I'm going to go out on a limb here and say that in general, I'd assume that having the limitations in the
join on clause would be better than all conditions in the
where clause. My understanding is that something like your first query will join the tables and then from those joins use the where clause. Whereas the second query will be more selective in the join and will search through less records.
Hard to come up with an elegant way to try to explain that...sorry. Of course, that is just my assumption.
http://dev.mysql.com/doc/refman/4.1/en/query-speed.html goes through the optimization processes.
Of course, it's all going to depend on the data in your tables, table size, indices, locking, etc. The best way is to run both and compare!