You seem to be raising a couple of puzzles here. One is how it has come about that human beings are sometimes motivated to help strangers, given that we might have expected evolution to produce beings concerned to promote the survival only of themselves and their kin. One immediate answer to this question is that, despite what some fanatical sociobiologists say, you can't explain everything about human behaviour as it is now merely by reference to our evolutionary history -- even if you allow in cultural as well as biological evolution. Evolution made it possible for impartial benevolence to develop, but it didn't necessitate it. Its emergence -- where it has emerged -- is to be understood primarily in terms of history rather than biology. The second puzzle is why we *should* help others even when doing so doesn't benefit us. Many people have believed egoism -- the view that the only reason we have to do anything is grounded in the promotion of our own well-being. Philosophy has so far failed to...

It might be helpful to follow a strand of British empiricism and to think about 'morality' as a social phenomenon, involving various 'sanctions' such as blame, guilt, shame, and so on. (So in that respect it is rather like law, though the sanctions there are somewhat different.) Your worry is that some moral principle you accept -- that it's wrong to cause serious suffering merely for fun, say -- has emerged only because of the evolutionary advantage conferred on groups which accept something like that principle. So it seems quite contingent which principles we come to believe -- as you imply, in different circumstances we might accept different principles. But, to pick up Alex's point, we have the capacity to stand back from our 'morality' and assess whether we have independent reason to accept its principles. In which case, if you believe there is a reason not to cause suffering for fun, you may think that this justifies the moral principle which forbids it (as it would also justify a law forbidding it).