carbon, even in your example, evolution works through the individual's fitness. So, the population would change but only because the adaptation was to the advantage of the individuals.
That said, sometimes the individual is hard to define. Take your brothers for example, being a diploid species, the brothers are related by a factor of 1/2 (for comparison, identical twins would be related by a factor of 1.0 and you and I are related by a factor that is indistinguishable from zero).
So, while the dead red cardinal no longer has any direct fitness by virtue of being dead, he does have a measure of indirect (or more properly called "inclusive" fitness) via his brothers. BUT, because he is related to his brothers by only 1/2, he needs at least two brother to be successful to make up for himself (this is a bit super simplified since no one is paying tuition). So, if the red bird dies, for his trait, BUT at least 2 brothers, or 4 nephews, etc etc live because of that trait, it can go on via natural selection.
I would agree with everything Notropis said except that species are artificial. They can be problematic to define and somewhat arbitrary, but then so can individuals and populations and damn near everything else. That said, they are not completely anthropogenic figments of imagination.
carbon, when you start imagining lots of different scenarios it all gets pretty interesting how special conditions can make really weird things happen. Natural selection as defined by the Minnow (Notropis) points 1-4, can actually lead to species extinction in some cases. Some of which are not so crazy weird (and some that are). Evolution is a cool thing. How it goes down and what can happen is really interesting to puzzle out.
