I get the underlying theory just fine. It's a neat fictional example, but (and I'm not familiar with the underlying fiction) it would probably be extremely fitness-enhancing. A male light would probably be incredibly high status and have little difficulty producing offspring. If it were purely genetically determined, it seems like it'd be pretty hard to sustain - no one would want it for their own children. If it were recessive, it might work out better, but there still seems a substantial problem of free-loading.

Thus, this evolutionary explanation for homosexuality partly undermines itself: it's genetic, but it's not quite genetic and there's other stuff going on that determines whether or not it gets activated. So it's either genetics (actively-selected) + environmental factors or genetics (random noise) + environmental factors. That's not a very clear case for kin selection, to say the least.

My claim isn't that it couldn't possibly be related to kin selection. It's that, like many ev-psych claims, the evidence for "or something else is going on" is far too strong to make a definitive claim, particularly because the outcome is the exact opposite of what you'd expect from simpler evolutionary theory. Otherwise, you risk combining two theories in a way that can explain far too many outcomes. Any individual who fails to reproduce can divert resources to his siblings. You could just as easily say many negative traits that don't show up with absolute consistency are also advantageous. This seems like a stretch.

In retrospect I will admit that this example detracted from my overall point and was poorly chosen.