N = 6

Oh the magic of statistics. X and Y now both significant. Glory, fame, publication, etc etc.

(side note for stats junkies: although I am cranky about having needed to get this last N, I did take the opportunity to learn a bit more about the paired t-test vs the Wilcoxon paired test, and when to use each of them. Turns out any time it would be appropriate to use the paired t-test, you can also use the Wilcoxon, since it's ok to use nonparametric tests (Wilcoxon) even on normally distributed data. I knew this general statistical principle already, but still, the voice of my stats teacher rings in my ear: "You can't just do your stats a couple of different ways and pick the one that's most significant!!!" Which is just what I did. Screw the paired t-test, it's still yapping about p = 0.06.

Of course, every biologist I know does stats this way, because statisticians aren't biologists so what do they know? Also there's no way to find out, with an n of 6, if your data are normally distributed or not, so it doesn't seem all that illegitimate to try the parametric and non-parametric tests, unless you have really good reason to believe that your data shouldn't be examined with parametric stats.

But maybe there are biologists out there who actually follow the protocol of evaluating their experiments beforehand to decide on the N they'll need for statistical success, and sticking to it? I sure don't know any of them, but if you do that, I'd be interested to hear about it.)