The realization that GRBs are narrowly beamed implied that the actual rate of GRBs is much larger than the observed one .
There are 500 unobserved GRBs for each observed one .
The lack of a clear trigger makes it hard to detect these unobserved GRBs as orphan afterglows .
At late time , hundreds or thousands of years after a GRB , we expect to observe a GRB remnant ( GRBR ) .
These remnants could be distinguished from the more frequent SNRs using their different morphology .
While SNRs are spherical , GRBRs that arise from a highly collimated flow , are expected to be initially nonspehrical .
We ask the question for how long can we identify a GRBR among the more common SNRs ?
Using SPH simulations we follow the evolution of a GRBR and calculate the image of the remnant produced by bremsstrahlung and by synchrotron emission .
We find that the GRBR becomes spherical after \sim 3000 { yr } ( E _ { 51 } / n ) ^ { 1 / 3 } at R \sim 12 { pc } ( E _ { 51 } / n ) ^ { 1 / 3 } , where E _ { 51 } is the initial energy in units of 10 ^ { 51 } { erg } and n is the surrounding ISM number density in { cm } ^ { -3 } .
We expect 0.5 ( E _ { 51 } / n ) ^ { 1 / 3 } non-spherical GRBs per galaxy .
Namely , we expect \sim~ { } 20 non spherical GRBRs with angular sizes \sim \mu arcsec within a distance of 10Mpc .
These results are modified if there is an underlying spherical supernova .
In this case the GRBR will remain spherical only for \sim 150 { yr } ( E _ { 51 } / n ) ^ { 1 / 3 } and the number of non-spherical GRBRs is smaller by a factor of 10 and their size is smaller by a factor of 3 .