FWIW, HalfManHalfWit seems to have an excellent grasp of evolution and needs no lessons in it from anyone. Maybe we can page Darwin’s Finch to come in, agree, and settle it?
Anyway, just one thought regarding why there’s less cost (than one might think) to suicide attacks: it’s possible that multiple stings from a single bee are much less effective at deterring vertebrates than a single sting that injects all the bee’s venom at once. So if bees are going to inject all their venom in the first sting, there’s no benefit during the attack to having the bee survive the stinging.
(Now, suicide bees still have a cost after the attack, as the hive has fewer workers; I’m just pointing out that the cost during the attack is probably minimal).
I’m afraid it’s quite possible. In mammals, that’s true (at least for every species that we know of), but in other groups sex determination is not always genetic. For instance, the sex of crocodiles is determined by the temperature of the eggs during incubation. In some fish species, individuals change sex during their lives, as they grow larger they take on a different sex.
And, to directly address the issue raised, in (at least some species of) aphids, females parthenogentically reproduce more females throughout the summer, but in the fall the last generation develops as both male and female (which then fly around, mate, and lay eggs that hatch females in the spring).
This is of course not true at all in the case of social animals. Humans past the age of reproduction in traditional societies can have a big effect on the survival of their grandchilden and other descendents by providing care and imparting long-term knowledge. As has been pointed out, the hive is essentially a superorganism, and maintaining an adequate number of infertile workers for foraging and defense is vital to the ability of the queen and drones to produce fertile offspring. So the death of too many workers is something that can be selected against, by impairing the ability of the fertile members of the colony to reproduce.
Also, it’s worth pointing out that behavior isn’t “strictly random” in the case of non-social animals. Behavior which enhances survival until reproduction is accomplished will probably continue afterward. The behavior after reproduction might not be selected for, but it sure was selected for before reproduction, and since there’s no particular pressure to alter the behavior, it might well continue.