Ok, bob b, here's the beef. Semantic arguments are done. You want your proof.
The Luria-Delbruck fluctuation experiment done in 1943 still illustrates to this day that mutations are undirected and lead to selection. That is, specific mutations are not produced because they are needed by the organism but do result in increased hereditary fitness.
The experiment itself is rather simple, which is the beauty of the whole thing. It is often the simplest experiments and simplest observations that lead to large discoveries. This is one of them. In fact, this work was a majory contributor to Salvador Luria and Max Delbruck winning the 1969 Nobel Prize. Their experiment still stands as the gold standard of for demonstrating undirected mutation. Enough with the backpatting.
The initial observation was that when T1 virus was added to cultures of E. coli that the culture would turn from cloudy to clear in about a 20-30 minute time frame. This was due to the virus infecting the cells making them burst. However, after a while the culture would become cloudy again. What they found is that the bacteria that recolonized the culture were resistant to the virus, as were their descendents. They formed two hypotheses to explain this observation: (from the 1943 paper)
1. Mutation: There is a finite probability for any bacterium to mutate during its life time from "sensitive" to "resistant". Every offspring of such a mutant will be resistant, unless reverse mutation occurs. The term "resistant means here that the bacterium will not be killed if exposed to virus, and the possibility of its interaction with virus is left open.
2. Acquired Hereditary Immunity: There is a small finite probability for any bacterium to survive an attack by the virus. Survival of an infection confers immunity not only to the individual but also to its offspring. The probability of survival in the first instance does not run in clones. If we find that a bacterium survives an attack we can not from this information infer that close relatives of it, other than descendants, are likely to survive the attack.
What they did to test these hypotheses was grow several parallel cultures. They would then spread these separate cultures onto separate plates that contained virus. The next day they would count the number of colonies on each plate, with each colony representing a single bacterium at the time of plating. If the mutation hypothesis was right, and if these mutations occurred in a random generation before exposure to virus, then the number of resistant colonies should vary wildly between parallel cultures. As an added control, they plated each culture multiple times so that error introduced in the transfer and plating of bacteria could be controlled for. If the acquired hereditary immunity hypothesis was right, then there should be very little variation in the number of resistant clones in each parallel culture.
What did they find? In any single culture, the number of colonies on multiple plates was nearly the same. However, there was wide variance between any two cultures. This means that the acquired hereditary immunity hypothesis was wrong and the mutation hypothesis was right. What they had discovered was that mutations conferring virus resistance occurred in a random generation before exposure to virus. The variance seen between cultures was due to the randomness of the mutation. If it occurred in an early generation then there would be many resistant colonies upon exposure to phage. If the mutation occurred in a later generation then there would be relatively few colonies. Because any two colonies were grown independent of each other, there was wide variance between any two cultures.
Note: If my explanation is a little hard to follow, you might want to check out the wikipedia page:
http://en.wikipedia.org/wiki/Luria-Delbruck_experiment
You will also notice that this experiment was done a decade or so before DNA was even discovered. It was later found that T1 phage resistance was due to mutations in the tonB gene, a surface protein responsible for vitamin B12 uptake in E. coli. What was happening is that the T1 virus would bind to the tonB protein which allowed the virus to infect the bacterial cell. Mutations in this gene prevented the virus from ever attaching to the bacteria. This is consistent with the observations made by Luria and Delbruck. They observed that one could separate the bacteria from the phage, evidence that the virus did not attach to the bacteria. Another interesting aspect is that tonB has a high "turnover rate". That is, it does not stay on the surface very long and must be replaced on a constant basis. This means that tonB proteins on the surface would not be transferred to descendents and that all tonB had to be made by each offspring. Luria and Delbruck were lucky that their model had these features because it made it possible to test the two hypotheses.
Again, this is not a proof of evolution in general but you seem caught on the topic of natural selection and mutation, so I focused there.