I agree with JD that one should never raise misconceptions (I prefer
to call them "preconceptions"--less pejorative and besides, most of
them have been developed over the years by their own observations,
irrespective how how flawed those observations may be) in class. If
you bring up a common preconception then it is more likely to imprint
on other students and reinforce in those already suffering from it.
As John says, its best to teach the correct conceptions and ignore
the incorrect ones--until they are raised by a student. Then deal
with them as quickly as possible, preferably by showing the students
how the conception is incorrect and why, perhaps with a clearly
designed demo that can be done in the spur of the moment. Major
potential pitfalls probably need to be addressed, but carefully, as
it is more likely that students will remember the pitfall than the
correct idea (in analogy with the recent finding that talking about
"myths" about any topic--see the Wash. Posts, weekly Saturday column
"5 Myths about . . ."--tends more to reinforce the myths than debunk
them).
More important is how to deal with your misconceptions when they
become apparent to the students. I always found it best to
immediately admit it and quickly correct it, apologize for the error
and move on.