Right then; since with vertical dropouts this should not be possible, you need to do some further thinking. When you remove the wheel, is there evidence from the marks left by the hub axle locknuts that one has shifted in the dropout? If the wheel has moved, the only way is for the right-hand locknut to slide forward due to the pull of the chain. If this has not happened, then the wheel has not shifted. If it has happened, then there must be enough slack in the dropout to permit it. If so, then you should be able to recreate the movement by slackening the q/r and pulling the rim at the front of the wheel towards the left-hand chainstay.

What happens?

I am guessing then that the slot in the r/h dropout is wider than in the l/h one, allowing the wheel to sit either centrally in the frame (when the axle sits against the back of the r/h dropout slot and in the centre of the l/h slot) or off to one side (when the axle sits hard against rhe front of the r/h slot and centrally in the l/h slot).

If this is right, the solution is to file the front face of the l/h dropout slot so when the wheel is pulled hard against the front face of both dropouts it sits centrally. I have done this myself on several frames and can confirm both that it works and that on occasion it is needed.

I am referring to the cycle as viewed when riding it. If you are looking at the cycle from the left-hand side, then I mean the left-hand edge of the left-hand dropout... If you think about what is happening and what you want to happen, it should all make sense.

Just get a nice standard 2nd cut file and remove a bit at a time until the wheel sits centrally as described.

Thanks Richard,I just used a bit of sandpaper and it seems to have done the trick.

I was also wondering,should the Q/R be greased or not,opinions seem to differ?