Context : Detecting post-merger features of merger remnants is highly dependent on the depth of observation images .
However , it has been poorly discussed how long the post-merger features are visible under different observational conditions .

Aims : We investigate a merger-feature time useful for understanding the morphological transformation of galaxy mergers via numerical simulations .

Methods : We use N-body/hydrodynamic simulations , including gas cooling , star formation , and supernova feedback .
We run a set of simulations with various initial orbital configurations and with progenitor galaxies having different morphological properties mainly for equal-mass mergers .
As reference models , we ran additional simulations for non-equal mass mergers and mergers in a large halo potential .
Mock images using the SDSS r band are synthesized to estimate a merger-feature times and compare it between the merger simulations .

Results : The mock images suggest that the post-merger features involve a small fraction of stars , and the merger-feature time depends on galaxy interactions .
In an isolated environment , the merger-feature time is , on average , \sim 2 times the final coalescence time for a shallow surface bright limit of 25 mag arcsec ^ { -2 } .
For a deeper surface brightness limit of 28 mag arcsec ^ { -2 } , however , the merger-feature time is a factor of two longer , which is why the detection of post-merger features using shallow surveys has been difficult .
Tidal force of a cluster potential is effective in stripping post-merger features out and reduces the merger-feature time .

Conclusions :