(This was first posted in sci.stat.consult, but I realize now that
the present group is more appropriate. Sorry for the cross-posting.)
Does anybody know of a treatment (preferably a text) of occupancy
numbers for the case in which they appear in an arbitrary order? (I think
that is equivalent to the urns being indistinguishable as well.) Feller
gives a few examples but no general expressions.
I have come across references to the book by Johnson & Kotz: "Urn
Models and Their Application" (Wiley, 1977). It is now out of print but I
could try to get an interlibrary loan if someone thought it would be
useful.
--
Norman B. Grover
Jerusalem, Israel