Using the high sensitivity of XMM-Newton , we have studied the X-ray emission of the two proto-typical late-stage merger remnants , NGC 3921 and NGC 7252 .
In the case of NGC 7252 , this is complemented by archival Chandra data .
We investigate the nature of the discrete X-ray point source populations and the hot diffuse gas components in these two galaxies , and compare them in the light of their different merger ages and histories .

We detect 3 candidate ultra-luminous X-ray point sources in NGC 3921 and at least 6 in NGC 7252 , for which we have high spatial resolution Chandra data .
These have luminosities ranging from \sim 1.4x10 ^ { 39 } -10 ^ { 40 } erg s ^ { -1 } ( for H _ { 0 } = 75 km s ^ { -1 } 
Mpc ^ { -1 } ) .
We expect these ULXs to be high mass X-ray binaries , associated with the recent star formation in these two galaxies .

Extended hot gas is observed in both galaxies .
We have sufficient counts in the XMM-Newton data to fit two-component hot plasma models to their X-ray spectra and estimate the X-ray luminosities of the hot diffuse gas components to be 2.75x10 ^ { 40 } erg s ^ { -1 } and 2.09x10 ^ { 40 } erg s ^ { -1 } in NGC 3921 and NGC 7252 , respectively .
These luminosities are low compared with the luminosities observed in typical mature elliptical galaxies ( L _ { X } \sim 10 ^ { 41 - 42 } erg s ^ { -1 } ) , into which these merger remnants are expected to evolve .
We do not see evidence that the X-ray halos of these galaxies are currently being regenerated to the masses and luminosities seen in typical elliptical galaxies .
The mass of atomic gas available to fall back into the main bodies of these galaxies and shock-heat to X-ray temperatures is insufficient for this to be the sole halo regeneration mechanism .
We conclude that halo regeneration is most likely a long-term ( > 10 Gyr ) process , occurring predominantly via mass loss from evolving stars , in a sub-sonic outflow stage commencing \sim 2 Gyr after the merging event .