I looked in Knuth before asking and didn't see it.

But that is just the algorithm I am worried will distort the distribution at the end! For instance, if M is one, the probability of the first element being set is 1/N, but the probability of the last element being set is ...

(N-1)!/N!     Oh. (Whacks self on head.)

Knuth 3.4.2 Random Sampling and Shuffling, Algorithm S
Thanks, Abigail