In my experience there are a greater proportion of good undergraduate mathematicians that are left handed than one would assume if lefthandness were independent of mathematical competence (to a certain level). A discrepancy that seems to roughly accord, on a very small sample, to those who are better than average musically, and sportingly. Yet I also would contend that there are no more musically adept mathematicians than one would expect in any sample of a well educated population, and a damn sight fewer sportingly talented.
I have also noticed that, given the previous paragraph, there are fewer lefthanded lecturers than one might suppose, but then the sample becomes even less reliable than before.