FOM: replacing the separation scheme

[Colin Mclarty]

	I have to apologize for a mistake in my post earlier today
where I said I was giving a replacement axiom scheme for toposes.
In fact, I gave a separation axiom scheme, and the "generalized
elements" should be "global elements".

	Actually my book is the only publication I know of to
give a replacement axiom scheme. The main point is still that it
does indeed look much like the one for set theory. For any 
relation R stated in the (first order, external) language of the 
topos: 
	For any object A, if for each element x of A there is a 
	unique set B(x) such that x is R-related to B(x), then
	there is a function g:T-->A such that the fiber of g
	over any element x is the set B(x).

Colin