At first, I amused myself of thinking about it from a Lamarckian viewpoint, then realized that it'd make sense since those groups who are less likely to die from appendicitis before birthing age are more likely to pass the trait on. Although we've had appendectomies for a while now, we (as a race) have been snuffing it from appendicitis a whole lot longer. So there, Lamarck. Suck it.
Now, this is the tricky bit.
If that's true, and if we presume it'd also apply to personalities and behavioral traits (which has been often cited as happening -- particular personality/behavior traits existing for an evolutionary reason), and if we presume that the penchant for homosexuality is a trait one is born with... (drumroll) ...why wouldn't it be tapering off over the multi-centuries?
If you'll pardon the hyperbolic joke, if you're passing a penchant for homosexuality on to your kids, you're not doing this whole "gay" thing very well.
I know I've read some essays about it, but I can't recall what the general theory is for why something that all but eliminates the odds of procreation would continue on as a trait.