Interference seen in a member of an entangled pair

[Dbar_x]

Have you studied at the actual mathematical analysis of how erasing the information in the apparatus is supposed to restore interference

Please see my post in the thread How do you determine that a particle is/was entangled?

 

[zonde]

One more try at realistic explanation.

Say the same experiment http://arxiv.org/abs/quant-ph/9810080
There they say "The coincidence peak was nearly noise-free (SNR > 100) with approximately Gaussian shape and a width (FWHM) of about 2 ns. All data reported here were calculated with a window of 6 ns."
If we draw coincidence graph not for all data but only for 45° coincidences and if it has not one peak but two peaks with equal height (conditional +- and -+ peaks) then ... well, case solved.