Many things are fine in normal doses, but in higher than normal doses they can cause problems. Was thinking maybe then that the testosterone present for a male child causes a different reaction in the female body than a female child.
While it's true that excess amounts of any substance can cause problems, we're talking about the immune system in particular here. The woman's immune system has encountered the testosterone molecule every day of her life and even for some time while she was in the womb. So the molecule itself would be ignored by the immune system. The only way it could possibly cause a problem would be if there was a mutation in the child's DNA for producing testosterone, which caused it to be slightly different and marked as foreign. But that doesn't make sense to me because then you'd expect the rates to increase not based on numbers of male children but based on all children (since if the woman had a daughter, that child could get the mutation) or based on an assumption that male testosterone is different from female which doesn't seem likely to me.
So it would still be related directly to male children, and not female children?
Correct. They theorize that something is attached to the Y-chromosome that allows the mother's immune system to recognize it as a foreign substance.
I read a book called Born Cannibal that talked about this. And no it wasn't using "born cannibal" in a purely metaphorical sense. It pointed out cases where twins literally tried to eat each other in the womb. If I find the book, I'll post the relevant information.
If you do find it, I welcome the information but at the moment I remain entirely skeptical, because the idea given that a twin literally attempted to eat the other twin tends to go against the idea that a baby can't eat solid foods even when it is at full-term. I've heard it reported as "one twin ate the other" in cases where what really happened is they were competing for the mother's nourishment and one died.