Is Evolution Reversible

The local variation in a population is reversible (micro-evolution) but is macro-evolution reversible? Obviously not exactly but if conditions changed then would evolution back-track?Evolution doesn't have a direction so surely it is possible for it to go backwards?

Yes I understand that it would be almost impossible (but not impossible) to retrace exactly.However my question is really asking "are the previous versions lost or are they still possible". It should be easier to loose "advantages" rather than to gain them.Microevolution is "reversible" but how far back can it go?Ok so I know some people will say that evolution always goes forwards (like time) which is true but I'm asking if conditions changed then would the path be retraced if that gave a benefit. It would be "easier" to do this than to find a different solution (if an old solution worked then use it).

Yes it's a good example.Ok let me see if I can explain this.Looking at your diagram......You said that they started of at the top left as wingless.....however it appears that they started as winged and "immediately" lost their wings. The first winged version appears to be in existence today (top right).From your text it seems that the loss of wings occurred over 50 million years ago.It appears to me that those versions that regained their wings were not evolving wings but were "switching" them back on.I think there's one example in the diagram where they then lost the wings again after regaining them.Loosing wings is not just as simple as first would seem (the interfaces at the wing root need to change).So it looks to me like there were two options available right from the start of your diagram.........wings and wingless. These options carried on right to today and at different stages the options were turned on (I don't think wings or wingless re-evolved).Hence it looks like the options remained for 50 million years.So what I'm saying is that even after 50 million years it is possible to "backtrack". I don't think that any of the changes (winged to wingless) was a new evolution but rather a switching on of two mutually exclusive options. I think if the wings or wingless options had to re-evolve then they would have been significantly different to anything that had been in existence before (is evolution repeatable....I don't think so).I guess that asks the question........how much is "mutation" and how much of what we see is switching on or off of options?

Language "evolution" and the evolution of living things is not the same and it's not really a valid comparison.

I'm not sure whether your example was over 50M years or 50k years as you've suggested both.However if it's 50M then I'd suggest that it's very unlikely for a wing to re-evolve. When a feature is "lost" then the chances of it being duplicated closely is too high imo to be realistic. If it's very, very unlikely for evolution to go "backwards" then it's also very, very unlikely for it to follow the same steps and end up with the "same" solution. Evolution has no foresight - random mutation would not result in a feature re-evolving so there must be some memory involved. That's why I suggested there was a "switch" that "turns" wings or wingless on.If you lost the sprinkler from the end of your hose then the chances of another sprinkler occurring due to evolution would be almost impossible unless there were only a reduced number of possible options available.

I think it's more probable that the information is not lost and hence in this case evolution can reverse something that was done millions of years ago.You mentioned lack of imagination........well I can imagine many many ways that a problem could be addressed and if evolution is random (with selection) then I think that if a wing was "lost" then another competing feature could evolve to replace it. Using the same solutions does not imply randomness.

No I mean that it can reverse previous evolution as in RAZD's example. You can always argue that evolution like time does not go backwards but I meant that features could be dropped and reused. I'm not sure how far back this could go but RAZD's example implies millions of years.

I don't know. Maybe there was some breeding between two types??? However I think the traits remained in the population and if that's true then maybe evolution could regress.

How does selection determine a flying model?Evolution does not have any "knowledge" and hence even if a "flying model" was better, evolution couldn't end up with an almost identical result through random mutation.