Feathers on the other hand existed for the purpose of insulation so the more nested their structure the more they are suited for this purpose since they can better store the warm film of air (I think VadeRetro provided some examples of flightless dinosaurs with primitive feathers). Only later they also proved to be suitable for flight.
Also, a mutation doesn't have to be beneficial to spread in a population, a neutral mutation can do that as well. It may not be as likely to spread as a beneficial one but the chances for this are not zero. Especially in small populations this can happen quite often (-> founder effect, genetic drift).
Not correct. Neutral mutations cannot spread and likely will be lost within a few generations due to the laws of genetics. An allele with no survival benefit will just reproduce at the same rate as the population so it will never spread throughout a whole population. It will likely be lost due to the laws of chance. Since only one individual would originally have that mutation and at each reproduction there is only a 50% chance of the mutation to be reproduced with 2 chances of it being passed on in a stable sized population, it would eventually be lost due to the laws of chance (eventually you would get 3 procreations in a row that did not pass on the mutation).
Small populations also have a very big problem for evolution. Yes, they make mutations much easier to spread and that is the problem. Bad mutations are much more frequent by several orders of magnitude than good mutations as even evolutionists will agree. Therefore the tightly knit small population will be long destroyed or accumulate so many bad mutations that it will become unviable.