Who first proved that, over ZF, the statement
(1) The reals are well-orderable
is strictly stronger than the statement
(2) Every real-indexed family of nonempty sets of reals has a choice function?
The separation is of course implied by the consistency of large cardinals - (2)+$\neg$(1) is a consequence of AD$_\mathbb{R}$ - but a large-cardinal-free proof seems nontrivial, and I'm curious who first came up with such a thing. (To answer the implicit question, I've thought about it a little and can't come up with the proof myself, although I'm probably just being slow - I remember being told that the separation is provable with no consistency assumptions, and I think I recall being shown a fairly straightforward argument for it.)
Well, I googled the second statement, and the first article to come up was the following:
Which you can also find here. You're looking for Theorem 4 (which is in the third section).