically every time a pair is mated, the cross will have the various probabilities for each outcome. But there are many reasons (in addition to co-dominance) that this might not be the reflected in the percentages of each variation (plus, co-dominance can be corrected for in a Punnett square): first, when you do a Punnett square for one trait, you are assuming that the genes in the square are the only ones contributing to the outcome, and this is not always the case. For example, I think eight genes work in concert to produce human eye color. Each of those genes may have varying levels of dominace realtive to the others. Second, other genes that do not directly code for a trait can have a muting effect on the genes that do. Third, even if the crosses are being made in the correct proportions, the offspring resulting from some gene combos might not be coming to term. This is because some unrelated genes may be linked as "packages" that are inherited as a block, and because some alleles drastically reduce survivability in the wild. The body can recognize pairings that would have serious problems and can naturally abort offspring very early on in pregnancy. For example, if a neat color variation was linked to a fatal mutation, the body would likely destroy the fetus before the joey even moved to the pouch. A fourth possibilitiy could be that the dominant gene is very rare in the population, so it skews our interpretation of what is dominant. An example of this is polydactyly (=having more than five fingers or toes per hand or foot) in humans. The gene for polydactyly is dominant, but present in a very very low percentage of people, perhaps because it affected survivability as humans evolved. Finally, some mutations seem to be produced randomly fairly often because some genes just have a problem replicating themselves perfectly. So random origination of a mutation may amplify its presence in a population. I'm sure there are other possibilities, too.
Also: As someone who will get her M.S. in Zoology in May, I can assure you that to be a geneticist you need a LOT of math and statistics. You do math all the time. If her professor was a trained geneticist, he probably has a much better grasp on math than all of us (myself included), assuming none of us is a professional mathematician. <img src="/ubbthreads/images/graemlins/grin.gif" alt="" />