It doesn't matter what the supervisor's Ph.D. was in, originally: fields are protean and careers take all sorts of strange paths. What really matters is whether the supervisor's knowledge is sufficient to supervise the research that the student will be conducting. That depends on the path of their career, the precise focus of the student's research, etc.
My advisor, for example, had a Ph.D. in mathematics (since computer science didn't really exist as a separate field yet), but has worked in artificial intelligence, electrical engineering, physics, VLSI, biological modeling, and astronomy, and would be quite appropriate to supervise students in projects in any of those fields, if their focus and his background had sufficient overlap. Likewise, my background in artificial intelligence was no barrier to supervising a biology Ph.D. (which I have done), given the path of my own research over time.