I have gone through the previous posts on the subject in this forum but still have an issue. So please tell me what is wrong with my following (qualitative) reasoning: Assume the standard paradox is replaced with: Twin A at origin of x-axis in free space. Twin B accelerates away to speed near light speed (unspecified) along +ve x-axis. The x-axis has been priorly populated with identical clocks at each integer point and by an observer at each point using Einstein's method of clock synchroniazation such that any observation at points along the x-axis can be measured and then the results sent back to the origin. (I.e they will be 'time-stamped' such that 0 (origin) will know when they occured in his past at his clock time). Now according to S.R. the theory is based on relative velocities not on relative accelerations and the time-dilations observed are calculated using 'v' only. All calculations of any age difference are done this way and many books state that the actual acceleration values are not used in the calculations and can be made arbitrarily large such that Twin B can reach his velocity 'quickly'. Also I assume if we did factor in the acceleration, then we would only need to know the instantaneous relative velocities seen by Twin A at points along the x-axis during the accelerating phase to calculate the true time-dilation values at each point. To Begin: Twin B sets off with a number of identical clocks to Twin A. Upon reaching his target-relative-velocity his clocks appear slowed (and ARE slowed as he changed inertial frames) to observers along the x-axis. Now Twin B then decides to eject one of his clocks in the -ve x-axis direction at an identical acceleration value that he originally did himself such that when this ejected clock reaches relative-velocity-v w.r.t himself, the ejected clock then finds itself stationary w.r.t the x-axis. The ejected clock is now opposite one of the observers on the x-axis. Here's the problem: The ejected clock has initially been time-slowed w.r.t the x-axis; then time-slowed again w.r.t Twin B. So it has endured a compounded time slowing with no mechanism to actually set itself back to it's original 'rate' even though it ends up stationary back on the x-axis. Note. These time dilations must be 'real' and physical as are the Muon decay rates observed to be really time-dilated in the atmosphere of Earth. So from Twin B's frame,the ejected clock is experiencing a REAL time-dilation (changed inertial frames again). Yet the ejected clock is now stationary w.r.t Twin A and should tick at his rate. Now, I understand the formulae involved and the world-line explanation e.t.c. It just seems to me that anything that moves at various velocities w.r.t something else and comes back to rest keeps on compounding and building up time-dilations and will run slower and slower with no re-set?? So what is the answer? Thankyou. P.S. Any measurements done by either Twin on the others frame will be symmetric and both will claim each other is the one who has been time-dilated. Yet the claim made is that the one who actually changed inertial frames (accelerated) endures the REAL net effect of the time-dilation, CONTRARY to the S.R. Theory that it is ONLY relative velocities that are used to do the calculations. If the whole of the x-axis including Twin A were to accelerate to match Twin B's velocity so that they then were stationary w.r.t each other, then it seems that if a different accleration value was used, this would upset any theory that accelerations affect any time-dilation value as different dilations would result. The only way I see a resolution is that time-dilation effects are only calculated by referral and are not real. E.g. The Muon lifetime is not a 'proper' life'time'; it's value is found by taking it's assumed path length in the Earths atmosphere and divided by it's near c value of velocity to give a referred lifetime value. Again, the catch-22 situation. Whatever way I look at it, it makes sense upto a point, only to contradict itself when looking at all the views.